Datasets:

alon-albalak

/

licensed_pubmed

like 0

PMC514652

Background ========== Sphingosine-1-phosphate (S1P) and lysophosphatidic acid (LPA) are phospholipid growth factors which are present in normal serum and plasma. These lipids elicit diverse responses from a wide range of cell types, including enhanced cell survival, cell proliferation, induction of cytoskeletal changes and chemotaxis (reviewed in \[[@B1]-[@B4]\]. Some of these responses reflect activation of G protein-coupled receptors of the endothelial differentiation gene (Edg) family. The Edg receptor family is classified into two clusters based on ligand selectivity: S1P~1/2/3/4/5~(formerly Edg1/5/3/6/8) specifically respond to S1P whilst LPA~1/2/3~(formerly Edg2/4/7) respond to LPA \[[@B5]\]. Members of the S1P receptor family display higher sequence similarity to each other (approximately 40% identity) than to members of the LPA receptor family (approximately 30% identity). These homologies, coupled with observed differences in the structure of S1P and LPA receptor genes, suggest that these receptor families evolved from distinct ancestral genes. The S1P receptors contain a conserved glutamic acid residue present within the third TM that corresponds to glutamine in the LPA receptors. Interaction between distinct functional groups present on S1P and LPA with this residue was shown for the S1P~1~and LPA~1~receptors using computational modelling techniques \[[@B6],[@B7]\] and was demonstrated as the basis for the ligand preference displayed by the receptors. Experimental characterisation confirmed that replacement of glutamic acid with glutamine in S1P~1~changed ligand specificity from S1P to LPA, and the reciprocal mutation in LPA~1~resulted in recognition of both LPA and S1P \[[@B7]\]. In the present study, the role of this residue in determining ligand selectivity for the S1P~4~receptor was examined. Phylogenetic analysis of the Edg family of receptors indicates that S1P~4~is more closely related to other S1P receptors than receptors which respond to LPA. However, S1P~4~lies on the edge of the S1P family cluster and has been shown to bind S1P with lower affinity than other S1P receptors and hence it has been suggested that S1P is not the true endogenous agonist of this receptor \[[@B8]\]. We therefore decided to investigate whether replacement of this residue (E^3.29(122)^) with glutamine conferred LPA-responsiveness to the S1P~4~receptor and hence determine the role of this residue in this lower-affinity S1P receptor. To achieve this, we expressed wild type and E^3.29(122)^Q mutant S1P~4~receptors in CHO-K1 cells and studied responses to lysophospholipids using a \[^35^S\]GTPγS binding assay. Since CHO-K1 cells respond to LPA, we utilised fusion proteins constructed between the S1P~4~receptor and a pertussis toxin-insensitive Gα~i1~(C^351^I) G protein. Expression of these proteins in CHO-K1 cells followed by treatment with pertussis toxin prior to harvest allowed elimination of any signal due to stimulation of endogenous LPA receptors. Within this study, we also examined how the length of the LPA acyl chain affected potency at the mutant S1P~4~receptor, using a panel of naturally occurring LPA analogues. Computational models of complexes between the wild type or mutant S1P~4~receptor and S1P and LPA species were used to provide a molecular interpretation of the experimental findings. Results ======= Human HA-S1P~4~was mutated at position 122 to replace the naturally occurring glutamic acid with glutamine. The mutant and wild type receptors were stably expressed in CHO-K1 cells as in-frame GPCR-G protein fusions with pertussis toxin-insensitive Gα~i1~(C^351^I). Western blotting was used to detect expression of these fusion proteins. Membranes from HA-S1P~4~-Gα~i1~(C^351^I)- or HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I)-transfected cells contained a polypeptide with an apparent molecular mass of approximately 110 kDa, which reacted with anti-HA and anti-SG1 antibodies (Figure [1A](#F1){ref-type="fig"} and [1B](#F1){ref-type="fig"}) and was consistent with expression of the GPCR-G protein fusion. Confirmation of comparable cell-surface expression of these proteins was obtained via FACS analysis using an anti-HA antibody directly conjugated with fluorescein (Figure [1](#F1){ref-type="fig"} Panel III). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Expression of HA-S1P~4~-Gα~i1~and HA-S1P~4~(E^122^Q)-Gα~i1~(C^351^I) in CHO-K1 cells.**Membranes from untransfected CHO-K1 cells (lane 2) and CHO-K1 cells stably expressing HA-S1P~4~-Gα~i1~(C^351^I) (A) or HA-S1P~4~(E^122^Q)-Gα~i1~(C^351^I) (B) were analysed by Western blotting using anti-HA (panel I) or anti-Gα~i1~(panel II) antibodies. Visualisation of immunoreactive proteins was achieved using chemiluminescence after incubation of the blot with appropriate HRP-conjugated secondary antibodies. The position of each HA-S1P~4~fusion protein is indicated by an arrow. Cell-surface expressed HA-S1P~4~receptor was detected by FACS analysis (panel III) using a Fluorescein conjugate of the anti-HA antibody (blue line). Cells were also stained with an isotype matched control antibody (red line). No staining of untransfected CHO-K1 cells was observed using the Fluorescein conjugate of the anti-HA antibody (not shown). Data are presented as overlay histograms and are representative of at least five independent experiments. :::  ::: The response of HA-S1P~4~-Gα~i1~(C^351^I) and HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I) to S1P was assessed using membranes from cells transfected to express these proteins and treated with pertussis toxin prior to harvest. S1P promoted dose-dependent increase in \[^35^S\]GTPγS binding to membranes containing HA-S1P~4~-Gα~i1~(C^351^I) with an EC~50~of 355 nM ± 155 nM (n = 3); in contrast, membranes from HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I)-expressing cells demonstrated severely impaired response to S1P (Figure [2A](#F2){ref-type="fig"}). The EC~50~for S1P stimulation of the HA-S1P~4~-Gα~i1~fusion protein (355 ± 155 nM) was not statistically significantly different from that obtained using the unfused HA-S1P~4~receptor (439 ± 187 nM, Figure [2B](#F2){ref-type="fig"}) and compared favourably with published values for this receptor in HEK293T cells from two different research groups of 270 nM \[[@B9]\] and 790 nM \[[@B10]\]. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Sensitivity of HA-S1P~4~-Gα~i1~(C^351^I) and HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I) to S1P in \[^35^S\]GTPγS binding assay.**A. Membranes from CHO-K1 cells transfected with HA-S1P~4~-Gα~i1~(C^351^I) (*filled squares*) or HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I) (*open circles*) which had been cultured in the presence of 100 ng/mL pertussis toxin for 24 hours prior to harvest were stimulated with varying concentrations of S1P for 30 minutes at 30°C in the \[^35^S\]GTPγS binding assay. Gα~i~G proteins were immunoprecipitated after solubilisation and preclearance with non-immune serum. Data are the mean of three determinations ± SEM from a single experiment, and are representative of three such experiments performed. B. Dose-dependent stimulation of wild-type HA-S1P~4~(*filled squares*) and HA-S1P~4~-Gα~i1~fusion protein (*filled triangles*) by S1P measured as described in panel A. These data are representative of three such experiments performed and analysis of mean EC~50~values obtained for each protein showed them to be not statistically different. :::  ::: Structure activity relationships were determined by the \[^35^S\]GTPγS assay for S1P~4~or its E^3.29(122)^Q mutant using S1P and LPA species with 14:0, 16:0 and 18:1 acyl chains at a single (10 μM) concentration. Of the lysophospholipids tested, only S1P induced a strong response over basal levels (approximately 48% ± 5%) in membranes containing HA-S1P~4~-Gα~i1~(Figure [3A](#F3){ref-type="fig"}), whilst 18:1 LPA did not stimulate a statistically significant response; weak stimulation of \[^35^S\]GTPγS binding was observed with 14:0 and 16:0 LPA (approximately 11% ± 3% above basal in each case). In contrast, membranes containing HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I) showed at least weak response to each ligand (Figure [3A](#F3){ref-type="fig"}). The weakest agonist at the E^3.29(122)^Q mutant S1P~4~receptor was 18:1 LPA, which produced only 14% ± 2% stimulation over basal. S1P and 16:0 LPA gave approximately 21% ± 4% and 19% ± 4% stimulation over basal response, respectively. The best agonist for the E^3.29(122)^Q mutant was 14:0 LPA, which gave a 40% ± 2% enhancement over basal levels. The stimulation promoted by 14:0 LPA was statistically different from that produced by 18:1 LPA (p \< 0.01). Sensitivity of the receptor to stimulation by each form of LPA, and particularly 14:0 LPA, was markedly increased after introduction of the E^3.29(122)^Q mutation and indicated that this position was important in influencing HA-S1P~4~ligand preference. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Ligand preference of HA-S1P~4~(E^122^Q)-Gα~i1~(C^351^I) in \[^35^S\]GTPγS binding assay.**Membranes were stimulated for 30 minutes at 30°C with lysophospholipid ligands in the \[^35^S\]GTPγS binding assay and Gα~i~G proteins immunoprecipitated after solubilisation and preclearance with non-immune serum. (A) Membranes from CHO-K1 cells transfected to express HA-S1P~4~-Gα~i1~(C^351^I) or the mutant HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I) and incubated with 100 ng/mL pertussis toxin for 24 hours prior to harvest, were left untreated (basal), or treated with vehicle or 10 μM concentrations of 18:1 LPA, 16:0 LPA, 14:0 LPA or S1P. Data are the mean of three determinations ± SEM from a single experiment and are representative of three such experiments performed. Statistical significance from the basal responses of each set of membranes tested is denoted by \* (P \< 0.05) or \*\* (P \< 0.01); \#\# denotes statistical significance from the response to 18:1 LPA (P \< 0.01) for the HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I)-transfected membranes. (B) Membranes from CHO-K1 cells transfected to express HA-S1P~4~(E^122^Q)-Gα~i1~(C^351^I) and incubated with 100 ng/mL pertussis toxin for 24 hours prior to harvest, were stimulated with various concentrations of S1P (*crosses*), 18:1 LPA (*open circles*) or 14:0 LPA (*filled triangles*). Data are the mean of three determinations ± SEM and are representative of three such determinations performed. :::  ::: These results indicate that introduction of the E^3.29(122)^Q mutation in the S1P~4~receptor confers LPA-responsiveness, and that a short form of LPA was a more effective agonist than the intermediate and longer forms, when tested at this single concentration. Dose response curves were constructed for ligand-induced activation of the E^3.29(122)^Q S1P~4~mutant by the 14:0 and 18:1 forms of LPA as well as S1P (Figure [3B](#F3){ref-type="fig"}). An EC~50~could only be determined for the 14:0 form of LPA as S1P and 18:1 LPA caused minimal stimulation at only the highest concentration tested. The EC~50~value for activation of HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I) was calculated to be 3.8 ± 1.4 μM. However, since a plateau of maximal stimulation was not achieved, interpretation of this EC~50~value needs caution. This result clearly showed that 14:0 LPA was a weak agonist of HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I) and hence confirmed the involvement of residue 122 in S1P~4~ligand preference. Similar results were obtained using a second CHO-K1 clone expressing this fusion protein (not shown). Computational modelling of the S1P complex with the wild type S1P~4~receptor identifies the best S1P binding site within the TM with the phosphate group at the extracellular end (Figure [4A](#F4){ref-type="fig"}). Ion pairs appear between the phosphate group of S1P and two cationic amino acids, R^3.28(121)^and K^5.38(202)^. An additional ion pair occurs between the cationic ammonium of S1P and E^3.29(122)^. Hydrophobic residues from TM2, TM3, TM5 and TM6 line the binding pocket and surround the alkyl chain of S1P. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Computational models of wild type S1P~4~and its E3.29(122)Q mutant with S1P and LPA species.**Computational models of the complexes between the wild type S1P~4~or its E^3.29(122)^Q mutant with S1P or various LPA species generated by Autodock 3.0 and minimised using the MMFF94 forcefield in the MOE program. Complexes in each panel are shown from the same viewpoint with the extracellular end of the receptors oriented to the top of the figure. Standard element color codes are used with grey, white red, blue and magenta representing carbon, hydrogen, oxygen, nitrogen and phosphorous. Ribbons are shaded from red at the amino-terminus to blue at the carboxy-terminus. (A) Model of the complex between S1P (spacefilling) and the wild type S1P~4~receptor. Residues in the receptor involved in ion pairs with S1P are shown as stick models and labelled. (B) Superimposition of the wild type S1P~4~complex with S1P (orange) and the E^3.29(122)^Q S1P~4~mutant complex with 14:0 LPA (green). For clarity, the only position at which the modelled amino acid position is shown for both receptor models is 3.29(122). Other residues had very similar optimised positions in the two model structures. (C) Superimposition of wild type S1P~4~complexes with 18:1 LPA (cyan), 16:0 LPA (yellow) and 14:0 LPA (green) on E^3.29(122)^Q mutant complexes with 18:1 LPA (blue-green), 16:0 LPA (gold) and S1P (orange). For clarity, the only position at which modelled amino acid position is shown for both the wild type and mutant receptor models is 3.29(122). Other residues had very similar optimised positions in all model structures. (D) Space-filling models which represent the minimised extended conformation of each structure were constructed using SYBYL 6.9 software (Tripos Inc., St. Louis, MO., U.S.A.). The distance between phosphorus and terminal carbon atoms was predicted for each structure listed from top to bottom: 18:1 LPA, 27.0 Å; 16:0 LPA, 26.7 Å; 14:0 LPA, 24.2 Å; S1P, 24.0 Å. :::  ::: The best complex of 14:0 LPA in the E^3.29(122)^Q S1P~4~mutant receptor model has striking similarity to the best complex of S1P in the wild type S1P~4~receptor model (Figure [4B](#F4){ref-type="fig"}). Both models demonstrate ion pairing between the phosphate group and two cationic amino acids, R^3.28(121)^and K^5.38(202)^. Each ligand interacts with the amino acid at position 3.29(122), S1P by an ion pair with the carboxylate of the wild type glutamate and 14:0 LPA by a hydrogen bond with the mutated glutamine. Multiple hydrophobic residues surround the nonpolar tails of the lipid ligands. The superimposition of the two complexes (Figure [4B](#F4){ref-type="fig"}) also demonstrates that the ligands occupy almost identical volumes. Common interactions and overlap volumes are qualitatively consistent with the experimental findings that these ligands give similar 48% and 40% maximal stimulation over basal for S1P at the wild type and 14:0 LPA at the mutant receptor, respectively. In contrast to the complexes of 14:0 LPA with E^3.29(122)^Q S1P~4~and S1P with wild type S1P~4~, the remaining complexes show much less common volume (Figure [4C](#F4){ref-type="fig"}). Most complexes exhibit the phosphate interactions described for 14:0 LPA with E^3.29(122)^Q S1P~4~and S1P with wild type S1P~4~. Of particular interest is the observation that the best complexes generated by Autodock for the 18:1 LPA species with wild type S1P~4~has a very high positive van der Waals interaction energy, \> 3000 kcal/mol, compared to values well under 200 kcal/mol for every other complex studied. In the best complexes found for 16:0 LPA and 18:1 LPA in both constructs, the terminal six to eight carbons of the hydrophobic tails fold into L-shaped conformations quite different from the extended conformations observed in the S1P complex with wild type S1P~4~or the 14:0 LPA complex with the E^3.29(122)^Q S1P~4~mutant. The terminal carbons in several complexes curl out of the receptor between TM5 and TM6 (Figure [4C](#F4){ref-type="fig"}) due to the restricted length of the binding pocket. These results suggest that the complete lack of S1P~4~activation in response to 18:1 LPA is likely due to failure to form a complex. The strongest complexes formed, S1P with wild type S1P~4~and 14:0 LPA with the E^3.29(122)^Q S1P~4~mutant, have complementary interactions with the residue at position 3.29(122). These strong complexes give the most robust activation. Weak complexes are formed for other combinations due to mismatched interactions with position 3.29(122) or excessive length of the hydrophobic tail. The presence of hydrophobic tails of 16:0 or 18:1 LPA between transmembrane domains may additionally impair the conformational change necessary for full agonist responses. Discussion ========== Parental CHO-K1 cells respond to LPA in functional assays, reflecting expression of endogenous LPA~1~(G. Holdsworth, *et al*., manuscript in preparation). For this reason, fusion proteins between wild type or mutant HA-S1P~4~and the pertussis toxin-insensitive Gα~i1~(C^351^I) G protein were used in these studies. Expression of these proteins in CHO-K1 cells followed by treatment with pertussis toxin prior to harvest allowed elimination of any signal due to stimulation of endogenous LPA receptors. McAllister *et al*. \[[@B11]\] (.(adopted a similar approach for studies of the LPA~1~receptor. We examined the role of residue E^3.29(122)^in controlling S1P~4~ligand selectivity using functional and computational methods. This residue, which is conserved throughout the S1P receptors, has been shown to control ligand specificity for the related S1P~1~receptor \[[@B7]\]. Introduction of the E^3.29(122)^Q mutation severely affected the response of S1P~4~to S1P: in dose-response experiments S1P caused minimal stimulation at only the highest concentration of ligand used. This is in agreement with published observations for activation of the equivalent S1P~1~mutant \[[@B7]\]. 14:0 LPA was able to induce dose-dependent stimulation of S1P~4~(E^3.29(122)^Q) with an EC~50~of approximately 3.8 μM but only promoted minimal stimulation of the wild type S1P~4~receptor. The modelled complexes of 14:0 LPA with E^3.29(122)^Q S1P~4~and S1P with wild type S1P~4~demonstrate nearly identical volumes occupied by the two ligands and very similar interactions between these ligands and their respective receptors. Of particular importance are amino acid residues at positions 3.28(121), 3.29(122) and 5.38(202), which either ion pair with the phosphate or interact with the 2-amino or 2-hydroxyl group in S1P and 14:0 LPA, respectively. The importance of interactions with amino acids at positions 3.28 and 3.29 has been previously noted for the S1P~1~\[[@B6],[@B7]\] and LPA~1,2,3~\[[@B7],[@B12]\] receptors. The S1P~4~receptor exhibits marked constitutive (agonist-independent) activity (Figure [5](#F5){ref-type="fig"}) which was unaffected by the introduction of the E^3.29(122)^Q mutation (data not shown). This indicates that the mutation perturbs S1P recognition without affecting the ability of the receptor to spontaneously adopt an active conformation. Similar observations have been reported for the β~2~AR, where a mutation in the sixth transmembrane domain abolished agonist activation but not constitutive activity \[[@B13]\]. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Constitutive activity of HA-S1P~4~and HA-S1P~4~-Gα~i1~(C^351^I).**Comparison of basal, vehicle and stimulated (10 μM S1P) GTPγS binding for membranes prepared from CHO cells transfected to express HA-S1P~4~or HA-S1P~4~fused to PTx insensitive Gα~i1~. Where indicated, cells were treated with 100ng/ml PTx for 24 hours prior to harvesting. PTx treatment of HA-S1P~4~-transfected membranes prevented activation by S1P and also caused a dramatic reduction in basal signalling, indicative of constitutive activity. In contrast, when the HA-S1P~4~-Gα~i1~fusion-expressing membranes were treated with PTx, there was only a slight reduction in basal signalling and the receptor still responded to exogenous S1P, indicating that the receptor signalled via the tethered, PTx-insensitive G protein. Basal signalling in PTx-treated HA-S1P~4~-Gα~i1~-expressing membranes exceeded that seen with membranes from untransfected CHO cells. :::  ::: Unlike S1P, which exists as a single species *in vivo*, the term LPA actually refers to a family of molecules that take the general form 1-*o*-acyl-2-hydroxy-*sn*-glyceryl-3-phosphate. Naturally occurring forms of LPA contain acyl chains of differing lengths, with differing degrees of saturation. Investigations into the effect of the length and degree of saturation of the acyl chain of LPA have been undertaken for the LPA receptors \[[@B14],[@B15]\], but limited SAR information is available for S1P receptors (22). The LPA-responsive E^3.29(122)^Q S1P~4~mutant facilitates structure activity relationship (SAR) studies due to the greater availability of LPA analogs relative to S1P analogs. Comparison of space-filling models of the structures of S1P and three analogues of LPA (Figure [4D](#F4){ref-type="fig"}) revealed that 14:0 LPA most closely resembled S1P in terms of apparent length. \[^35^S\]GTPγS binding assays demonstrated greater agonist activity of 14:0 LPA at the mutant receptor relative to 18:1 or 16:0 LPA. This SAR indicates a length restriction for the S1P~4~agonist binding site. Model complexes of 16:0 and 18:1 LPA contained alkyl chains that fold at the bottom of the binding pocket, defined by a cluster of hydrophobic amino acids. Three of these differ either in position of sidechain branching or size relative to LPA receptors and the other S1P receptors. Position 2.46, I88 in S1P~4~, is leucine in LPA~1--3~and other S1P receptors. Residue I6.40(256) is larger than the valine found in the other four S1P receptors, LPA~1~and LPA~3~. Finally, I7.51(305) corresponds to the smaller valine in S1P~2~and S1P~3~and the much smaller alanine in LPA~2~. These findings provide a molecular explanation for a similar SAR observed using para-alkyl amide analogs of S1P \[[@B9]\]. SAR obtained with the S1P~4~mutant are in contrast to that shown by LPA receptors, which exhibit the general trend of 18:1 ≥ 16:0 \> 14:0 for potency and maximal stimulation \[[@B15]\]. Since mutation of residue 122 in the S1P~4~receptor from the naturally occurring glutamic acid to glutamine conferred responsiveness to 14:0 LPA and severely affected responses to S1P, our observations support the hypothesis that this conserved residue in the third transmembrane domain of the S1P receptors is involved in ligand recognition. This is in contrast to a recent paper describing models of several GPCRs, including S1P~4~, which had been generated using novel first principle methods \[[@B16]\]. In this model of S1P~4~, interactions between S1P and residues T^7.34(127)^and W^7.37(291)^and E^7.30(284)^were observed. Interaction of E^7.30(284)^with the ammonium group of S1P appeared to control ligand selectivity since the other residues appeared to interact with the phosphate group, which is present on both LPA and S1P. It is therefore surprising that none of these residues are conserved throughout the S1P or LPA receptor families. The data presented here support the assertion that glutamic acid residue 3.29 present in the third transmembrane domain of the S1P receptors controls ligand selectivity and suggest that the S1P~4~model described by Vaidehi *et al*. \[[@B16]\] is inaccurate. The current study provides new information for the development of more selective S1P receptor agonists. In particular, an S1P analog with its hydrophobic chain extended by either 2 or 4 carbons would be a very poor agonist of the S1P~4~receptor. On the other hand, the activation of the S1P~1~-E^3.29(121)^Q mutant by 18:1 LPA \[[@B7]\] indicates that a chain-extended S1P analog should retain agonist activity at the S1P~1~receptor. S1P receptor agonists with differing selectivity profiles will be useful tools to more completely map the physiological and pathophysiological roles of these receptors. Conclusions =========== These studies confirm that glutamic acid residue 3.29, present in the third transmembrane domain of the S1P receptors is important for the selective recognition of S1P, versus the closely related lipid, LPA. Mutation of E3.29 to glutamine diminished response to S1P and allowed structure activity studies using the diverse available LPA species. The mutant S1P~4~receptor is stimulated most strongly by LPA 14:0 and is not activated by the longer LPA 18:1, in contrast with a previous report on the analogous S1P~1~receptor mutant that responded to LPA 18:1. Thus the S1P~4~receptor ligand binding pocket is shorter in length than the S1P~1~ligand binding pocket. Methods ======= Residue nomenclature -------------------- Amino acids within the TM of S1P~4~can be assigned index positions to facilitate comparison between receptors with different numbers of amino acids, as described by Weinstein and coworkers \[[@B17]\]. An index position is in the format x.xx. The first number denotes the TM in which the residue appears. The second number indicates the position of that residue relative to the most highly conserved residue in that TM which is arbitrarily assigned position 50. E3.29, then, indicates the relative position of glutamate 122 in TM3 relative to the highly conserved arginine 143 in the E(D)RY motif which is assigned index position 3.50 \[[@B17]\]. Materials --------- Materials for tissue culture were supplied by Invitrogen Ltd. (Paisley, Scotland, U.K.). Foetal bovine serum was obtained from Helena Biosciences Ltd., (Sunderland, U.K.) or PAA Labs GmbH., (Linz, Austria). Pertussis toxin was purchased from CN Biosciences Ltd., (Nottingham, U.K.). Lysophosphatidic acid (18:1, 16:0 and 14:0) and S1P were from Avanti Polar Lipids Inc., (Alabaster, AL., U.S.A.). The SG1 antiserum was produced previously \[[@B18]\]. All other chemicals were from Sigma Aldrich Company Ltd., (Gillingham, Dorset, U.K.) or BDH Ltd., (Poole, Dorset, U.K.) unless stated otherwise. Construction of receptor expression plasmids -------------------------------------------- The S1P~4~coding sequence was cloned from a human PBMC cDNA library using the sense primer 5\'-GAGAGA**GCGGCCGC**CACCATG[TATCCATATGATGTTCCAGATTATGCT]{.underline}AACGCCACGGGGACCCCGGTG-3\', which contains a *Not*I restriction site (bold) and the haemagglutinin HA epitope tag (YPYDVPVYA, underlined) immediately after the initiator methionine, and the antisense primer 5\'-GAGAGA**GAATTC**[GGC]{.underline}GATGCTCCGCACGCTGGAGATG-3\', which contains an *Eco*RI restriction site (bold) and changes the S1P~4~stop codon to alanine (underlined). A C^351^I mutant of the Gα~i1~G protein (previously produced, \[[@B19]\]) was amplified using PCR with the sense cloning primer 5\'-GAGAGA**GAATTC**GCCA CCATGGGCTGCACACTGAGCG-3\', which contains the *Eco*RI restriction site (bold), and the antisense cloning primer 5\'-GAGAGA**GGATCC**TTAGAAGAGACCGATGTCTTTTA G-3\', which contains a *Bam*HI restriction site (bold). After digestion of each PCR product with the appropriate restriction enzymes, fragments were ligated into the pIRESpuro mammalian expression vector (Invitrogen Ltd.) to generate an in-frame fusion between HA-S1P~4~and Gα~i1~(C^351^I). The E^3.29(122)^Q mutation was introduced into the S1P~4~sequence in parallel PCR reactions. Complementary oligonucleotides were designed across the residue which was to be mutated such that each primer contained the necessary base change to mutate residue 122 to glutamine (underlined in each primer): sense mutational primer: 5\'-CAGTGGTTCCTACGG[CAG]{.underline}GGCCTGCTCTTCAC-3\'; antisense mutational primer: 5\'-GTGAAGAGCAGGCC[CTG]{.underline}CCGTAGGAACCACTG-3\'. Mutational sense or antisense primers were used in parallel PCR reactions with the appropriate antisense or sense cloning primer, with HA-S1P~4~plasmid DNA as template. Equimolar amounts of each purified PCR product were mixed and amplified in a further reaction, using the cloning primers described above. The resultant product was digested with the appropriate restriction enzymes and ligated with the Gα~i1~sequence in the pIRESpuro expression vector to generate an in-frame fusion between HA-S1P~4~(E^3.29(122)^Q) and Gα~i1~(C^351^I). Cell culture and transfection ----------------------------- CHO-K1 cells were maintained at 37°C with 5% CO~2~in Dulbecco\'s modified Eagle\'s medium (DMEM), supplemented with 10% foetal bovine serum (FBS), 2 mM L-glutamine and non-essential amino acids. Sub-confluent cell monolayers were stably transfected to express either HA-S1P~4~-Gα~i1~(C^351^I) or HA-S1P~4~(E^3.29(122)^Q)-Gα~i1~(C^351^I) fusion proteins using Lipofectamine reagent (Invitrogen). 72 hours post-transfection, cells were seeded in media supplemented with 7.5 μg/mL puromycin and the resultant clones examined for expression of cell surface receptor using FACS analysis. Clonal cell lines were expanded in complete DMEM containing 7.5 μg/mL puromycin and were transferred to serum free DMEM approximately 24 hours prior to harvesting. Where indicated, 100 ng/mL pertussis toxin was included in the serum free medium. It should be noted that we initially expressed S1P~4~in RH7777 cells, which are unresponsive to S1P and LPA and have been commonly used for studies of Edg family receptors \[[@B20]\]. Unfortunately, our attempts to detect activation of S1P~4~expressed in these cells using a variety of functional assays were unsuccessful. Therefore, we used CHO-K1 cells as an alternative host in these studies; expression of functional S1P~4~in CHO-K1 cells has also been reported by Mandala et al. \[[@B21]\]. FACS analysis ------------- The amino-terminal HA-epitope tag was detected using a fluorescein conjugate of the anti-HA antibody, clone 3F10 (Roche Molecular Biochemicals Ltd., Lewes, U.K.). Cells were harvested non-enzymatically and washed with FACS buffer (PBS containing 3% FBS and 0.1% NaN~3~) then stained with the 3F10 antibody (or an isotype matched control) for 40 minutes at 4°C in the dark. After washing with FACS buffer, cells were analysed using a FACScalibur flow cytometer (BD Biosciences, Oxon., U.K.). Preparation of cell membranes ----------------------------- Cells were harvested non-enzymatically, washed with PBS and resuspended in \"assay buffer\" (20 mM Hepes, pH 7.4, 3 mM MgCl~2~, 100 mM NaCl), supplemented with \"complete\" protease inhibitors (Roche Molecular Biochemicals Ltd.). Cells were homogenised in a nitrogen cavitation chamber (500 psi for 15 minutes). Unbroken cells and nuclei were pelleted by centrifugation (500 × *g*, 10 minutes, 4°C) and the supernatant fraction was centrifuged at 45,000 × *g*for 45 minutes at 4°C. Membrane pellets were resuspended in assay buffer, titurated through a fine gauge needle and stored at -80°C until required. Immunoblot analysis ------------------- Samples were resolved by SDS-Page on 4--20% Tris-Glycine gels (Invitrogen) and were transferred to Immobilon-P membrane (Millipore Ltd., Herts., U.K.). The membrane was blocked using 2.5% Marvel in PBS before incubating with primary antibodies which had been diluted in PBS/0.1% Tween-20 containing 1% Marvel. The high affinity rat anti-HA antibody was diluted 1 in 500; the anti-G~αi1~antibody (Autogen Bioclear Ltd., Wilts., U.K.) was diluted 1 in 1000. Immunoreactivity was detected using an appropriate horseradish peroxidase-conjugated secondary antibody, diluted 1 in 10,000 in PBS/0.1% Tween-20 containing 1% Marvel, followed by detection using SuperSignal reagents (Perbio Science Ltd., Cheshire, U.K.). \[^35^S\]GTPγS binding assay ---------------------------- \[^35^S\]GTPγS binding experiments were performed essentially as described previously \[[@B22]\]. Briefly, membranes were incubated with or without the indicated ligand for 30 minutes at 30°C in assay buffer containing \[^35^S\]GTPγS (100 nCi/point), saponin (20 μg/point) and 0.1 μM GDP. 18:1 LPA was prepared as a 2 mM DMSO stock whilst 16:0 and 14:0 LPA were prepared as 2 mM stock solutions in 1:1 ethanol:water per supplier recommendation due to their poor solubility in DMSO. S1P had previously been dispensed as thin film aliquots (dissolved in MeOH and the solvent evaporated under nitrogen) in brown glass vials and stored at -70°C prior to use. Lipids (S1P or LPA forms) were diluted in assay buffer containing 1% fatty acid free BSA, such that the final concentration of BSA in the assay was 0.1%. Following incubation, membrane protein was solubilised with 1.25% NP-40 and 0.4% SDS and after pre-clearance using non-immune serum, Gα~i1/2~subunits were immunoprecipitated with SG1 antiserum, used at a dilution of 1 in 200. Non-specific binding was determined by the addition of 100 μM GTPγS. Bound radioactivity was measured using liquid scintillation counting. Experimental data analysis -------------------------- Numerical data are expressed as means ± standard error, shown as error bars in the appropriate figures. Statistical comparisons were made using one-way ANOVA with Dunnett\'s multiple comparison post test. Receptor model development -------------------------- A model of human S1P~4~(GenBank™ accession number AAP84350) was developed by homology to the experimentally-validated model of S1P~1~\[[@B23]\]. Alignment of the S1P receptor sequences was performed using the MOE software package (version 2003. 01 ed. Chemical Computing Group, Montreal, Canada). The alignment was optimised by the manual removal of gaps within the TM, and alignment in the region of TM5 was shifted one position to correctly orient K5.38(202) toward the interior of the helical bundle (Pham, *et al*., unpublished data). A preliminary model was generated by homology modelling using default parameters and subsequently manually refined to optimise interhelical hydrogen bonding. Cis-amide bonds present in the loop regions were converted to the trans conformation by manual rotation followed by the minimisation of two residues on either side of the amide linkage to a root mean square (RMS) gradient of 0.1 kcal/mol·Å using the MMFF94 forcefield \[[@B24]\]. After these manual refinements, the receptor model was optimised using the MMFF94 forcefield to an RMS gradient of 0.1 kcal/mol·Å. A model of S1P~4~with the E^3.29(122)^Q mutation was developed by performing the appropriate mutation in MOE, and saturating the residue with hydrogen atoms. To allow the sidechains of the other residues in the binding pocket to adapt to the presence of the new moiety, the backbone atoms of the receptor were fixed and the receptor was optimised to an RMS gradient of 0.1 kcal/mol·Å using the MMFF94 forcefield \[[@B24]\]. Ligand model development ------------------------ Computational models of the naturally-occurring stereoisomers of 14:0 LPA, 16:0 LPA, 18:1 LPA, and S1P were built using the MOE software package. The -1 ionization state for the phosphate functionality was chosen for all ligands, and the +1 ionization state was chosen for the amine moiety of S1P. Previous docking studies using the -2 ionization state for phosphate in related systems yield essentially identical geometries as studies using the -1 ionization state. These ligands were geometry optimised using the MMFF94 force field \[[@B24]\]. Docking ------- Using the AUTODOCK 3.0 software package \[[@B25]\], 14:0 LPA, 16:0 LPA, 18:1 LPA, and S1P were docked into the S1P~4~wild type and S1P~4~E^3.29(122)^Q mutant receptor models. Each docking box was centered near F^3.33(126)^with dimensions of 30.75 × 23.25 × 23.25 or 32.25 × 23.25 × 23.25 Å for shorter (S1P and 14:0 LPA) or longer (16:0 and 18:1 LPA) ligands, respectively. At least 20 putative complexes were generated for each receptor:ligand pair using docking parameters at default values with the exception of the number of energy evaluations (2.5 × 10^8^), generations (10000) and maximum iterations (3000). Resultant complexes were evaluated based on final docked energy, Van der Waals interaction energies from the MMFF94 forcefield as well as visual analysis. The complexes with the lowest final docked energies and others of interest were geometry optimised using the MMFF94 force field \[[@B24]\], and the lowest energy complex after minimisation was chosen as the final complex structure. Abbreviations ============= CHO, Chinese hamster ovary; Edg, endothelial differentiation gene; ERK, extracellularly regulated kinase; FACS, fluorescence activated cell sorter; G protein, guanine nucleotide-binding protein; GPCR, G protein-coupled receptor; HA, haemagglutinin; LPA, lysophosphatidic acid; MAP, mitogen-activated protein kinase; PBMC, peripheral blood mononuclear cell; PTx, pertussis toxin; SIP, sphingosine-1-phosphate; TM, transmembrane domain Authors\' contributions ======================= G Holdsworth performed and interpreted all studies with experimental S1P~4~fusion proteins and drafted the manuscript. D Osborne performed and interpreted docking studies to generate all mutant complexes with all LPA species and S1P and all wild type complexes with LPA species. TC Pham generated the homology model of the human S1P~4~receptor. J Fells performed docking studies of S1P with the wild type S1P~4~receptor. G Hutchinson and G Milligan participated in the design and coordination of the experimental studies with S1P~4~fusion proteins. A Parrill participated in the design and coordination of the modelling studies and edited the manuscript. Acknowledgments =============== We gratefully acknowledge the assistance of Jim Turner (Celltech R&D Ltd.) for production of lipid space-filling models and Gabor Tigyi (University of Tennessee Health Sciences Center) for critically reading the manuscript. This work was supported in part by grants from NIH (1 RO1 CA92160-01) and the American Heart Association (awards 0050006N and 0355199B). The Chemical Computing Group generously donated the MOE program.

PubMed Central

2024-06-05T03:55:47.173847

2004-8-6

PMC514697

Background ========== Biological relevance -------------------- Pseudoknots have been shown to be functionally relevant in many RNA mediated processes. Examples are the self-splicing group I introns \[[@B1]\], ribosomal RNAs, or RNaseP. Recently, pseudoknots were located in prion proteins of humans, and confirmed for many other species \[[@B2]\]. With the current increased interest in the universe of RNA functions \[[@B3]\], algorithmic support for analysing structures that include pseudoknots is much in demand. Previous algorithmic work ------------------------- Well established algorithms for the prediction of RNA secondary structures (MFOLD \[[@B4]\], RNAfold \[[@B5]\]) are commonly based on a thermodynamic model \[[@B6]\], returning a structure of minimal free energy, called MFE-structure for short. In spite of their importance, pseudoknots are excluded from consideration by these programs for reasons of computational complexity: While folding a sequence of length *n*into unknotted structures requires *O*(*n*^3^) time and *O*(*n*^2^) space, finding the best structure including arbitrary pseudoknots has been proved to be NP-complete \[[@B7],[@B8]\]. In fact, the proof given in \[[@B8]\] uses a scoring scheme based on adjacent base pairs only, simpler than the MFE model because it neglects entropic energies from loops. These complexity results leave two routes to achieve practical algorithms. The first route is to consider pseudoknots in full generality, but resort to an even more simplistic energy model. An *O*(*n*^4^) time and *O*(*n*^3^) space algorithm for base pair maximization has been given in \[[@B7]\], and an *O*(*n*^3^) time algorithm based on maximum weight matching in \[[@B9]\] and \[[@B10]\]. The second route is the one followed here: We retain the established thermodynamic model, but restrict to a more tractable subclass of pseudoknots. For some quite general classes of pseudoknots, polynomial time algorithms have been designed: Rivas and Eddy achieve *O*(*n*^6^) time and *O*(*n*^4^) space \[[@B11]\]. This algorithm is available, and, in spite of the high computational cost, it is actually used in practice. We build upon this work and shall call it *pknotsRE*for later reference. Further improvements have been shown to be possible for yet more restricted classes, e.g. the non-recursive simple pseudoknots considered by Lyngsø and Pedersen \[[@B12]\] with *O*(*n*^5^) time and *O*(*n*^4^) space, but to our knowledge, no implementations are available. Recently, an *O*(*n*^4^) time and *O*(*n*^3^) space algorithm based on the technique of \[[@B7]\], that uses a thermodynamic model has been reported in \[[@B13]\]. While it handles simple pseudoknots consisting of more than two helices, it is restricted to non-recursive pseudoknots. Thus, this class of pseudoknots and the class presented here have a nonempty intersection, but neither of them contains the other. Our contributions ----------------- The new contributions reported here are the following: • We present an algorithm *pknotsRG*for folding RNA secondary structures including pseudoknots under the MFE model which requires *O*(*n*^4^) time and *O*(*n*^2^) space. • The algorithm considers the class of simple recursive pseudoknots, further restricted by three rules of canonization. Each simple recursive pseudoknot has a canonical representative that is recognized by *pknotsRG*. • While this class is more restricted than the one of the Rivas/Eddy algorithm, practical evaluation shows that our algorithm finds the same pseudoknots, while the length range of tractable sequences is increased significantly. • We provide an evaluation of the class of pseudoknots introduced here against known examples from the literature. • We perform a rigorous evaluation of our algorithm on 212 sequences from PseudoBase \[[@B14]\] plus 7 other structures and compare our results with those obtained with *RNAfold*and, where feasible, with *pknotsRE*. Results ======= It is not easy to relate the classes of pseudoknots recognized by the different algorithms mentioned above. We refer the reader to the review by Lyngsø and Pedersen \[[@B8]\], which compares these classes by means of examples. The starting point of our work is the algorithm *pknotsRE*by Rivas and Eddy. It recognizes pseudoknots that can be nested and can have unlimited chains of helices involved in crosswise interactions. The drawback of this powerful, but computationally expensive algorithm is the following paradox: Pseudoknots with complex helix interactions naturally require longer primary sequence than simpler ones. The high runtime complexity of *O*(*n*^6^), however, as well as the space consumption of *O*(*n*^4^) restricts the use of this algorithm to a maximal sequence length of around 150 nucleotides. Most of the pseudoknots predicted belong to a much simpler structural class and do not exhibit chains of crosswise interactions. The algorithm developed here achieves time complexity *O*(*n*^4^) and space complexity *O*(*n*^2^). The runtime improvement, compared to *pknotsRE*, results from an idea of canonization, while the space improvement results from disallowing chained pseudoknots. These improvements extend the range of tractable sequences to a length up to 800 nucleotides, and we can locate pseudoknots up to this size in even longer sequences. Simple recursive pseudoknots ---------------------------- Following the terminology of \[[@B7]\], a *simple*pseudoknot is a crosswise interaction of two helices, as shown in Figure [1](#F1){ref-type="fig"}. In simple *recursive*pseudoknots, we allow the unpaired strands *u, v, w*in a simple pseudoknot to fold internally in an arbitrary way, including simple recursive pseudoknots. Let us call this class sr-PK. More complex knotted structures like triple crossing helices or kissing hairpins, as shown in Figure [4](#F4){ref-type="fig"}, are excluded from sr-PK. We will show later how they can be integrated in our approach and outline the increased computational cost of doing so. For the main part of this paper, we concentrate on the class sr-PK. Anticipating the complexity of a DP algorithm --------------------------------------------- Thermodynamic RNA folding is implemented via dynamic programming (DP). We start with a semi-formal discussion of how to estimate the efficiency of a DP algorithm for folding (or any kind of motif search) *before*it is written in detail. We consider elements of RNA structure as sequence motifs of different types: hairpins, bulges, multiloops, etc. The following notation is taken from the algebraic dynamic programming approach \[[@B15]\]. By an equation m = f \<\<\< a \~\~\~ b \~\~\~ c \| \| \| g \<\<\< c \~\~\~ a we specify that the sequence motif *m*can be composed in two alternative ways: The first case, labelled by *f*, requires adjacent occurrences of motifs *a*, *b*and *c*. The second case, labelled by *g*, requires adjacent occurrences of motifs *c*and *a*. When motif *m*is to be scored, *f*and *g*are seen as the scoring functions that combine the local score contribution of each case with the scores of sub-motifs *a*, *b*, and *c*. What is the computational effort of locating motif *m*in an input sequence *x*of length *n*, say at sequence positions *i*through *j?*First we assume that all motifs can have arbitrary size between 0 and *n*. The algorithm must consider all boundary positions (*i, j*) for motif *m*, which requires *O*(*n*^2^) steps at least. In case *g*, it must consider all boundary positions *k*where motifs *c*meets *a*, such that the runtime for case *g*is in *O*(*n*^3^). In case *f*, there are two such moving boundaries *k*and *l*between the three sub-motifs, so we obtain *O*(*n*^4^) overall for motif *m*. This can be improved if there is an upper bound on the size of some motif involved. If motif *a*is a single base, for example, the exponent of *n*decreases by 1 in both cases. Furthermore, if motif *b*is (say) a loop of maximal size 40, then one factor of *n*is reduced to a constant factor and overall asymptotic runtime is now *O*(*n*^2^). Sometimes a motif description can be restructured to improve efficiency by reducing the number of moving boundaries. Whether or not this is possible does not depend on the motif structure, but on the scoring scheme! This is a somewhat surprising observation from \[[@B15]\], where such optimizations are studied, and where also the line of reasoning exercised here is given a mathematical basis. In the sequel, we shall exploit another source of efficiency improvement. If the lengths of two sub-motifs are coupled somehow, say *a*and *c*have the same length, then the boundaries *k*and *l*in case *f*are related by *k*- *i*= *j*- *l*. When iterating over *k*, we can use *l*:= *j*- *k*+ *i*(rather than *k*≤ *l*≤ *j*)and save another factor of *n*. Canonization ------------ When the search space of a combinatorial problem seems to be too complex to be evaluated efficiently, heuristics are employed. Canonization restricts the search space in a well-defined way, arguing that all the relevant solutions in the full search space have a representative that is canonical, and hence, nothing relevant is overlooked. One such technique is the purging of structures that have isolated basepairs. Here the plausibility argument refers to the underlying energy model, where base pairings without stacking have little or no stabilizing effect. This canonization does not affect efficiency, but it achieves a significant reduction of the search space (figures in \[[@B16]\]), which renders the enumeration of near-optimal solutions \[[@B17]\] much more meaningful. We shall introduce three canonization rules that reduce class sr-PK to the class of *canonized simple recursive pseudoknots*, csr-PK. Using the notation introduced above, the motif definition of a simple recursive pseudoknot is given by knot = knt \<\<\< a \~\~\~ u \~\~\~ b \~\~\~ v \~\~\~ a\' \~\~\~ w \~\~\~ b\' with boundaries at sequence positions *i, e, k, g, f, l, h, j*as shown in Figure [2](#F2){ref-type="fig"}. Segment *a*forms a helix with *a*\', and *b*with *b\'*. Segments *u*, *v*, and *w*can have arbitrary structures, including pseudoknots. Naively implemented, we can expect a DP algorithm of time complexity *O*(*n*^8^) according to our efficiency estimation technique introduced above. We now apply canonization. Note that it only applies to helices forming pseudoknots; other helices are unaffected. We first present the technical aspects; the discussion of these restrictions is deferred to the next section. ### Canonization rule 1 \(a) Both strands in a helix must have the same length, i.e. \|*a*\| = \|*a*\'\| and \|*b*\| = \|*b*\'\|. (b) Both helices must not have bulges. Note that (b) is a stronger restriction and trivially implies (a). Under the regime of Rule 1 we may conclude: *f*= *l*- (*e*- *i*) *h*= *j*- (*g*- *k*) We are left with 6 out of 8 boundaries that vary independently, and runtime is down to *O*(*n*^6^). ### Canonization rule 2 The helices *a*, *a\'*and *b*, *b\'*facing each other must have maximal extent, or in other words, compartment *v*must be as short as possible under the rules of base pairing. We observe that the maximal length of *a*and *a\'*is fixed once *i*and *l*are chosen. The maximal helix length *stacklen*(*i, l*) can be precomputed and stored in an *O*(*n*^2^) table. The same observation holds with respect to the other helix, and we fix *e*= *i*+ *stacklen*(*i, l*) *g*= *k*+ *stacklen*(*k, j*). Thus, we are left with only four independently moving boundaries -- *i, k, l, j*--, and can hope to obtain an algorithm with runtime O(*n*^4^). Scores of pseudoknots found between *i*and *j*are stored in table *knot*(*i, j*), and hence the space requirements are O(*n*^2^), which is the same asymptotic space efficiency as in the folding of unknotted structures. A subtlety arises when both helices, chosen maximally, compete for the same bases of *v*, or in other words, the length of *v*would become negative. This case is addressed by ### Canonization rule 3 If two maximal helices would overlap, their boundary is fixed at an arbitrary point between them. Let *m*and *m\'*be the helix lengths so determined. We finally obtain *e*= *i*+ *m* *g*= *k*+ *m\'* The language of pseudoknots in class csr-PK can be defined by a simple context free grammar over an infinite terminal alphabet. Let *a*^*k*^denote a terminal symbol of *k*times the letter *a*. The grammar uses a single nonterminal symbol *S*and its productions are  for arbitrary *k, l*≥ 1. For example, the simple pseudoknot of Figure [1](#F1){ref-type="fig"} is represented as the string .. \[\[\[\...\...{{..\]\]\]\]\...\...\....}}. This grammar is useful to judge how different an experimentally determined structure is from class csr-PK. It is not useful for programming, since it is ambiguous and does not distinguish the fine grained level of detail required in the energy model. Canonical representatives ------------------------- A careful discussion is required to show that each simple recursive pseudoknot, if not canonical by itself, has (a) a canonical representative of (b) similar free energy. Rule 1 (b) affects the length of helices that are considered in forming the pseudoknot. Let there be a pseudoknot between *i\'*and *j\'*. It is not canonical if one of the two helices contains bulges. However, there must be at least one pair of shorter helices without bulges at *i, j*with *i\' ≤ i*and *j*≤ *j\'*, which serves as a canonical representative, albeit with somewhat higher free energy. Rule 2 is justified by the fact that the energy model strongly favours helix extension. Clearly, for each family of pseudoknots delineated by *i, k*, *l*, *j*there is a canonical one with maximal helices, whose free energy is at least as low -- except for the following case: The maximal helices compete with the internal structure of *u*, *v*and *w*. It may be possible to contrive a structure where shortening (say) helix *a -- a\'*by one base pair allows to create two pairs with new partner bases in *u*and *v*, resulting in a structure which has slightly lower energy. Still, the free energy of the canonical pseudoknot must be very similar. Finally, Rule 3 requires a decision where to draw the border between two helices facing each other and competing for the same bases. An arbitrary decision here can only slightly affect free energy, as the same base pairs are stacked either on the a -- *a\'*or the *b -- b\'*helix. Let *E*(*s*) denote the free energy computed for structure *s*. Summing up, we have shown that for each simple recursive pseudoknot *K*, there is a canonical one *C*in the search space. While we cannot prove that *E*(*C*) ≤ *E*(*K*), we have argued that this is likely, and if not, the energies will at least be close. Still, there might be another, energetically optimal canonical structure *S*(knotted or not) such that *E*(*K*) \<*E*(*S*) \<*E*(*C*). In this case, if only the \"best\" structure *S*is reported, neither *K*nor its canonical representative *C*is observed. (A remedy to this is the computation of near-optimal structures.) Finally, let us add that the implementation described below is actually slightly more general that the \"pure\" csr-PK model described above: We do allow a single nucleotide bulge in either helix of a pseudoknot, which complicates the program, but does not affect asymptotic efficiency. Evaluation of the class csr-PK ------------------------------ To evaluate how well the class csr-PK covers known pseudoknots, we considered 212 pseudoknot structures from PseudoBase. The observations are shown in Table [1](#T1){ref-type="table"}. We find 172 simple recursive pseudoknots, and 40 of more general shapes. We find that 135 out of the 172 pseudoknots lie in csr-PK, i.e. they are their own canonical representatives. 11 more fall into the relaxed csr-PK, where we allow a single nucleotide bulge in Canonization Rule 1. Thus, we cover 146 out of 212 (68%). 26 simple recursive pseudoknot do not fall in class csr-PK, since they contain isolated basepairs, non canonical basepairs or one of the helices has not maximal extent. Considering the remaining 20% complex pseudoknots, note that often pseudoknots in more general classes also have a good representative in csr-PK. For example, the pseudoknot of Hepatitis delta virus (Figure [3](#F3){ref-type="fig"}) consists of four interacting helices of shape *a -- b -- c -- d -- c\' -- a\' -- d\' -- b*\', where helix *d -- d\'*is very short -- only two base pairs. Deleting it, helix *c -- c\'*is no longer interacting with other helices, and the pseudoknot falls within class csr-PK. Better than optimal ------------------- There are many reasons why \"the\" MFE structure may only be part of what we want to know about a molecule\'s foldings. To deal with the problem when the optimal (knotted) structure is non-canonical, and its canonical representative is dominated by an unrelated structure, we provide two means: First of all, our algorithm is non-ambiguous, the prerequisite for a non-redundant enumeration of near-optimal structures \[[@B16]\]. We can let the program to report the *k*best structures. Secondly, we shall provide three variants of our program: *pknotsRG-mfe*computes the mfe structure (or the k best), pseudoknotted or not. *pknotsRG-enf*picks out from the folding space the energetically best structure that contains at least one pseudoknot. *pknotsRG-loc*computes the energetically best pseudoknot that can be formed locally, i. e. somewhere in the sequence. \"Best\" is defined here as minimal free energy per base, to avoid a built-in bias towards large pseudoknots. The best local pseudoknot motif is included by adding two cases: bestPK = skipleft \<\<\< base \~\~\~ bestPK \|\|\| bestPKl bestPKl = skipright\<\<\< bestPKl \~\~\~ base \|\|\| knot These clauses have time complexity *O*(*n*^2^) and preserve the non-ambiguity of the algorithm. If desired, an enumeration of near-optimal \"local\" pseudoknots is also feasible. Predictive accuracy ------------------- We first consider the predictive accuracy achieved by our approach. We have already evaluated the class csr-PK against the known pseudoknots, and we know that our algorithm correctly implements this class in its search space. What is really tested in the following is the adequacy of the current thermodynamic model (which our algorithm shares with *RNAfold*and in an older version with *pknotsRE*), and the results in this section may improve if this model is further improved in the future. We test our algorithm on the set of sequences listed in Table [2](#T2){ref-type="table"}, including 212 sequences from PseudoBase. Although there is some redundancy on the sequence level, there is a good reason why we found it important to use all available sequences for testing: Even near identical sequences can have different MFE structures, or a small change may prevent successful pseudoknot prediction. In contrast to \[[@B13]\] we did not restrict the evaluation to the class of pseudoknots recognized by our program. It is also instructive to retain the difficult cases, and see whether the predictions catch at least some aspect of a more general pseudoknot. We compare our results to the output of *RNAfold*, as a representative for RNA folding tools without pseudoknot folding capability, and to *pknotsRE*where computationally feasible. For each predicted structure we count the number of correctly and falsely predicted base pairs (TP and FP). Let BP be the number of basepairs in the reference structure from the database or literature. We define the sensitivity as (TP/BP), selectivity as (TP/TP+FP). In Table [3](#T3){ref-type="table"} we list the prediction accuracy for our sequence set. For all sequences we enhance the prediction accuracy with respect to *RNAfold*. Both, the sensitivity and the selectivity increase. Compared to *pknotsRE*our results are slightly better, probably because we are using the newer and subtler energy model. For example, for the sequence of hepatitis delta virus, our algorithm predicts all helices except for the very short helix 5 (see Figure [3](#F3){ref-type="fig"}), while the other programs miss more than 50% of the basepairs. We also folded 14 randomly selected human tRNAs (third line in Table [3](#T3){ref-type="table"}) and found only one false positive pseudoknot. Interestingly, the pseudoknotted structure has two helices (9 bp) in common with the true clover-leaf structure, while the structure computed by *RNAfold*has only one helix (4 bp). For all programs the overall prediction accuracy for tRNAs is not very high. tRNAs are a known hard case for structure prediction because they contain many modified bases. Since we use the same energy model as *RNAfold*and our algorithm does not introduce spurious pseudoknots, predictions of *RNAfold*and *pknotsRG*for unknotted structures are identical. Of course, if there is more than one optimal structure, each of the optimal alternatives may be reported and thus the same folding can not be guaranteed. Computational performance ------------------------- Clearly, we are able to fold sequences that are longer than *pknotsRE\'s*limit of 150 nucleotides. Short sequences up to 100 nucleotides are folded within a minute. Long sequences (400 bp) take about 2 hours (see Table [4](#T4){ref-type="table"}). If we restrict the maximal pseudoknot size to a reasonable constant, say 150 nucleotides, we can further increase the running time. The algorithm runs now in *O*(*cn*^3^) with a rather large constant *c*. This enables us to fold sequence of length 1000 in approximately 12 hours. We can further observe that the space requirements scale quadratically with the input size, as expected. For a fair comparison, the reader should keep in mind that the extra time spent by *pknotsRE*is not strictly wasted: It is spent on assuring that the optimal folding of the input RNA sequence does not contain pseudoknots with chained interacting helices of lower free energy than the reported structure. *pknotsRG*does not consider such structures and hence cannot make this assertion. Discussion ========== In the following, we discuss extensions of the implemented model and their expected computational cost Bulges, triple crossing and kissing hairpins -------------------------------------------- Canonization Rule 1 can be relaxed further to allow larger bulges inside the helices forming a pseudoknot. As long as their number (and hence the length difference of the two arms of a helix) is bounded by a constant, asymptotic efficiency is not affected. Two examples of non-simple pseudoknots are shown in Figure [4](#F4){ref-type="fig"}. We can incorporate them into our algorithm adding the definitions kiss = kss \<\<\< a\~\~\~u\~\~\~b\~\~\~v\~\~\~a\'\~\~\~w\~\~\~c\~\~\~x\~\~\~b\'\~\~\~y\~\~\~c\' triple = trp \<\<\< a\~\~\~u\~\~\~b\~\~\~v\~\~\~c\~\~\~w\~\~\~a\'\~\~\~x\~\~\~b\'\~\~\~y\~\~\~c\' Canonization can be applied as above, with Rule 3 becoming more sophisticated for the triple interaction case. This would yield an algorithm of runtime *O*(*n*^6^), bringing runtime back to the efficiency class of the Rivas/Eddy algorithm. But note that the space requirements remain *O*(*n*^2^). This is due to the fact that we now consider three interacting helices, but not arbitrary chains. Folding long sequences ---------------------- RNA folding *in vivo*as *in vitro*must be understood as a hierarchical process, where small structures in close vicinity form first, and then combine to larger ones \[[@B18]\]. The folding path becomes relevant, and the longer a sequence, the more unlikely it is that its folding path leads to a global energy minimum. In other words, the longer the sequence, the less reliable are the results of minimum free energy folding. *pknotsRG*gives us the possibility to test this using a fairly large structure containing pseudoknots that have been proved experimentally. We considered the sequence of the group I intron from Tetrahymena thermophila (419 NT) (V01416). The MFE-structure found was quite different from the \"true\" structure taken from the literature. We hand-coded the experimental structure and evaluated its stability in our energy model. The result was striking: the experimental structure (-132.26 kcal/mole) was significantly far from the possible minimum of free energy (-155.64 kcal/mole). So far in fact that it seems infeasible to detect the structure by scanning the space of near-optimal structures. This could be interpreted as the energy model being incorrect, but since it works well for short sequences, we suggest that this is an indication that the kinetics of folding already have a strong influence with this size of sequence, at least when pseudoknots are involved. While we have achieved a considerable speedup for predicting small pseudoknotted structures, it seems that minimum free energy approach is not meaningful with the largest structures which it now can handle algorithmically. However, the situation changes when we are looking for particular structural motifs (see below). Conclusion ========== We presented an algorithm *pknotsRG-mfe*, based on the MFE-model, for finding the best RNA structure including the pseudoknot class csr-PK. This requires *O*(*n*^4^) time and *O*(*n*^2^) space. The algorithm variant *pknotsRG-enf*returns the energetically best structure that contains a pseudoknot (interesting when the MFE structure is unknotted), while *pknotsRG-loc*reports the best pseudoknot (under a length-normalized energy score) *somewhere*in a sequence. We achieve a high prediction accuracy for moderate length sequences, whereas long sequences, at least when pseudoknots are involved, seem to have a folding scheme that cannot be modelled with minimum free energy folding. Algorithm *pknotsRG*is based on a simpler grammar model than the crossed interaction grammars \[[@B19]\] underlying *pknotsRE*, as well as the communicating grammars underlying the recent approach by Cai \[[@B20]\]. It requires only a minor extension over the ADP tree grammars that are applicable to a wide range of sequence analysis problems \[[@B21]\]. Furthermore, the grammar is not only a theoretical backup, explaining the underlying model. With minor annotation for the sake of efficiency, the grammar actually constitutes executable code. This means that *pknotsRG*can serve as a template for a new class of programs we call thermodynamic matchers. Many functionally important RNAs like RNase P or group-I-introns have known structures that include pseudoknots. The search for such motifs using combinatorial matchers like RNAmotif \[[@B22]\] is hampered by the problem that a motif description is either too specific and misses relevant instances, or else it is too vague and produces a large number of different matches to the same sequence. We suggest to develop *thermodynamic matchers*, which are RNA folding programs, based on the established MFE model, but specialized to the particular structural motif at hand. Such a matcher returns the optimal way to fold a sequence into the motif structure, together with the free energy of this folding. Comparing this energy to the MFE of an unrestricted folding can give us a hint with respect to the significance of such a match. Methods ======= Choice of implementation method ------------------------------- Using the ideas presented so far, our folding algorithm can be implemented in any language suitable for dynamic programming, say FORTRAN or C. However, we are interested in a reusable implementation that can be integrated without change in specialized folding programs called thermodynamic matchers. Therefore *pknotsRG*was implemented using the method of algebraic dynamic programming (ADP) \[[@B15],[@B23]\]. RNA folding in ADP ------------------ In ADP, the search space of a DP problem is defined on a declarative level, specified by clauses like the ones we have already seen above. Together they form a tree grammar, defining a tree language whose elements are all the candidates in the search space. In our case, the candidates are RNA structures represented as trees. The typical DP recurrences are implicit in this description. Scoring is achieved by interpreting the operators (e.g., *knt, skipleft, skipright*) that build the trees as scoring functions. The grammar needs to be annotated with respect to tabulation and the application of the objective function (in our case, minimization). The advantage of this method is its high level of abstraction. No subscripts, no errors. The perfect separation of search space definition and evaluation allows the same grammar to be used for different kinds of analyses, e. g. folding space statistics. Relevant algorithmic properties such as non-ambiguity and efficiency can be studied on this level of abstraction. Last not least, an ADP program can be executed as is, avoiding the explicit formulation of DP recurrences (and a whole universe of programming errors). A significant, but constant factor of speedup can be gained by explicitly formulating the recurrences and implementing them in a lower level language. Automating this process is part of our current work. We start from an ADP algorithm for folding RNA secondary structures (excluding pseudoknots) provided by Dirk Evers \[[@B24]\]. We show ADP clauses defining the closed substructures: stacks, hairpins, bulges, and multiloops, adding an alternative for pseudoknots. region denotes an arbitrary sequence of (unpaired) bases.  The shown code abstracts from efficiency annotation and the treatment of dangling bases. The complete algorithm is found on the ADP WWW pages \[[@B25]\]. It is based on the standard MFE model with dangling bases, is non-ambiguous and requires *O*(*n*^3^) time and *O*(*n*^2^) space. A size constraint of 30 is used to bound loop length in internal loops. Closed substructures are defined such as to avoid lonely base pairs. While all this is easily expressed within the standard ADP framework, our new algorithm requires extensions which are now explained. Adding pseudoknots ------------------ The implementation strictly follows the outline given in the methods section, except that a considerable amount of detail related to the energy model has to be taken care of. While ADP bans the use of subscripts, our canonization ideas require to explicitly manipulate subscripts. We show the concrete pseudoknot code, but explain only the essential points. A subscript pair (*i, j*) denotes input sequence positions *inp*~*i*+1~.. .*inp*~*j*~. \[\...\] denotes lists, and \<- denotes enumerating a list of alternative values. knot (i, j) = \[pk energy a u b v a\' w b\' \| k \<-\[i+2 .. j-1\], l\<-\[k+1 .. j-2\], These line chooses *k*and *l*from the interval (*i, j*), and put together the results from *a*, *u*, *b*, *v, a\'*, *w, b\'*under the scoring function *pk*. Each helix must have a minimum length of two bases. Due to stereochemical reasons one base in the front part and two bases in the back part are left explicitly unpaired; these bases should bridge the stacks. This consideration is taken over from *pknotsRE*. The next definitions implement canonization rules 1, 2 and 3. They determine the helix lengths, finally computed into the variables *m*and *m\'*. If either of them is smaller than 2, a pseudoknot is not possible at this particular location.  The function *cut*shortens the helix *b -- b\'*as much as necessary in case of overlapping helices. The next lines define the pseudoknot components *a*through *b*\', plus the local energy contribution. To avoid an extra factor of *n*in time complexity, the energies of maximal length helices are also precomputed in table *stackenergy*. If the helix *b -- b\'*must be chosen shorter than maximal to avoid overlap, a correction term has to be subtracted. This explains the negative term in the energy computation.  Left to be defined are the interior structures front, middle, and back. For reasons of space, we only show the definition of *front*. For a full implementation of the algorithm see [additional file 1](#S1){ref-type="supplementary-material"}.  This case takes care of a potentially dangling base from the *b*-helix, and if the remaining region is not empty, an arbitrary list of substructures (*comps*) is recognized. *idd, frd*and *pul*are the corresponding functions from the energy model. Overall, the energy of a pseudoknot consists of stabilizing and destabilizing terms. Where possible, we use the values from the current thermodynamic energy model \[[@B6]\]. As stabilizing terms we count the nearest neighbour stacking energies of the pseudoknot helices and contributions of dangling bases at both ends of each helix. If the length of the middle part *v*is smaller or equal to 1, the pseudoknot helices stack coaxially on each other and we further add the appropriate stacking energy. In \[[@B11]\] a pseudoknot initiation parameter of 7 kcal/mole is proposed. However, we found out, that setting this value to 9 kcal/mole performs better with the new energy model. Our observation supports the similar choice made by Dirks and Pierce \[[@B26]\]. Finally, we penalize each unpaired nucleotide inside a pseudoknot loop with 0.3 kcal/mole. This seems to be the best approximation of the values given in \[[@B27]\]. Of course, if the pseudoknot is recursive the energy of the subcomponent is taken into account as well. The first clause (knot) chooses *k, l*inside (*i, j*), computes *m*and *m\'*using the precomputed maximal helix information, and passes these boundaries to the pseudoknot compartments. Methodically, this is a use of inherited attributes with the underlying tree grammar, and appears to be a novel technique in dynamic programming, at least in its grammar oriented tradition \[[@B19],[@B28]-[@B30]\]. The relative effort of implementing the three variants of *pknotsRG*can be judged from the sizes of the tree grammars required, which are summarized in Table [5](#T5){ref-type="table"}. Availability ------------ The three variants of the algorithm *pknotsRG-mfe, pknotsRG-enf*, and *pknotsRG-loc*are available as executables and source code on the Bielefeld Bioinformatics Server \[[@B31]\]. Authors\' contribution ====================== RG had the initial idea for the algorithm. JR developed and evaluated the software. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 Source code of pknotsRG-mfe ::: ::: {.caption} ###### Click here for file ::: Acknowledgement =============== We gratefully acknowledge the help of Dirk Evers, whose ADP code for folding unknotted structures served as a starting point for our implementation. Marc Rehmsmeier helped with some delicate algorithmic aspects. Peter Steffen acted as a semi-automated compiler of ADP code into C. We also thank Elena Rivas for discussing effects of canonization at an early stage of this work. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **A simple pseudoknot.**A simple pseudoknot, formed by helices *a -- a\'*and *b -- b\'*, with intervening sequences *u, v, w*. If the internal parts *u, v, w*contain further secondary structures, we call it a simple recursive pseudoknot. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **The boundaries of a pseudoknot.**Eight moving boundaries delineating a simple recursive pseudoknot. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **The pseudoknot of Hepatitis Delta Virus.**On the left side the structure proposed by \[32\], on the right the MFE-structure predicted by *pknotsRG*. Our algorithm misses only the short helix 5 and adds 3 basepairs to helix 4. Image created with RnaViz2 \[33\]. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Kissing hairpins and triple helix interaction.**Two examples of more complex pseudoknots with three helices: Kissing hairpins (left) and triple helix interaction (right). :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Class membership of 212 pseudoknots from PseudoBase. The sequences were determined by comparative sequence analysis and/or by experimental techniques. The largest class of pseudoknots is simple recursive or even canonical simple recursive. ::: simple recursive pseudoknots ------------------------------ -------------- ----------------- ------------------- -------------- ------- csr-PK 1-nt bulge Rule 2 violated isolated basepair G-A basepair total 135 11 6 17 3 172 complex pseudoknots internal loop triple helix four helices kissing hairpins large bulge total 23 12 1 3 1 40 ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Sequences used for testing ::: Sequence Length BP Reference ------------ ---------- ---------- ---------------- PseudoBase variable variable \[14\] 7 HIVRT 35 11 \[34\] 14 tRNAs 71--82 18--19 HDV 87 32 \[32\] ag-HDV 91 25 \[32\] TYMV 86 24 \[35\] TMV (up) 85 25 \[36\] STNV 252 69 GenBank:M64479 ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Evaluation of predictive accuracy *TP*= true positives, sens. = sensitivity, *FP*= false positives, sel. = selectivity, *K*= (number of correct predicted pseudoknot helices)/ (expected number of pseudoknot helices). False negatives can be derived from: *FN*= *BP*-- *TP*. For HIVRT, PseudoBase and tRNA the average over all sequences is taken. ::: RNAfold pknotsRE pknotsRG-mfe ------------ ------ ----------- ------------ -------------- ----------- ----------- ----- ------------ ------------ ----- PseudoBase 13.1 7.1(54.2) 4.4(61.9) \- 9.2(72.3) 3.8(70.7) \- 10.2(77.3) 3.5(74.7) \- HIVRT 11 4.7(42.8) 1.7(73.3) 1/2 11(100) 0(100) 2/2 11(100) 0(100) 2/2 tRNAs 18.7 9.4(50.4) 12.5(43.0) 0/0 \- \- \- 9.8(52.3) 12.2(44.5) 0/0 HDV 32 12(37.5) 16(42.9) 2/4 14(43.8) 16(46.7) 2/4 29(90.6) 2(93.5) 3/4 ag-HDV 25 4(16.0) 24(14.3) 1/2 24(96.0) 9(72.7) 2/2 21(84.0) 11(65.6) 2/2 TYMV 24 17(70.8) 6(73.9) 1/2 24(100) 1(96.0) 2/2 23(95.8) 2(92.0) 2/2 TMV 25 13(52.0) 8(61.9) 3/6 13(52.0) 9(59.1) 3/6 20(80.0) 4(83.3) 6/6 STNV 69 26(37.7) 54(32.5) 2/8 \- \- \- 42(60.9) 37(53.2) 5/8 ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Performance results Performance results for random RNA sequences and comparison to *pknotsRE*. The performance of *pknotsRG-loc*is similar to *pknotsRG-mfe*and therefore not shown. The bounded version restricts pseudoknots to a maximum length of 150 nucleotides; measured on an UltraSPARC CPU with 450 MHz and 4 GB main memory. Time is measured in (h:m:s), memory in MB. ::: Length pknotsRE pknotsRG-mfe pknotsRG-enf pknotsRG-enf bounded -------- ---------- -------------- -------------- ---------------------- --------- ----- --------- ----- 40 17.4 \- 0.7 1 0.9 2 0.9 2 80 21:11 38 9.5 5 8.7 5 8.8 5 100 1:23:50 80 20.0 8 32.5 10 34.5 9 200 \- \- 6:46 36 8:03 42 6:33.6 42 300 \- \- 33:07 80 44:58 102 20:48.4 90 400 \- \- 1:48:08 146 2:28:18 184 45:28.9 155 ::: ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Program sizes ::: Program variant Nonterminals Productions Tables ----------------- -------------- ------------- -------- pknotsRG-mfe 27 74 8 pknotsRG-loc 27 68 8 pknotsRG-enf 40 119 11 :::

PubMed Central

2024-06-05T03:55:47.176774

2004-8-4

PMC514698

Background ========== While similarity (homology) of DNA sequence between organisms can be used to propose *potential*gene functions, transcriptional regulation, and protein pathways (e.g., \[[@B1]\]), there are often major differences in the protein products, functions, and pathway involvement of genes with nearly identical sequences \[[@B2]\]. Consequently, sequence homology may be viewed as a means of generating an initial \"draft\" hypothesis for the gene network of a newly sequenced organism that can be built upon using high throughput experimental techniques such as DNA chips and microarrays for mRNA transcript profiling \[[@B3]\], protein abundance profiling with mass spectroscopy and 2-D gel electrophoresis \[[@B4]\], and protein-protein and protein-DNA binding assayed using SELDI mass spectrometry \[[@B5]\] and protein chips \[[@B6]\]. In addition, new genetic technologies, in particular small interfering RNA (siRNA) for selective gene suppression facilitate high-throughput massively parallel perturbation of the gene and protein networks of biological systems \[[@B7]\]. Given the potential scale and complexity of experiments and resulting data sets, biologists need a modeling and simulation framework to optimally design experiments and interpret results. The problem is not simply one of \"reverse engineering\" to find the optimal \"best fit\" gene, protein, and/or metabolite interaction model to explain a set of experimental results; rather, modeling should suggest the range of hypotheses that can potentially explain the results of one experiment and select the optimal next experiment to reduce the number of possible alternative hypotheses, with the goal of converging to a biological system model that can be used to predict the effect of molecular perturbations. A major challenge of modeling biological systems is that conventional methods based on physical and chemical principles require data that is difficult to accurately and consistently obtain using either conventional biochemical or high throughput technologies, which typically yield noisy, semi-quantitative data (often in terms of a ratio rather than a physical quantity) \[[@B3]\]. In particular, microarray gene expression ratios are ultimately obtained from pixel counts of relatively messy images. Boolean networks (e.g. \[[@B8]\]) are computationally simple and do not depend on precise experimental data, and thus they are suitable for handling both the complexity of biological networks and the challenge of generating and comparing multiple hypothetical networks as described in the above scheme. However, Boolean models have inadequate dynamic resolution to accurately describe the behavior of a biological network \[[@B9]\]. In contrast, differential equation models (e.g., \[[@B10]\]) can be computationally expensive and sensitive to imprecisely measured parameters. Even the lower throughput RT-PCR method for gene expression measurement (as described in \[[@B10]\]) cannot produce quantitatively precise data that can be accurately mapped to actual mRNA concentrations in the sample. Because of computational limitations, continuous modeling approaches (e.g., \[[@B10],[@B11]\]) are limited to finding the single model that best fits experimental data given some set of constraints, such as a maximally sparse gene interaction network \[[@B11]\]. Fuzzy logic \[[@B12]\] provides a mathematical framework that is compatible with poorly quantitative yet qualitatively significant data. Fuzzy logic is a natural language for linguistic modeling, thus it is consistent with the qualitative linguistic-graphical methods conventionally used to describe biological systems. Fuzzy models are rule-based; accordingly, there is a potential scalability problem as the number of antecedents (\"inputs\" to the rule) and variable states (\"resolution\" of inputs and rule outputs) increase, causing combinatorial explosion. Non-scalable conventional fuzzy logic has previously been used to analyze microarray data \[[@B13]\]. However, because of the nonlinear scalability of the modeling method and resulting computational expense of generating rules for multiple inputs, this method allows for only one possible positive and one possible negative regulator for each gene, thus yielding few biologically meaningful insights and experimentally testable hypotheses. The problem of rule set combinatorial explosion is addressed by the union rule configuration (URC) developed by Combs and Andrew \[[@B14]\], which allows for linear growth in rule set complexity with both resolution (number of states) and number of inputs (rule antecedents) at the cost of having to represent nonlinear relationships between inputs as hidden layers \[[@B15]\]. In the linear (URC) fuzzy logic scheme, there are distinct fuzzy rules for each individual input to a given output. For example, given input variables A and B to an output C, there would be a set of rules relating A to C (e.g., \"If A is LOW then C is LOW\", \"If A is HIGH then C is HIGH\") and another set of rules relating B to C (e.g. \"If B is LOW then C is HIGH\", \"If B is HIGH then C is LOW\"). After each rule is applied individually, the intermediate evaluations of the fuzzy state of the output variable (\"node\") are aggregated by a fuzzy union (logical OR) operation (e.g. by summing or taking the resulting memberships in the fuzzy sets defining the state of the output). This contrasts with conventional fuzzy logic (or the \"intersection rule configuration\"), which has rules relating all combinations of inputs evaluated by a fuzzy intersection (logical AND). For the example with inputs A and B to output C, rules would read as, e.g., \"If A is LOW and B is LOW then C is LOW\", \"If A is LOW and B is HIGH then C is HIGH\", etc., leading to a combinatorial explosion avoided by the URC. The utility of linear (URC) fuzzy logic has been demonstrated in its ability to qualitatively model the *lac*operon of *E. coli*\[[@B16],[@B17]\]. In our previous work, a URC fuzzy logic model was constructed from existing qualitative biological knowledge about the interaction of genes and limited quantitative data on protein and metabolite concentration and enzyme kinetics, showing the power of linear fuzzy logic to describe complex multi-component regulation. Here, the linear fuzzy logic method is extended to tackle the inverse problem of gene network reconstruction from real quantitative microarray data where there are many inputs. This involves both methods for mapping the experimental data to the fuzzy logic membership functions and a useful implementation of the URC fuzzy logic to represent the gene networks. In addition, a robust algorithm for performing searches through the exponentially large space of possible gene networks is presented. To address the problem of generating *all*plausible hypothetical network models that explain an experimental data set, we are initially proceeding with an exhaustive search of possible gene interactions to find those that fit the data within some error threshold. Thus, the problem we are tackling is of exponential complexity with O(*m*^*N*^) growth in the number of possible rules for the behavior of a given \"output\" gene of a gene interaction node, where *N*is the number of (input) genes that can possibly control it and *m*is the number of possible rules describing the effect of each single input gene on the output gene. On the other hand, if a linear fuzzy logic scheme is not used, the problem would grow at an unacceptably high O(*m*^*N\^N*^) rate. The number of possible rules for each gene-gene interaction (*m*) is given by *n*^*n*^, where *n*is number of fuzzy sets that describe the state of a variable. Hence, we will constrain the size of the problem by (i) setting the minimum number of fuzzy sets to three, the minimum for meaningful resolution, (ii) limiting the number of possible input genes that are allowed to control the output of the output gene at each node of the fuzzy network model, and (iii) not allowing nonlinear gene interactions which would require hidden layers. The last condition is not particularly severe, as a typical nonlinear interaction (e.g., \"xor\") interaction between two regulatory proteins at a gene is mediated by an intermediate complex between the proteins that can be represented as an independent node in a network model. Therefore, \"hidden layers\" may generally be avoided by including more biological detail as explicit nodes in the model: for example, explicitly including the temporary interaction between proteins within a scaffolded cellular signal transduction complex, or by incorporating as model nodes the various topological states of a region of DNA influencing transcription factor binding, or in general, adding sufficient biological detail such that interactions between inputs can be linearly modeled. We apply partially scalable, linear fuzzy network modeling to a data set commonly used for demonstrating computational methods in systems biology, microarray experiments of yeast cell cycle gene expression \[[@B18]\]. These data were obtained in 1998, prior to subsequent technical and statistical advances to improve data quality. However, to keep our case study as general as possible and demonstrate the ability of the fuzzy logic approach to handle other similarly noisy data sets, we do not do any data processing other than the fuzzy modeling process (described in the Methods). Exhaustive search is used as a brute force \"reverse engineering\" method to find all possible gene network models that fit the data for a set of twelve genes known to participate in the yeast cell cycle. We show that the search converges to a small number of models describing the expression of each gene within a fit tolerance. Models found from the data for one particular yeast cell cycle time series are also capable of qualitatively predicting data from another time series experiment (i.e., one using a different cell synchronization method). In addition, given the constraints of the search algorithm (described in more detail in the Methods) and our limitation to pure transcriptional data, we find that the best fitting fuzzy network models collectively recover some direct and indirect functional relationships between genes predicted by interactions found by previous biochemical experiments as well as quantitative and statistical methods based on transcriptional correlations. Results ======= Yeast cell cycle data set ------------------------- As a proof of concept, we have used exhaustive search to generate fuzzy gene networks based on yeast (*Saccharomyces cerevisiae*) cell cycle microarray time series data sets presented in \[[@B18]\] (which included data from \[[@B19]\]). Researchers frequently use these data sets to demonstrate and validate statistical and clustering analysis (e.g., \[[@B20],[@B21]\]), mathematical modeling \[[@B22],[@B23]\], and reverse engineering methods \[[@B21],[@B24]\]. Biological details of the yeast cell cycle transcriptional network and some computational methods for its analysis are reviewed in \[[@B25]\]. *S. cerevisiae*cell cycle regulatory protein-DNA interactions were also the subject of a recent extensive experimental study \[[@B26]\] and there is a large amount of previously obtained biological knowledge on the interaction of yeast cell cycle proteins, i.e., information contained in the Yeast Proteome Database \[[@B27]\] and the KEGG pathway database \[[@B28]\]. Consequently, predicted transcriptional network models we derive for the Spellman *et al*. \[[@B18]\] data set can be tested against numerous independent data sets and compared with models obtained using other methods. We focus on the 12 key yeast cell cycle genes listed in Table [1](#T1){ref-type="table"} with descriptions taken from the Yeast Proteome Database. The protein products of these genes have been extensively studied using conventional biological techniques and are known to regulate each other and play key roles in controlling cell cycle. Consequently, observed correlations between the genes of Table [1](#T1){ref-type="table"} in cell cycle microarray data are most likely the result of real biological activity rather than noise. In addition, cell cycle gene subsets similar to this one have been the subject of other recent gene network modeling and reverse engineering publications (e.g., \[[@B21],[@B24]\]). Figure [1](#F1){ref-type="fig"} shows the current understanding of the interactions of the cell cycle protein subset. There are three sets of gene expression time series in \[[@B18]\] measured for cells synchronized by different methods, called the *cdc15*, *alpha*, and *cdc28*sets. We fit models on the basis of the *cdc15*data set since it contains the least number of missing data. Time points in the *cdc15*set for which there is missing data for one or more of the 12 genes are excluded from the rule search. We perform an exhaustive search with a maximum of 4 inputs per node, as detailed in the Methods. A Microsoft Excel workbook with the complete fitting data set is provided in [Additional File 1](#S1){ref-type="supplementary-material"}, including all the fuzzy rule models for each gene obtained from exhaustive search with an *E*~MIN~threshold of approximately 0.6. Results of fitting to data -------------------------- Figure [2](#F2){ref-type="fig"} shows the number of rule models found in the exhaustive search that fit the expression time series of the *CLN1*gene (using the *cdc15*data set) at different error tolerance levels (*E*~MIN~, as defined by Equation 4 in the Methods). It shows typical behavior for the exhaustive rule search. The number of fuzzy models that fit a gene expression time series decreases exponentially as the fit tolerance (*E*~MIN~) increases, up to a maximum tolerance above which no models fit the data. A successful search generally converges to a small number of distinct models at the maximum fit tolerance, representing \"plausible\" hypothetical transcriptional networks that can explain the available data. In some cases, though it did not occur for any of the genes analyzed here, the search fails and there are a large number of models with similar poor fit scores and no suitable subset of \"plausible\" models. The plausible model subset generally contains common rule patterns. For example, Table [2](#T2){ref-type="table"} lists the models for *CLB5*expression with the highest fit scores found in the exhaustive search. The rules are in the format used for the example described in the Methods section. Table [3](#T3){ref-type="table"} shows three models for each gene in the network: the best fitting rule (highest score) and the two highest scoring rules with different combinations of input genes. The scores for each of the three models are provided in corresponding rows at the bottom of the table. Figure [3](#F3){ref-type="fig"} shows the best fitting interaction network diagrams for each node gene from Table [3](#T3){ref-type="table"}. To test whether the linear fuzzy gene network models found for one set of experimental data (i.e., *cdc15*synchronization time series) can accurately predict another set of results for the same system, we analyzed the microarray time series for *alpha*cell synchronization presented in \[[@B18]\]. There are some missing values for some genes at some time points in the *alpha*data set, which are set to zero and could potentially lead to discrepancies between the modeling and experimental data only at those points. Figure [4](#F4){ref-type="fig"} shows the predicted time series for the expression ratio of four genes (*CLN1*, *CDC28*, *SWI6*, and *CLB5*) given the highest (except for *CLN1*, second-highest) scoring models in Table [3](#T3){ref-type="table"}. (The second-best fitting model is used for *CLN1*because it consists of four inputs, including all three inputs in the highest-scoring model along with another gene. Thus, it represents a more general \"consensus\" model for *CLN1*.) These models fit the original *cdc15*training data with very different calculated tolerances (as measured by the fit error *E*) ranging from 0.510 (*CDC28*, Figure [4B](#F4){ref-type="fig"}) to 0.930 (*CLN1*, Figure [4A](#F4){ref-type="fig"}). Discussion ========== Using exhaustive search, we found linear fuzzy networks that predict *cdc15*cell cycle microarray data for the expression of most of the twelve yeast genes we analyzed. The rule search typically converged to a small set of \"plausible\" models at a given fit error (*E*) tolerance for each gene (with exponential convergence as shown in Figure [2](#F2){ref-type="fig"}). Even for genes for which no highly fitting model could be found, such as *SWI4*, the best model (fitting at *E*= 0.620) predicts the qualitative behavior of independently measured *alpha*time series data (Figure [4C](#F4){ref-type="fig"}). Moreover, models that are more predictive (*E*\> 0.8) of the *cdc15*training data provide quantitatively accurate predictions of the *alpha*data (Figures [4A](#F4){ref-type="fig"} and [4D](#F4){ref-type="fig"}). Notably, these consistently good fits for the *alpha*data set were achieved using exactly the same arctangent data normalization and fuzzification scheme applied to the *cdc15*data set. This suggests that the fuzzy processing methods described here can be generally applied for data sets obtained from different microarray experiments, provided a roughly symmetric distribution of Log2 ratios about 0, such that sets **1**and **3**both remain meaningful -- though the ratios could be re-centered if necessary. In general, our results demonstrate the ability of qualitative fuzzy rule models to interpret the results of quantitative data and make predictions that can be statistically analyzed. Consequently, these models can be used to pose experimentally testable hypotheses. Measurements of mRNA expression from microarray experiments complement information from additional gene knockout, DNA-protein and protein-protein experiments. A model based on pure transcriptional data will thus necessarily contain indirect relationships between proteins and miss other direct purely protein-protein interactions. However, gene network models can suggest functional roles and relationships for genes and proteins, and these models are necessary in complex system analysis to design and interpret further experiments that will specifically determine protein function and identify actual chemical interactions. To see what biological insights can be derived from fuzzy gene network models, we can examine areas of agreement and discrepancies between the best-fitting models found in our exhaustive search (shown in Table [3](#T3){ref-type="table"} and Figure [3](#F3){ref-type="fig"}) and the current understanding of the yeast cell cycle network (summarized in Figure [1](#F1){ref-type="fig"}). Focusing on *CLN1*, we found positive regulation by *CLN2*and negative regulation by *CDC20*(Figure [3](#F3){ref-type="fig"}), which are correlations expected from biological knowledge (as shown in Figure [1](#F1){ref-type="fig"}) and found by Soinov, *et al*. using a supervised learning method \[[@B21]\]. In addition, the model for *CLN1*includes a direct connection with *CDC28*and an indirect connection with *MBP1*(through regulation of the SBF complex) that are consistent with their relative positions in the cell cycle (Figure [1](#F1){ref-type="fig"}). The best-fit model for *CLN1*depended solely on a positive interaction with *CLN2*, revealing the strong co-transcriptional connection between *CLN1*and *CLN2*. The connection between *CLB5*and *CLB6*was also found in the model for *CLB5*. Other successfully found interactions include the negative regulation of *CDC6*by *CLB6*and the positive regulation of *CLB6*by *MBP1*. Some biologically accurate relationships were found that were absent from the supervised learning analysis of \[[@B21]\]. Notably, the model successfully recovers the apparent inhibition of *CLB5*by *CDC20*, which is not shown in Figure [1](#F1){ref-type="fig"} (based on the KEGG pathway) but arises from cdc20 protein presenting clb5 protein to proteases for degradation (as included in the model of \[[@B22]\], references within). There are several biological relationships that are not found in the best-fitting networks of Figure [3](#F3){ref-type="fig"}, such as an interaction between *SWI6*and *SWI4*(which form a multiprotein complex). The best-fitting models for *SWI4*include a repressing action by *MBP1*, which is inconsistent with biological knowledge (Figure [1](#F1){ref-type="fig"}) suggesting that *MBP1*and *SWI4*activity should correlate (since they act at the same stage in the cell cycle). However, closer examination of the Spellman data set reveals that the amplitude of *MBP1* transcription varied within a small range, and the measurement could have been very noisy, resulting in a potential error by the algorithm. (It should be noted that no correlation is identified between *MBP1*and *SWI4*by the supervised learning algorithm in \[[@B21]\].) In general, determining which relationships found in the fuzzy gene network represent biologically accurate interactions is a question that must be resolved by analyzing other data sets or from new experiments. The multiple plausible hypothetical input gene combinations can be used to optimally design experiments to add most information for least effort (time and cost) to revise fit errors and produce a new, more realistic set of hypothetical networks. Conclusions =========== In this work, we describe partially scalable, linear fuzzy logic models for biological network modeling. We demonstrate our approach by developing network models that accurately predict transcriptional data from typically noisy and semi-quantitative microarray experiments. Looking at the transcription network alone provides us with a view of the system at the \"gene interactions\" level. As measurement technology rapidly advances, the methods we describe can be extended to comprehensive heterogeneous data sets. To address the problem of analyzing the complex results of an exhaustive fuzzy model search and designing optimal experiments, we are currently developing pattern recognition methods to better visualize and interpret potentially large sets of models. In addition, we are considering stochastic methods to accurately sample and characterize the \"space\" of all possible fuzzy models to (a) more efficiently identify the subset of plausible models and (b) identify common patterns among all the models to gain a better understanding of the system and its evolution. While it is tempting to develop methods to obtain a single \"optimal solution\" as in a classic inverse problem, this is not appropriate for complex biological systems. Scarcity of both data and biological understanding mean that at best experiments will merely limit the space of potential solutions. Biological system analysis is a *dynamic*reverse engineering problem, requiring continuous acquisition of new experimental data -- data that should be acquired from experiments designed and informed by continuous modeling. Linear fuzzy rule network models are a promising methodology for an integrated modeling and experimental approach. Since fuzzy rule models are enumerable, methods developed for combinatorial optimization can be extended to them. Moreover, linear fuzzy network models can simultaneously contain both quantitative and qualitative information, providing a common framework for a broad range of biological data, including mass spectrometry analysis, RT-PCR, single cell imaging, metabolite profiling, and other technologies yet to be developed. Methods ======= Converting between numerical data and fuzzy sets ------------------------------------------------ We use three fuzzy sets, Low (or **1**), Medium (**2**), and High (**3**) to represent the magnitude of gene expression, as defined in Figure [5](#F5){ref-type="fig"}. *Fuzzification*(conversion to fuzzy representation) of a numerical datum *x*is performed by finding the corresponding fuzzy set memberships *y*~1~, *y*~2~, and *y*~3~(with values ranging from 0 to 1.0) given the linear functions shown in Figure [5](#F5){ref-type="fig"}, where  *Defuzzification*(conversion back to numerical representation) is performed using the \"simplified centroid method\" \[[@B29]\], with point set definitions shown in Figure [5](#F5){ref-type="fig"}. Following a fuzzy rule evaluation that returns fuzzy set memberships **y**= \[*y*~1~ *y*~2~ *y*~3~\] in sets **1**(Low), **2**(Med), and **3**(High) respectively, the estimated numerical result of the evaluation, , is given by the centroid for the points located at -1, 0, and +1 for each set respectively, or  The fuzzy set definition and centroid defuzzification of Equations 1 and 2 were selected to maximize computational efficiency during exhaustive search: all 27 rules can be represented by easily implemented algebraic functions and it is possible to design the implementation to avoid as many costly if/then comparisons as possible. In addition, the scheme perfectly reproduces monotonic linear positive and negative interactions (i.e., the functions *f*(*x*) = *x*and *f*(*x*) = -*x*are quantitatively equal to monotonic fuzzy rules, which can be written using notation from the following section as \[1 2 3\] and \[3 2 1\] respectively) so it generally will not introduce systematic error in the model. To apply this scheme for defuzzification and fuzzification scheme, experimental data must be projected on to the interval -1.0 through +1.0. Thus, log base 2 expression ratios are normalized by taking the arctangent of each ratio and dividing by π/2, yielding a symmetric transformation covering the desired interval. Previous work normalized expression ratios by the maximum value found in the experiment \[[@B17]\] or used different fuzzy set definitions for each variable \[[@B16]\], but those approaches suffer from a lack of universality across data sets and makes it difficult to compare and integrate data from different experiments. On the other hand, the arctangent method is defined across infinity, so no data will be \"out of range\". It also takes into account the fact that gene expression ratios often \"saturate\", and the difference between different degrees of high and low ratios are not necessarily biologically significant (this is because of the optical methods for measuring microarrays and the exponential error introduced using RT-PCR). When used in conjunction with the overlapping fuzzy set mappings shown in Figure [5](#F5){ref-type="fig"}, these \"middle\" values will tend to land in the Medium set (**2**). Comparing fuzzy predictions to numerical data --------------------------------------------- The fuzzy rule relating the input of a single gene to an ouptut node gene can be expressed as a rule vector **r**. For example, the rule **r**= \[3 2 1\] corresponds to the linguistic rules: If *Input*is Low (**1**) then *Output*is High (**3**) If *Input*is Med (**2**) then *Output*is Med (**2**) If *Input*is High (**3**) then *Output*is Low (**1**) Given the fuzzified expression of an input gene **y**= \[*y*~1~ *y*~2~ *y*~3~\] obtained using Equation 1 and the general fuzzy rule **r**= \[*r*~1~ *r*~2~ *r*~3~\], the resulting fuzzified expression of the output gene **z**will be:  In general, node behavior is the result of *N*input genes acting on the output gene simultaneously. In the linear fuzzy logic scheme, the rule for each input gene is evaluated separately, leading to intermediate outputs **z**^*i*^:  These intermediate fuzzy values are summed algebraically to obtain the final resulting fuzzy value for node gene expression:  This result is defuzzified using Equation 2 to evaluate the output of the node. For three fuzzy sets, there are 3^3^or 27 possible rules describing the effect of a single gene on another gene. Thus, if there are *N*input genes for a node, there are 27^*N*^total possible rule combinations describing the behavior of the node gene. In general, no rule combination will be an exact fit to real experimental data. Given some tolerance to fitting error, there will be multiple possible rule combinations, representing plausible hypothetical gene network models. In our present work, we define the error of the fit for the *M*data of the output gene **x**= {*x*~1~,*x*~2~,\...,*x*~*M*~} as  where  is the set of defuzzified numerical predictions (typically log expression ratios) and  is the mean of the experimental data set **x**. A perfect fit results in a maximum *E*of 1.0. This error score was chosen because while it is quantitative, it emphasizes the correlation in qualitative behavior between the fit and prediction instead of the absolute numerical fit, which can be difficult to model with the limited resolution of three fuzzy sets. We can use the fitting error to rank these models, and use rule patterns consistent throughout all plausible models as a basis for constructing the template of a final network model that can be tested experimentally. Example of fuzzy rule evaluation -------------------------------- As an example to illustrate fuzzy gene networks using a simple rule combination, we consider three genes (*G1*, *G2*, *G3*) with log base 2 expression ratios measured at three different times: *G*1 = {-3.0 0 +3.0} *G*2 = {0.3 0 -0.3} *G*3 = {+1 0 -1.0} Using the arctangent normalization to project the ratios on \[-1,1\], we obtain *G*1 = {-0.795 0 +0.795} *G*2 = {+0.186 0 -0.186} *G*3 = {+0.500 0 -0.500} which can be fuzzified using Equation 1 to yield: *G*1 = {\[0.795 0.205 0\] \[0 1.0 0\] \[0 0.205 0.795\]} *G*2 = {\[0 0.814 0.186\] \[0 1.0 0\] \[0.186 0.814 0\]} *G*3 = {\[0 0.5 0.5\] \[0 1.0 0\] \[0.5 0.5 0\]} with vectors for each time point containing set membership in Low (**1**), Medium (**2**), and High (**3**). Consider the following rules for *G1*and *G2*as input genes to *G3*: *G*1:*G*3 = \[3 2 1\] *G*2:*G*3 = \[1 2 3\] where the rules can be written in English as If *G1*is Low (**1**) then *G3*is High (**3**) If *G1*is Med (**2**) then *G3*is Med (**2**) If *G1*is High (**3**) then *G3*is Low (**1**) If *G2*is Low (**1**) then *G3*is Low (**1**) If *G2*is Med (**2**) then *G3*is Med (**2**) If *G2*is High (**3**) then *G3*is High (**3**) Now, the evaluations of the rules taken individually are *G*1:*G*3 = {\[0 0.205 0.795\] \[0 1.0 0\] \[0.795 0.205 0\]} *G*2:*G*3 = {\[0 0.814 0.186\] \[0 1.0 0\] \[0.186 0.814 0\]} The sum of two intermediate outputs (Equation 3) is the predicted fuzzy behavior of *G3*for the three time points, which can be defuzzified using the point-centroid method (Equation 2) and transformed back to real numbers on \[-1,1\]: *G*3 = {\[0 1.019 0.981\] \[0 2.0 0\] \[0.981 1.019 0\]} = {0.491 0 -0.491} These numbers can be transformed back to a Log2 expression ratio by inverting the normalization (multiplying by π/2 and taking the tangent): *G*3 = {0.97 0 -0.97} Finally, we use Equation 4 and the experimental data for *G3*to calculate the fit error for this rule combination:  which compares to a maximum *E*= 1.0 for a perfect fit. Exhaustive network search ------------------------- In general, a possible model for a node can include any combination of the genes available to act as inputs. In the work described here, we consider potential interactions of 12 genes. Thus, a rule for any one gene can include as inputs any combination of any number of up to all 11 other genes. Since each input gene can influence the node by any one of the 27 possible fuzzy rules, there are approximately 10^16^possible rule combinations for each of the 12 genes, making the exhaustive search method practically impossible. Thus, the number of possible inputs to a node must have a maximum constraint to make exhaustive search tractable. Studies of network topology through the experimentally observed association of proteins suggest that in many cases only few regulatory proteins are observed to directly influence the expression of a gene \[[@B26],[@B30]-[@B32]\]. For our transcriptional network searches, we use the constraint of up to 4 input genes to any node. Thus, for each node gene, each of the other 11 genes occurs as an input alone and also in combination with any of up to 3 of the other genes as multiple inputs. Our use of this input constraint does not necessarily restrict the full range of interactions that can be found for the genes in our network, since *all*possible combinations of 1 through 4 of the genes are searched sequentially. For example, in our fitting of rules to *CLN3*, we considered the following potential input combinations: *SIC1*alone, *SIC1*and *CLN1*together, *SIC1*-*CLN1-CLN2*, *SIC1-CLN1-CLN2-CLN3*, *CLN1*, *CLN1-CLN2*, *CLN1-CLN2-CLN3*, *CLN1-CLN2-CLN3-SWI4*, *CLN2*, *CLN2-CLN3*, etc. If we include all combinations from 1 through 4 of the genes taken from the 11 total possible inputs, then the total search space for each of the 12 genes consists of approximately 10^8^rules (taking about 10 minutes on a PowerMac G4 using a single 450 MHz processor). Simulation files used to generate all the data presented here are available from the authors upon request. Authors\' contributions ======================= BAS originated the concept of applying scalable (URC) fuzzy logic to modeling biological systems, implemented the scheme described within on the data set, and was the primary author. JPF conceived of the approach to use exhaustive search for biological network reconstruction. JNQ and AAQ developed and initially implemented the method of combinatorial input selection for the exhaustive network search, and AAQ contributed to the text. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 Microsoft Excel spreadsheets of simulation results. See descriptive text in the workbook. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ Thanks to Drs. J. Kercher, A. Golumbfskie, B. Pesavento, and K. S. Kim for reviewing drafts of the manuscript and providing important insights. Allan Gu helped with software integration. This work was performed under the auspices of the U.S. Department of Energy by the University of California, Lawrence Livermore National Laboratory under contract No. W-7405-Eng-48. This research is funded by the Laboratory Directed Research and Development (LDRD) Program at Lawrence Livermore National Laboratory (LLNL). The LDRD Program is mandated by Congress to fund director-initiated, long-term research and development (R&D) projects in support of the DOE and national laboratories mission areas. The Director\'s Office LDRD Program at LLNL funds creative and innovative R&D to ensure the scientific vitality of the Laboratory in mission-related scientific disciplines. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Known Yeast Cell Cycle Interactions.**Schematic of known yeast cell cycle interactions between protein products of the twelve genes we studied (Table 1). An arrow indicates a positive interaction and a closed circle indicates a negative interaction. A dashed line indicates a protein-protein (e.g., phosphorylation, etc.) interaction and solid lines indicate direct transcriptional regulation. Genes are shown adjacent to each other when they are treated as a cluster in the KEGG pathway diagram (in the case of SBF and MBF the genes code for protein subunits). A connection inside a cluster indicates regulation of the whole cluster simultaneously. The interactions are located along an approximate time axis through the stages of the cell cycle (G1, S, G2). :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Fuzzy Rule Search Convergence.**Histogram of the number of fuzzy rules for *CLN1* expression at different fit error (*E*) levels. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Best Fit Model of Yeast Cell Cycle Interactions.**Diagram of the best fitting gene interactions for each node gene, as shown in Table 3. Solid lines ending in an arrow represent positive interactions (induction, e.g., \[1 2 3\], \[1 3 3\], etc.), dotted lines ending in a closed circle represent monotonic negative interaction (repression, e.g., \[3 2 1\], \[3 1 1\], \[1 1 2\], etc.), and dash-dot lines ending in a diamond represent an interaction that is both positive and negative (e.g., \[3 1 3\], \[1 3 1\], etc.). :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Fuzzy Model Fit to Numerical Data.**Comparison of experimental (solid black line) gene expression ratios from the *alpha*cell synchronization data set (Spellman, *et al*., 1998) and the predicted time series of the best-fitting fuzzy model (grey line) for genes *CLN1*, *E*= 0.930 (**A**), *CDC28*, *E*= 0.510 (**B**), *SWI6*, *E*= 0.620 (**C**), and *CLB5*, *E*= 0.881 (**D**). :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Numerical-Fuzzy Data Conversion.**Fuzzification (top) and defuzzification (bottom) schemes used in this analysis, showing conversion to and from values in the interval \[-1,1\] on the x-axis and fuzzy sets LOW (**1**), MED (**2**), and HIGH (**3**). See text for details. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### List of Genes in the Fuzzy Network Model Subset of genes involved in the yeast cell cycle selected for our demonstration network. Descriptions taken from the Yeast Protein Database (Costanzo *et al*., 2000; URL: <http://www.proteome.com/YPDhome.html>). ::: **GENE NAME** **ORF** **DESCRIPTION** --------------- --------- ----------------------------------------------------------- SIC1 YLR079W inhibitor of the Cdc28-Clb protein kinase complex CLN1 YMR199W G1/S-specific cyclin CLN2 YPL256C G1/S-specific cyclin CLN3 YAL040C G1/S-specific cyclin SWI4 YER111C transcription factor, subunit of the SBF factor SWI6 YLR182W transcription factor, subunit of SBF and MBF (check this) CLB5 YPR120C B-type cyclin CLB6 YGR109C B-type cyclin CDC6 YJL194W initiates DNA replication, active late G1/S CDC20 YGL116W cell division control protein CDC28 YBR160W cyclin-dependent protein kinase MBP1 YDL056W transcription factor, subunit of MBF ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Best Fitting Rules for *CLB5*Each row represents rules relating inputs (in columns) to *CLB5*. Rule format described in Methods; e.g., \[3 3 1\] for *CDC20*means \"If *CDC20*is **1**(Low) Then *CLB5*is **3**(High); if *CDC20*is **2**(Med) then *CLB5*is **3**(High); if *CDC20*is **3**(High) then *CLB5*is **1**(Low).\" ::: **Fit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{#T3 .table-wrap} Table 3 ::: {.caption} ###### Three Best-Fitting Models for Each Gene\* Rules are in the same format as in Table 2 (see Methods). In this case, the genes in rows are inputs to genes in columns. The three best combinations of rules are in consecutive rows for each input; i.e., the best fit rule combination is the first row of each input block with the fit score (*E*) in the first row of the last block in the table. \* Models are in rows, e.g., the best-fitting model for *SIC1* has the input *CLB5* with the rule (1 3 3), *MBP1* with rule (3 1 1), and *CDC6* with rule (1 2 3) corresponding to the fit error *E*= 0.689. ::: *OUTPUTS* --------------- ----------- ------- ------- ------- ------- ------- ------- ------- ------- ------- ------- ------- **SIC1** \-\-- 2 2 3 3 2 1 1 1 3 \-\-- 2 2 3 2 1 3 1 3 3 3 2 1 1 2 3 \-\-- 2 2 3 2 1 3 1 3 3 3 2 1 1 2 3 **CLB5** 1 3 3 \-\-- 1 3 3 1 3 3 \-\-- 1 1 3 1 3 3 \-\-- 1 1 3 3 2 1 **CDC20** 3 3 1 \-\-- 3 3 2 3 2 1 1 3 3 2 3 1 \-\-- 3 3 1 2 3 1 \-\-- 3 1 1 1 3 3 **CLN3** 3 1 2 \-\-- 3 3 1 2 2 3 3 3 2 1 2 3 \-\-- 3 2 1 2 1 3 1 1 3 3 1 3 \-\-- 3 3 1 1 1 3 **SWI6** 1 3 1 3 2 1 \-\-- 3 1 2 **(SBF)** 1 3 1 3 1 1 \-\-- 1 2 1 3 1 2 2 3 1 1 3 1 3 2 1 \-\-- 1 2 3 3 1 1 **CLN1** 3 3 1 \-\-- 1 2 3 1 2 3 1 1 2 3 1 1 3 2 1 \-\-- 1 3 3 1 2 3 3 1 1 1 2 3 3 1 1 \-\-- 1 3 3 1 3 3 3 1 2 **CLN2** 1 1 3 3 1 1 3 1 3 1 3 3 \-\-- 1 1 3 1 1 3 3 2 1 3 1 3 1 3 3 \-\-- 1 1 3 1 1 3 3 1 3 1 3 3 \-\-- 1 1 3 **CLB6** 1 2 3 1 1 3 1 1 3 \-\-- 1 1 3 1 3 3 3 1 3 1 2 3 1 1 3 1 1 3 1 1 3 \-\-- 1 1 3 1 3 3 3 3 1 1 1 3 1 1 3 \-\-- 1 2 3 1 3 3 **SWI4** \-\-- 3 2 1 **(SBF)** 1 1 3 2 3 3 \-\-- 3 3 2 \-\-- 3 2 1 1 2 3 **CDC28** 3 1 3 \-\-- \-\-- 3 1 2 \-\-- **MBP1** 3 1 1 3 3 1 2 1 3 3 3 1 \-\-- **(MBF)** 3 1 1 2 3 1 1 2 3 3 3 1 \-\-- 3 2 1 3 3 1 2 3 2 1 1 3 3 3 1 2 2 3 \-\-- **CDC6** 1 2 3 1 2 3 1 1 3 3 2 3 \-\-- 1 3 3 1 2 3 1 3 1 3 2 3 3 2 3 \-\-- 1 2 3 2 2 3 2 1 3 3 2 3 \-\-- **Fit (*E*)** 0.689 0.881 0.703 0.792 0.620 0.943 0.877 0.714 0.782 0.510 0.624 0.764 0.643 0.875 0.702 0.768 0.599 0.930 0.837 0.706 0.768 0.508 0.624 0.758 0.631 0.852 0.695 0.746 0.596 0.916 0.835 0.705 0.763 0.498 0.622 0.720 :::

PubMed Central

2024-06-05T03:55:47.181158

2004-8-10

PMC514699

Background ========== In a relatively short time microarrays have become a well established technique, widely used by researchers. Microarrays offer nothing less than to be able to monitor simultaneously the expression levels for thousands of genes. The RNA molecules from the biological sample are called *targets*, and the polymers of nucleic acids fixed on the surface are called *probes*. The very large number of genes represented on each microarray requires the use of computer based approaches. Although such approaches currently constitute a very rich and active area of research, for many data analyses this step can be summarized simply: under the prior assumption that for the large majority of the genes represented on a microarray the expression will not vary significantly across experiments, the main focus is be to isolate the few genes of interest from the rest. Short oligonucleotide arrays are a particular type of microarrays. Short oligonucleotide arrays are constituted of short probes (oligonucleotides) with several probes designed to match different part of the target sequences. The use of techniques originating from the micro-electronics industry proved very successful in the making of short oligonucleotides arrays \[[@B1]\], helping the *Affymetrix*company establish itself as one of the primary manufacturers for microarrays. The pre-processing of oligonucleotide array data differs from other microarray data, specifically because probe intensities associated with each gene are generally summarized by an expression value, or expression index. As it contributes to the computation of the expression values, this step alone is of importance. Different algorithms have been suggested to replace and improve Affymetrix\'s original algorithms \[[@B2]\], including E. Lazaridis *et al*.\'s *playerout*\[[@B3]\], Li and Wong\'s model \[[@B4]\], medianpolish \[[@B5]\], and Affymetrix\'s own improvements to its algorithms \[[@B6]\]. However, information about the individual probes was not disclosed until a few years ago. Only with the release of the probe sequences for a significant number of Affymetrix chips, data analysis approaches considering the nature of individual probes have been made possible. The use of the chemical nature of the probes, on which depends the binding energy with complementary sequences, has already been suggested to improve pre-processing of Affymetrix data at the probe level \[[@B7],[@B8]\]. The annotation for sequenced genomes have progressed considerably since the design of the Affymetrix chips, even the most recent ones, and matching the latest transcriptomic (or genomic) data available with the chip designs is an obvious thing to do. We have performed such a remapping for a few Affymetrix chips, and we show that the resulting probe-to-gene mapping can differ substantially from the original Affymetrix mapping. This can affect the interpretation of experimental data. As annotations of genomes continue to evolve, it is also desirable to have a framework to perform and handle up-to-date probe-to-gene mapping. We provide an open source and documented implementation to do so. Results ======= The results obtained are subdivided in two main categories: the matches between probes and reference sequence obtained, and the difference in the outcome of an analysis when using an alternative mapping. Probes matching multiple RefSeq entries --------------------------------------- A fair number of probes were found to match several reference sequences, as shown in Figure [1](#F1){ref-type="fig"}. For example, the RefSeq *NM\_001544.2*is found to have 21 matching probes. Eleven of these matching probes also match another reference sequence: *NM\_022377.1*. A quick look at the annotation reveals that both reference sequences are two different transcripts variants of the same gene \'Homo sapiens intercellular adhesion molecule 4, Landsteiner-Wiener blood group (ICAM4)\' and that the same probes are found matching these two sequences. However the ten remaining probes matching *NM\_001544.2*are also found matching a fairly large number of other reference sequences (from a little less than 300 to almost 600 reference sequences, depending on the probe). We found that these probes are designed to match sub-sequences frequently found in mRNA: ALU repeats. All the probes matching ALU repeats are in the *official mapping*grouped in a common probe set, called \'human ALU\'. Besides \'human ALU\' probes, other probes matching multiple reference sequences were found. In that case, the reference sequences matching a given probe have closely related annotations, or even identical annotations. Complex mixtures of partial overlaps for the probe sets can then be observed. As an example, the probe 88322 is found matching the reference sequences *NM\_017445.1*, *NM\_003519.3*, *NM\_003520.3*, *NM\_003521.2*, *NM\_003525.2*, *NM\_003528.2*, *NM\_080593.1*and *XM\_301109.1*, annotated \'H2B histone family, member S (H2BFS), mRNA\', \'histone 1, H2bl (HIST1H2BL), mRNA\', \'histone 1, H2bn (HIST1H2BN), mRNA\', \'histone 1, H2bm (HIST1H2BM), mRNA\', \'histone 1, H2bi (HIST1H2BI), mRNA\', \'histone 2, H2be (HIST2H2BE), mRNA\', \'histone 1, H2bk (HIST1H2BK), mRNA\' and \'similar to Histone H2B 291B (LOC350694), mRNA\' respectively. Such matches in each case require expert annotators to curate the *alternative mappings*, so we chose to simply ignore the probes matching several reference sequences in the rest of this study. Other multiple matches are more easy to handle, and potentially more harmful when included in an analysis. Some probes are found to hybridize to several unrelated reference sequences, as shown in Figure [2](#F2){ref-type="fig"}. Only 290 probes that were labeled mismatches in the *official*mapping, were found to be legitimate perfect match probes in the *alternative*mapping. Reference sequences matching all the probes from probe sets ----------------------------------------------------------- We also found a significant number of reference sequences for which all the matching probes belong to one probe set in the *official mapping*. When a \'one to one\' association can be established between a probe set of the *official mapping*and a probe set in an *alternative mapping*, which means that a given reference sequence matches all the probes associated with one *Affymetrix ID*, we conclude a complete agreement between the *alternative mapping*and the *official one*(See Figure [3](#F3){ref-type="fig"}, top). That is the case for 6274 out of 17426 reference sequences (17426 is the number of reference sequences for which at least one matching probe was found). When the association is \'one-to-many\', in the sense that several complete probe sets in the *original mapping*are matching one reference sequence, one could conclude that, alternative splicing events left aside, some probe sets are redundant (See Figure [3](#F3){ref-type="fig"}, bottom). We obtain 1168, 212, 38, 4 and 2 reference sequences for which the matching probes are coming from 2, 3, 4, 5 and 6 *original*probe sets respectively. The Figure [4](#F4){ref-type="fig"} shows that it represents a significant part of the cases. This represents 8% of the *original*probe sets of a HG-U133A that are potentially redundant. Effect on the outcome of an analysis ------------------------------------ Naturally the expression values, or expression indexes, computed from the probes intensities are sensitive to differences in the mapping: different probes will give different summary expression values, which can have an effect on the outcome of an analysis. To verify it, we performed a standard exploratory analysis of Affymetrix data (two samples, looking for the genes that are significantly differentially expressed). The probe level intensities were pre-processed and expression values computed. The *original mapping*was used to obtain a first set of expression values (set Affy), while the *alternative mappings*made from matching NCBI\'s reference sequences was used to obtain two more sets of expression values (sets Alt1 and Set Alt2), using all the matching probes or all the probes matching uniquely respectively. In other words, the set Alt2 differs from the set Alt1 in the sense that probes in Alt1 matching several reference sequences were removed from Alt2. The set Affy describes 22283 probe sets, the set Alt1 describes 18076 probe sets (for a total of 184735 probes) and the set Alt2 11640 probe sets (for a total of 153257 probes). The number of mismatch probes in the *official mapping*that are found in the *alternative*mappings is low: 290 in Alt1 and 87 in Alt2. For each one of the three sets, \'significantly differentially expressed genes\' (SDEGs) are searched for: a Welch\'s two-sample t-test is performed on all the expression values in each set, and the selection for significant p-values done as described by Ventura and collaborators \[[@B9],[@B10]\] (qvalue set to 1%). The number of SDEGs obtained in the sets Affy, Alt1 and Alt2 are 163, 163, and 103 respectively. Table [1](#T1){ref-type="table"} shows how many of these represent identical probe sets. However, there is also a significant fraction of probe sets for which no simple equivalence could be made. As shown when discussing the matches, the situation is rather complex and a detailed examination for each case would be needed before a comparison between the mappings is possible. The presence of mismatch probes in the *alternative*mappings does not appear to have much influence. A legitimate concern can be that some of the probe sets in the *alternative mappings*only contain one or two probes, therefore the results obtained with theses probe sets may be dubious. In fact, only very few of these probe sets contain a small number of probes, as shown in Figure [5](#F5){ref-type="fig"}. Moreover, the minimal acceptable number of probes in a probe set has been reported to be lower than the number of probes commonly used \[[@B11]\]. Validation of the results found with our *alternative mappings*will have to be done *in silico*through the curation of the mappings by expert annotators and experimentally with techniques like RT-PCR. Software for inter-exchange mappings ------------------------------------ We present a complex situation, with new probe sets built on matches between NCBI\'s RefSeq reference sequences and the sequences of the short oligonucleotide probes found on commercial arrays. This would be of little practical use for the research community without an easy access to the data or the tools used to obtain them. The framework developed in the package \'affy\', an open-source and documented collection of data structures and functions for the analysis of *GeneChip*oligonucleotide arrays at the probe level, is currently used by a growing number of researchers. We take advantage of the features it offers by providing *alternative mapping*objects that can be \'plugged in\', and used instead of the original ones, whenever wanted. The Bioconductor package \'altcdfenvs\' contains helping functions, and documentation explaining how to achieve this. Discussion ========== We performed a matching of the Affymetrix probes against the latest reference sequences from NCBI\'s RefSeq data bank. A number of probes appear to match a large number of reference sequences, hence to match a large number of transcripts. When analyzing a real data set with state-of-the-art processing methods, we observe that the outcome of an analysis can be influenced by inaccuracies in the probes mapping. This is a potential problem when more and more people use \'high-throughput\' procedures, to select \'significantly important genes\' in an automated or semi-automated fashion. We introduce an alternative mapping between probes and identifiers, our identifiers being NCBI\'s RefSeq IDs, and offer for download alternative mappings for the Affymetrix chips HG-U95Av2 and HG-U133A. The *affy*package is a free environment to work with *Affymetrix*data at the probe level. It has capabilities to include alternative mapping and use them in a simple way. Our work aims at showing potential problems. Biological expertise remains necessary to discuss the exact nature of each match. *Affymetrix*offers at their NetAffx web site \[[@B12]\] a tool that allows visualization of probes matching to multiple sequences. As the annotation of the human genome improves over time, our environment allows to update the probe-to-gene mapping accordingly and analyze microarray data using the best biological data and knowledge available. The environment for alternative mappings we present could also be of use when using *GeneChips*designed for a certain organism with mRNA known to differ (a mutant or a slightly different organism for example). Conclusions =========== We built new \'probe to gene\' mappings by matching the sequences for the probes found on human Affymetrix GeneChip arrays against NCBI\'s RefSeq database of sequences. Solely observing the distribution of the probes matching reference sequences, and comparing the results with the mapping provided by the manufacturer of the chips, we found potential problems such as probes matching several reference sequences and probes from different probe sets in the official mapping matching all the same reference sequence. Depending on the mapping used, the outcome of the data analysis will change as different genes may be selected as differentially expressed. We suggest that a good mapping of probe to genes changes in time, follows the most recent updates in sequence databases (the databases of sequences are constantly modified, with sequences corrected, and sometimes hypothetical gene sequences withdrawn). We like to picture the current situation, where the mappings are frozen, as a Dorian Gray-like syndrome: the apparent eternal youth of the mapping does not reflect that somewhere the \'picture of it\' decays. We developed a set of open-source tools, perfectly integrated to an already existing working environment for Affymetrix arrays. The tools are documented and made available to the research community. They let one reproduce our results, and build other mappings. Methods ======= Reference sequences and probe sequences --------------------------------------- The database of reference sequences used for remapping the Affymetrix chips is NCBI\'s RefSeq (first release of NCBI\'s RefSeq, dated June 30th, 2003), freely available for download on the RefSeq website \[[@B13]\]. Only sequences tagged as *Homo sapiens*mRNA are considered. The Affymetrix chip types HG-U95Av2 and HG-U133A, both designed for the study of the human transcriptome, were used. While similar features were observed for both, we focus on the HG-U133A to describe our results, as this is the chip with the most recent design. The probe sequences for the chips are freely available on the Affymetrix website \[[@B14]\], as well as on the Bioconductor website \[[@B15]\] as meta-data packages. In total, there are 495930 probe sequences on the chips of type HG-U133A (245965 perfect match probes and 245965 mismatch probes). Probe matching -------------- *Affymetrix GeneChips*have several probes per probe set, and a probe set usually represents a gene. Each probe is 25 bases long, and on most arrays probes are grouped in pairs. A probe pair is constituted of a perfect match (pm) probe, designed to match perfect a target gene sequence, and a mismatch (mm) probe, designed to measure non-specific hybridization. The mismatch probe differs from its associated perfect match probe only in the 13th base. In some cases, a probe set represents a fragment of a gene, e.g, 3-prime and 5-prime extremities of the same gene are used as internal control for the efficiency of the reverse transcription. We prefer the term \'probe set ID\' to \'gene\' since probe sets are not always genes. We call the association \'probes -- probe set ID\' a *mapping*. We refer to the grouping of probes in probe sets given by *Affymetrix*as the *official mapping*, while a grouping of the probes matching a reference sequence into a probe set is referred to as an *alternative mapping*. The short length of the probes, as well as several authors reporting a successful use of perfect matches only \[[@B3]-[@B5],[@B16]\], suggest that the hybridization signals coming from exact matches alone are able to capture the expression signal reliably. We consider all the probes (including *mm*) as potential *pm*probes to perform the matching. The matching of a probe to a sequence is done using the Bioconductor package *matchprobes*. It performs an exact string matching, as done by the standard C library *string:*only complete sequence identity between a probe and a fragment of a reference sequence is counted as a match. Experimental data ----------------- We collected pancreatic tumor tissue from 8 patients and normal pancreatic tissue from 5. Tumor samples were collected from patients undergoing surgery for pancreatic tumors, quick frozen in liquid nitrogen and kept at -80°C until RNA extraction (the study was approved by the ethical committee for Copenhagen). The tissue was homogenized using a Polytron (kinematica. AG, Littau-Luzern, Schwitzerland). 5 μg RNA was extracted from each sample and labeled. The RNA was extracted according to the Trizol protocol (Invitrogen). The samples were applied to Affymetrix HG-U133A GeneChips according to manufacturer\'s instruction. Clinical results from this study will be published elsewhere (manuscript in preparation). Data processing --------------- The probe level data are pre-processed using the affy software package \[[@B17]\]. No background correction is performed, and probe level intensities are normalized using the *vsn*\[[@B18]\] normalization method. Summary expression indexes are computed using the method *medianpolish*, using only perfect match probes. The exclusive use of perfect match probes is needed to have an identical pre-processing step for the *official mapping*and *alternative mappings*, to allow comparison of the results obtained. Availability ------------ All the datasets and software used are freely available, and on a wide range of platforms (Linux, Microsoft Windows, MacOS X, other UNIX-like operating systems). The original material used to obtain our results, software and data, is available from the web page: Alternative mappings <http://www.cbs.dtu.dk/staff/laurent/download/maprefseq/>. The package altcdfenvs, new with the Bioconductor release 1.4 of May 2004 and available on the Bioconductor website, is designed to help researchers to build their own alternative mappings, using their own set of reference sequences and the probe sequences for the Affymetrix chip of their choice. Authors contributions ===================== LG conceived, developed and implemented the concept of alternative mapping for Affymetrix chips, and drafted the manuscript. MM and LFH provided original experimental data. SK provided guidance with the choice of a source for reference sequences, and with comments and input on the manuscript. Acknowledgements ================ The authors wish to thank three anonymous referees for crucial comments, and for believing the manuscript could be improved enough. The work of Laurent Gautier was funded by a grant from the Danish Biotechnology Instrument Center. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Histrograms for the number of probes per probe set.**Both plots are histogram, the values on the y axis are relative frequencies. (top) Histogram of the number of probes on the U133A chip matching a particular RefSeq transcript. A clear peak can be seen for 11 probes, which corresponds to the number of probes in a probe set most commonly found in the *original mapping*. This set is called Alt1 in the last part of the section Results. (bottom) Histogram of number of RefSeq transcripts matching a single probe (log scale). Most of the probes can be seen matching only one RefSeq. In both plots, the probes associated to human \'ALU\' repeats were filtered out. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Reference sequences matching the probe 122174.**The reference sequences matching the probe 122174, assigned to the probe set *211697\_x\_at*and annotated \'Homo sapiens RNA-binding protein LOC56902 mRNA, complete cds\' in the *original mapping*. NCBI\'s reference sequences are annotated \'paraneoplastic antigen MA2 (PNMA2), mRNA\', \'LOC202934 (LOC202934), mRNA\', \'hypothetical protein LOC283507 (LOC283507), mRNA\', \'sarcalumenin (SRL), mRNA\', \'LOC284095 (LOC284095), mRNA\', \'putative 28 kDa protein (LOC56902), mRNA\' and \'sodium channel, voltage-gated, type III, beta (SCN3B), mRNA\' respectively :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Probes, probe sets and reference sequence.**Reference sequence matching all the probes from probe sets in the *official mapping*. The reference sequence is represented by a long dark cylinder, while the matching probes are represented by red or yellow fragments of cylinder. The wire frame represents the grouping of probes in a probe set in the *official mapping*. (Top:) All the probes matching the reference sequence constitute a probe set in the *official mapping*. (Bottom:) All the probes matching the reference sequence constitute two different probe sets in the *official mapping*. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Scatter plots for the number of probes matching a reference sequence.**Scatter plot of the total number of probes matching a reference sequence against the number of probes remaining after removing the probes matching several reference sequences. Colored areas are displayed to indicate the z-axis, the number of probes occupying each spot in the graph. A grid-like pattern can be observed in the lower-left corner of the plot. The size of the cells is 11 probes, which is the number of probes contained in most of the probe sets in the *official mapping*. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Histograms of the number of probes per probe in the SDEGs.**Distribution of the number of probes per probe set in the sets of significantly differentially expressed genes for the set Alt1 (top) and the set Alt2 (bottom). The probe sets for which an identical probe set could be found in the Affy set are represented in red. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Number of probe sets Number of complete probe sets (see Figure 3, top) in common between the sets of significantly differentially expressed genes (SDEGs) Affy, Alt1 and Alt2 (163, 103 and 103 probe sets respectively). ::: Affy Alt1 Alt2 ------ ------ ------ ------ Affy 61 54 Alt1 61 95 Alt2 54 95 :::

PubMed Central

2024-06-05T03:55:47.185521

2004-8-14

PMC514700

Background ========== Knowledge about the metabolic network of an organism is essential for understanding its physiology and phenotypic behavior. A comprehensive understanding of the metabolic network at the system level is particularly important for both biotechnological and biomedical research and is now made possible by rapid advances in genome sequencing and functional genomics. *In silico*reconstruction of metabolic networks from genome sequences of organisms represents a starting point for a systematic analysis of metabolism \[[@B1]-[@B3]\]. The functionality of the potential metabolic network of a given organism can then be further experimentally studied by system perturbations at both physiological and genetic levels \[[@B4]\]. Several methods and tools have been recently developed for the reconstruction, visualization and analysis of metabolic networks. These include general static metabolic network tools such as the Kyoto Encyclopedia of Genes and Genomes (KEGG) \[[@B5]\], the Boehringer Mannheim metabolic charts \[[@B6],[@B7]\] and dynamic or potentially dynamic tools such as WIT \[[@B3]\], MPW \[[@B8]\], EcoCyc \[[@B9],[@B10]\] and PathFinder \[[@B11]\]. Metabolic network reconstruction is generally based on the identification of metabolic enzymes and the corresponding biochemical reactions in a specific organism. For this purpose the EC numbers of all possible enzymes need to be determined. The set of EC numbers of an organism may be obtained from the genome annotation. This conventional approach of metabolic network reconstruction is briefly summarized in Fig. [1A](#F1){ref-type="fig"}. It covers three successive steps: (1) gene finding, (2) database searching and function assignment and (3) metabolic reconstruction. In the first step, genes or coding sequences (CDSs) are predicted from the genome data using programs such as Glimmer \[[@B12]\], GeneMarkS \[[@B13]\], ZCURVE \[[@B14]\] or CRITICA \[[@B15]\]. Then, coding sequences are used as queries and compared to sequence databases such as GenBank, GenPept and SWISS-PROT or to databases of protein domains and functional sites such as InterPro \[[@B16]\], PROSITE \[[@B17]\], Pfam \[[@B18]\] etc. Based on the similarity, the function of the database entry may be assigned to a CDS as its annotation. From these function assignments the metabolic network can be constructed. This three-step approach is used for example in the WIT system \[[@B3]\]. Efforts were also made in some bioinformatic systems such as WIT to reconstruct metabolic networks from incomplete genome data \[[@B3]\]. The program suite ERGO™, a commercial version of WIT, integrates over 400 finished and unfinished genomes into a comprehensive network of metabolic and non-metabolic pathways \[[@B19]\]. No details about the WIT approach have been published and merely some information about 55 annotated genomes (Status: March 2004) is publicly available on the website of WIT \[[@B20]\]. The three-step method starting from gene finding has several drawbacks for the reconstruction of the metabolic network from incomplete or unfinished genome sequences. In unfinished sequences of a genome, especially in sequences with a low genome coverage (e.g. less than 4 fold), there may be many sequencing errors that do not warrant an accurate prediction of genes \[[@B21]\]. For example, the start or stop positions of CDSs may not be accurately predicted. Protein sequences translated from these CDSs may be completely wrong because of coding frame shifts. Fusion CDSs may be predicted, to which the function assignment is difficult. Moreover, a CDS that normally appears as one CDS in other organisms may be predicted as several smaller fragmented CDSs. On the other hand, existing CDSs may not be found at all, either because of sequencing errors or because of limitations of the gene finding software. For eukaryotes, the prediction of CDSs is even more difficult because of the existence of introns. To avoid these problems, alternative methods are required for directly reconstructing metabolic networks from unfinished genome data. Sequencing and annotation are still time and resource consuming. An as early as possible exploitation of the genome data is of importance for functional genome research. In this work, we propose a method to identify coding sequences for proteins (particularly for metabolic enzymes) directly from unannotated low-coverage genomic data for *in silico*reconstruction of the metabolic network. The method is demonstrated with genome data from two organisms. A program combining automatic prediction and function assignment of CDSs with a visualization and comparison of metabolic networks of different organisms is also delivered. Principle of the new method =========================== The principle of the new method is schematically shown in Fig. [1B](#F1){ref-type="fig"}. In comparison to the conventional three-step method (Fig. [1A](#F1){ref-type="fig"}) our method can be called a two-step approach. To avoid the separate step of gene finding in the conventional methods, we propose to reverse the searching relationship between public databases and the query sequence: gene or protein sequences from public databases are taken as queries, while the sequences in the unannotated genome of a given organism are treated as a local database that can be searched using a standalone algorithm of BLAST \[[@B22]\]. This results in the prediction of possible CDSs in the genome and simultaneously their functions. Functional information about these CDSs is then used to reconstruct the metabolic network. Thus, our method can significantly simplify the process of CDS prediction and metabolic network reconstruction. By skipping over the separated steps of gene-finding and function assignment, our method can avoid or relax some of the problems of the traditional methods mentioned above. Results and Discussion ====================== Evaluation of IdentiCS for identifying protein coding sequences from genome sequences of *S. typhimurium*LT2 ------------------------------------------------------------------------------------------------------------ To examine the applicability of the method proposed, the well-annotated genome sequences of *S. typhimurium*LT2 \[[@B23]\] are used as \"standard of truth\" for statistic evaluation. According to the annotation given in the KEGG database, *S. typhimurium*LT2 has 4449 CDSs, out of which 1218 have been annotated with enzyme EC numbers. These CDSs encompass 656 different unique EC numbers. The reliability of CDS prediction by the program IdentiCS is evaluated by using a nucleotide database (KEGG genome) and a combined protein database (SWISS-PROT + TrEMBL + TrEMBL update) separately. The annotated coding sequences or proteins of *S. typhimurium*were filtered out of these databases before sequence alignment. All CDSs having an E-value less than 10^-10^were accepted and submitted for comparison with the KEGG annotation of *S. typhimurium*. The results are summarized in Table [1](#T1){ref-type="table"}. 92.6% and 97.7% (sensitivity) of the CDSs in the original annotation of *S. typhimurium*are identified by using the KEGG genome database and the whole protein database SWISS-PROT and TrEMBL, respectively. The sensitivity on the nucleotide level (91.1% and 98.2% for the two databases respectively) is similar as on the CDS level. These results suggest that the SWISS-PROT-TrEMBL based approach is more preferable than the KEGG genome based approach for our method. It is understood that the combined protein database SWISS-PROT and TrEMBL contains almost all of the known protein sequences available in public databases (including proteins *in silico*translated from nucleotide sequences) while the KEGG database contains only a limited number of sequenced and annotated genomes. The difference becomes more significant if the organism studied is evolutionarily far from any organism whose genome is completely sequenced and annotated. The specificity of the method is about 81--82% on the CDS level and 87.2--94.9% on the nucleotide level for the KEGG genome database and the whole protein database SWISS-PROT and TrEMBL. The moderate specificity on CDS level is due to the relatively high amount of additionally predicted CDSs (false positive). It should be mentioned that all the additionally predicted CDSs have quite strong statistic significance (most of them with an E-value 1E-20 -- 1E-40). These additional CDSs may be missed in the original annotation and could in fact represent good candidates for an improved annotation of the genome. The inconsistence rate by IdentiCS is as low as 0.35% for the KEGG genome database and 0.64% for the SWISS-PROT and TrEMBL protein database, indicating the reliability of our method. In the above mentioned evaluation of the method, the cut-off E-value is less than 1E-10 for the CDS prediction. Further the effects of different scoring parameters (i.e. bits score, E-value and identities) and their cut-off values on the CDS prediction by using the database SWISS-PROT and TrEMBL are examined. (Fig. 2A,2B,2C) show the distribution of the true positive CDSs, false positive CDSs, sensitivity and specificity as function of bits score, E-value and identities respectively. The major part of the false positive CDSs is found in the regions of bits score less than ca. 80 and E-value larger than 1E-15. The specificity increases with the bits score and E-value in a form of a saturation curve while the sensitivity decreases almost linearly. This indicates that the prediction performance can be further optimized by choosing appropriate cut-off parameters. If the CDSs with bits score less than 75 (their corresponding E-values are higher than 1E-15 in most cases) are rejected, the false positive will decrease by 40% (from 987 to 602) while the false negative will increase from 110 to 144 (Table [2](#T2){ref-type="table"}). The prediction specificity increases from 81.5% to 87.7% at the expense of a slight decrease in the sensitivity from 97.5% to 96.8%. The specificity of IdentiCS can be further improved by combining a third criterion, i.e. Identities \> = 25% (Table [2](#T2){ref-type="table"}). Thus, the sensitivity and specificity of IdentiCS for CDS prediction are satisfactory and can be balanced to certain extent by choosing proper scoring parameters. Identification of enzyme-coding sequences in *S. typhimurium*LT2 ---------------------------------------------------------------- For the reconstruction of metabolic network it is desired to know the enzyme-coding sequences, and especially the EC-number containing enzymes in an organism. The possibility to use the EC number-containing subset of the combined protein database SWISS-PROT + TrEMBL to identify enzyme-coding sequences in *S. typhimurium*LT2 and thus to further reduce the computation time for constructing the metabolic network is examined (Table [3](#T3){ref-type="table"}). 1894 of EC-number containing CDSs (EC-CDSs) are identified. Of the 1218 originally annotated EC-CDSs, 98.9% of them are identified with an annotation inconsistence rate of as low as 0.08%. The specificity appears to be relatively low due to the large number of false positives. However, if the prediction of EC-CDSs is compared to all the originally annotated CDSs, the number of true positive EC-CDSs is increased from 1204 to 1813 and the number of false positive EC-CDSs is decreased from 690 to 55, resulting in a specificity of 97.1%. The inconsistence rate still remains at a low level, indicating the prediction and function assignment for the additionally predicted EC-CDSs is correct and their EC numbers are missed in the original annotation. This can helps in reconstructing a more complete metabolic network. The KEGG genomes based prediction is also evaluated for its ability to predict the enzyme-coding sequences. 95.4% of the CDSs originally annotated to have an EC-number are correctly predicted and assigned with EC numbers. This value is slightly lower than the one based on the SWISS-PROT+TrEMBL database. The more complete protein databases are therefore more suitable for EC-CDS identification as well. Identification of enzyme coding sequences with different coverage of genome sequences of *K. pneumoniae* -------------------------------------------------------------------------------------------------------- Both the KEGG genomes and SWISS-PROT-TrEMBL databases are used to identify enzyme-coding sequences from the 3.9-fold and 7.9-fold coverage genome sequences of *K. pneumoniae*. From the 3.9-fold coverage genome, IdentiCS identified 1169 and 1342 EC-CDSs by applying the KEGG genome database and SWISS-PROT-TrEMBL databases, respectively, whereas from the 7.9-fold genome sequences 1158 and 1495 EC-CDSs, respectively. As in the case of *S. typhimurium,*IdentiCS identified 15% to 30% more EC-CDSs with queries from SWISS-PROT-TrEMBL than with queries from KEGG for the two versions of *K. pneumoniae*genome sequences respectively. The number of EC-CDSs identified for *K. pneumoniae*is comparable to that identified for *S. typhimurium*with the respective databases. They are also comparable to the number (1156) of annotated EC-CDSs of *E. coli*based on the KEGG genome database. With the method proposed by Ma and Zeng \[[@B24]\], the structure and evolution distance of the metabolic networks of these three organisms and other 47 bacteria are compared. The metabolic network of *K. pneumoniae*is found to be most similar to those of *E. coli*and *S. typhimurium*(data not shown). Thus, the predicted number of enzyme-encoding sequences for *K. pneumoniae*appears to be reasonable. With the same 3.9-fold coverage genome sequences of *K. pneumoniae*, the method of WIT predicted 2650 EC-CDSs which are twice the number of EC-CDSs in *E. coli*and *S. typhimurium*. The EC-CDSs predicted by WIT are significantly smaller and fragmented, possibly because of the presence of too many errors in the unfinished genome sequences. The fragmentation problem was overcome in our method that leads to a significant reduction in the number of identified EC-CDSs. The less false positive EC-CDSs will further simplify experimental design such as for microarray to examine the metabolic network. A comparison of the unique EC numbers of EC-CDSs identified from the two different versions of genome sequences and by the different approaches reveals that the results of these different combinations share over 80% of common EC numbers (Table [4](#T4){ref-type="table"}). The WIT version contains more EC numbers than other versions, obviously because the criteria used in our approaches allow a region to have only one function or an EC number whereas the method used by WIT allows more. With the 3.9-fold genome sequences both KEGG and SWISS-PROT based methods identified a certain number of EC numbers (44 and 81, respectively) that are not identified by WIT. Interestingly, the two different versions of genome sequences result in very close EC numbers. The KEGG based approach identified only 11 (1.68%) more and the SWISS-PROT-TrEMBL based approach identified merely 57 (7.76%) more by using the 7.9-fold genome sequences than by using the 3.9-fold genome sequences (Table [4](#T4){ref-type="table"}). This indicates that the 3.9-fold coverage genome sequences result in a fairly good estimation of enzyme-coding sequences for the purpose of an *in silico*reconstruction of the metabolic network of *K. pneumoniae*. It would be of interest to examine if this also applies to other organisms or even lower genome coverage. The use of lower genome coverage sequences for studying cellular metabolism will greatly accelerate the exploitation of genome sequencing projects. The EC numbers of *K. pneumoniae*as identified by the combination of 7.9-fold genome sequences and SWISS-PROT and TrEMBL databases are summarized in Table [5](#T5){ref-type="table"} in terms of different functional categories. More than half of the enzymes are involved in the metabolism of carbohydrates, amino acids, cofactors and vitamins. 22.2% -- 47.6% of the enzymes in the different KEGG metabolism categories can be found in *K. pneumoniae*(Table [5](#T5){ref-type="table"}). These values are comparable to the values of several evolutionarily closely related strains such as *Escherichia coli, S. typhimurium*, *S. typhi*, *Pseudomonas aeruginosa*and *Yersinia pestis*. Comparison of IdentiCS and CRITICA for identifying coding sequence from low coverage genome sequences ----------------------------------------------------------------------------------------------------- CRITICA is a well-developed program for the prediction of coding sequences \[[@B15]\]. It combines the comparative analysis of DNA sequences with noncomparative methods (i.e. dicodon bias). We compared CRITICA and IdentiCS for predicting CDSs from the two different versions of *K. pneumoniae*genomic sequences. The comparison is done on the basis of all the CDSs including non-enzyme coding sequences. The results are summarized in Table [6](#T6){ref-type="table"}. From the 3.9-fold coverage genome data, CRITICA predicts 6734 CDSs with a cut-off p-value = -4 suggested by Badger and Olsen \[[@B15]\], while IdentiCS predicted 5650 CDSs (with a cut-off E-value = 1E-10). 94.0% of the CDSs predicted by CRITICA are covered by the prediction of IdentiCS. O, In many cases two or more smaller CDSs predicted by CRITICA are covered by a CDS predicted by IdentiCS, obviously because of the relatively high sequencing errors in the 3.9-fold coverage genome data. CRITICA predicts 29 fusion coding sequences. Since they have similarities to two different functions, function assignment to this kind of fusion CDSs is uncertain. Half of the CRITICA-specific CDSs have p-values between 1E-4 and 1E-10. In comparison, of the 1348 CDSs merely predicted by IdentiCS, all have E-values less than 1E-10, 27% have E-values less than 1E-40; all the CDSs have an identity greater than 20% and 60% have an identity greater than 50%, indicating that the predictions by IdentiCS have a high confidence. From the 7.9-fold coverage genome data, CRITICA predicts 5135 CDSs. This number is much less than the CDS number predicted from the 3.9-fold coverage. This may be explained by the significant decrease of sequence errors in the 7.9-fold genome data. In contrast, CDSs predicted by IdentiCS are only 389 less than that predicted from the 3.9-fold genome data. 93.9% of the CRITICA predictions are covered by the 4512 CDSs predicted by IdentiCS. Only 8 CDSs predicted by CRITICA span two or more CDSs. This shows that the increase of sequence quality increases the precision of the prediction of CRITICA. Again, the IdentiCS-specific predictions have a high confidence: all with E-values less than 1E-10 and amino acid sequence identities greater than 20%, more than 50% with E-values less than 1E-20 and identities greater than 50%. The fact that in some cases fusion CDSs are predicted by CRITICA and in other cases many highly potential coding regions are not predicted as CDSs indicates a shortcoming in this algorithm for low quality contigs. When CRITICA finds a coding region with a high score, it tries to find the start and stop codons by extending this region to both upstream and downstream with the conditions of not decreasing the total score after extension. Sequencing errors, especially translation shifts, make it difficult for CRITICA to calculate the extension score correctly. In such cases, the algorithm used by IdentiCS does not need to locate the start and stop codons. Transcription frame shifts also have less interference to IdentiCS because it does not use predicted coding sequence as queries but uses entries from public database to search for coding sequences in the raw genome sequences of an organism. These features make IdentiCS more suitable for identifying possible protein-coding regions from low-coverage error-containing raw genome sequences than other available approaches. Reconstruction and visualization of metabolic networks for comparison --------------------------------------------------------------------- With the identified enzyme-encoding sequences discussed above the potential metabolic networks of *S. typhimurium*and *K. pneumoniae*can be reconstructed and compared to other organisms. The reconstruction of metabolic networks can be done in a similar way as based on CDSs from annotated genome sequences as recently described by Ma and Zeng \[[@B1]\]. Briefly, from the identified EC numbers of CDSs, the set of biochemical reactions involved in the organism can be established with the help of a reaction database (i.e. a revised version of LIGAND \[[@B5]\] or BRENDA \[[@B25]\]). From the reaction set, a connection matrix is obtained that can be used to represent the metabolic network as a directed graph for computational analysis. For a straightforward visualization of the biochemical reactions related to a specific organism and especially for comparing the metabolisms of different organisms, the KEGG metabolic maps act as a blueprint for visualization of metabolic networks in this work. Reconstruction, comparison and visualization of the metabolic network have been integrated in the program IdentiCS as a built-in component. It can work directly with the coding sequences and their functions predicted by IdentiCS or with existing annotation files such as the those in Microsoft Excel format or in GenBank flat file format. To use the KEGG maps for metabolic reconstruction, comparison and visualization information from the KEGG metabolic pathways is reformed into an Excel template. Metabolic network reconstruction is realized by mapping the identified enzymes (EC numbers) to the KEGG metabolic maps. Different similarity levels of the identified enzymes can be displayed in different colors. All identified enzymes in the map are marked and linked to their annotations. Web links to other Internet databases such as IUBMB \[[@B26]\], BRENDA \[[@B25]\], WIT \[[@B3]\], KEGG \[[@B5]\], Ecocyc \[[@B9],[@B10]\] and SWISS-PROT \[[@B27]\] are also integrated to offer the user a fast tool to access the relevant information. For metabolic network comparison, strain-specific colored boxes are drawn on the lower part of the EC number rectangle if that strain possesses this enzyme as demonstrated in Fig. [3](#F3){ref-type="fig"} for the glycolysis pathways. This box is linked to the original annotation of the enzyme in that strain. The background of the enzyme box becomes green if this enzyme is found in all the compared organisms. In such a way an informative metabolic network is generated which can serve as a starting point for functional and comparative analysis of the metabolism of organisms under study. Conclusions =========== The use of genome sequences from *S. typhimurium*and *K. pneumoniae*demonstrated the applicability and reliability of the new method proposed for *in silico*identification of protein coding sequences from unannotated genome sequences. The use of protein sequence databases SWISS-PROT and TrEMBL is more favorable than the use of KEGG genome database for identifying coding sequences and thus for metabolic network reconstruction. Furthermore, the method allows an adequate reconstruction of the potential metabolic network from sequence data with low coverage (e.g. \< 4 fold) of the bacterial genome as shown for *K. pneumoniae*. Together with the algorithms for the automatic annotation of sequences, the visualization and comparison of metabolic networks, the method and program developed in this work can accelerate the use of genomic data for studying cellular metabolism. Methods ======= Database preparation -------------------- The applicability of the method proposed above was examined with the genome sequences of two organisms, namely *Salmonella typhimurium*LT2 and *Klebsiella pneumoniae*. The genome of *S. typhimurium*LT2 has been completely sequenced and well annotated \[[@B23]\]. Thus, the annotated genome sequences of *S. typhimurium*LT2 serve as a reference to evaluate the accuracy of the proposed method. The sequences and annotation for *S. typhimurium*LT2 were downloaded from KEGG \[[@B28]\](version of Dec. 18. 2003). The genome of *K. pneumoniae*has been recently sequenced and the annotation is still in progress. Two different versions of the raw genome data of *K. pneumoniae*(3.9-fold whole genome shotgun coverage in 920 contigs and 7.9-fold coverage in 341 contigs) obtained from the Genome Sequencing Center of Washington University \[[@B29]\] were examined in this study. Each version of the raw genome data was formatted as a local database for BLAST \[[@B22]\]. Two types of databases are used in this work for the prediction and function assignment of CDSs for a given organism, namely the nucleic acid database from KEGG and the non-redundant protein sequence databases from SWISS-PROT, TrEMBL and TrEMBL updates. The reason to choose the genome database from KEGG as query but not from other nucleic acid databases such as GenBank or EMBL is that KEGG contains the most extensive EC numbers for enzymes that are needed for reconstructing metabolic networks. Therefore, the genome database of KEGG version can serve as an EC number source and be used for the purpose of comparative analysis of genome-based metabolism. In contrast, the flat data files from GenBank and EMBL do not contain the necessary enzyme index information in many cases. SWISS-PROT is human-curated and therefore more preferred. SWISS-PROT and its sister database TrEMBL (SWISS-PROT Release 42.7, TrEMBL Release 25.7, released on 15 Dec. 2003) were obtained from the Swiss Institute of Bioinformatics \[[@B30]\]. Not \"fasta\" format files but SWISS-PROT flat files were used because the enzyme EC numbers may not be included in the fasta format files available on the FTP site. Entries in the databases that do not contain EC numbers can be filtered out before the sequence alignment step to shorten the computational time if the purpose is merely to identify metabolic enzymes and to reconstruct the metabolic network. For identifying all possible CDSs, the complete SWISS-PROT and TrEMBL databases are used. Automatic prediction and annotation of protein-coding sequences --------------------------------------------------------------- The annotation process is based on similarity comparison as normally used in other annotation processes. The difference is that in our approach the gene or protein sequences from public databases are used as queries to search and locate similar ones in the raw genome sequences. When proteins from public database are used as queries, the tblastn algorithm in the BLAST program is applied that compares the query to all six translation frames of the unannotated DNA sequences. The dynamic translation of a small genomic database takes much less system resource than the translation of a large public database as in the conventional methods. Our method can thus be realized on a common PC system, especially when merely a subset of the public database is considered, for example for the purpose of identifying metabolic enzymes for metabolic network reconstruction. Because of sequence errors (especially the translation shift and abnormal stop codon) and the local alignment nature of the BLAST algorithm, the BLAST research may report several small alignments between different parts of the query protein or gene and different parts of a genomic contig even if there should be only one alignment in the reality. In this situation, the tfasty34 program in the FASTA3 suite \[[@B31]\] should give a better alignment since the translation shift is considered. But the tfasty34 program runs very slowly in our test and is therefore not used here for large-scale genomic alignment. In this work, fragment(s) of a genomic contig are joined to the genomic fragment that has the highest alignment score, resulting in a larger CDS fragment if: 1\. they are coded on the same strand of the same genomic contig as the highest score fragment. In other words, all of these fragments must be translated either in positive or in negative frames. 2\. the alignments have an identity level not lower than 80% of the identity level of the highest score alignment. 3\. the generated larger sequence region has alignment gaps or extensions not more than 20% of its length. Since many queries can be similar to the same region on a genomic contig and sometimes they may have different function annotations, the program must judge and choose one annotation for this region. The decision is made by applying the following criteria: 1\. Each region normally has only one function. Here the region represents a piece of nucleotides either on the positive strand or on the negative strand of a DNA molecule. The same physical position on different strands of a DNA molecule can belong to different regions, and can therefore have a different function assignment. Although there are examples in some viruses that a region can code different proteins depending on the transcription frame, it happens very rarely in other organisms. The user can assign a tolerance value (e.g. 60 bp) to allow two successive regions to overlap each other to some extent. 2\. Highest similarity principle. If a query gene or protein has a similarity to a CDS higher than other queries, then the function of this query gene or protein is assigned as the annotation of the CDS. Bits score is used as a measure for similarity first. If two queries have the same bits score, then the identity level in percentage is taken as a second measure for similarity. If both bits score and identity are the same and these two entries have different function annotation (rarely occurred) then both of their functions are assigned to that region. 3\. Closest evolutionary relationship. If two or more query genes or proteins are comparably similar (e.g. the difference between their identity levels is lower than 5%) to a CDS but have different function, the evolutionary relationship between these organisms is further considered. The annotation of the organism that is mostly related to the studied organism from the viewpoint of metabolic evolution is transferred to the unknown CDS. The evolutionary relationship between different organisms and the one studied is established with the method of Ma and Zeng \[[@B24]\] after the initial function assignment for the CDSs with the highest similarity criteria. In this way, the coding sequences of a genome are identified and annotated at the same time. No second large-scale sequence alignment is needed. Once all the software and databases are prepared, our program which is called IdentiCS ([Identi]{.underline}fication of [C]{.underline}oding [S]{.underline}equences from Raw Genome Sequences) can reconstruct the metabolic network of an organism with about 5 million base pairs of raw genome data. The computing time is less than 8 hours on a PC with 2.8 GHz Pentium 4 CPU and 512 MB memory. This program works together with Microsoft Excel under Windows environment. Statistic evaluation -------------------- For a more detailed examination of our method, the results are evaluated separately for the prediction of CDSs and their function assignment, although our method integrates these two aspects into one step. The terms true positive (TP), false negative (FN) and false positive (FP) are used to calculate the sensitivity and specificity of CDS prediction in comparison with CDSs in the original annotation. The terms \"sensitivity\" and \"specificity\" are defined according to Burset and Guigo \[[@B32]\]:  We also evaluated the terms TP, FN and FP on nucleotide level according to Burset and Guigo \[[@B32]\] and calculated the corresponding sensitivity and specificity as above. It should be mentioned that a true positive CDS does not necessarily mean that its function assignment is also correct. The terms consistence and inconsistence are used to describe whether a true positive CDS has the same function assignment as in the original annotation or not. Correspondingly, an \"inconsistence rate\" is used and defined as:  Authors\' contributions ======================= JS is the developer of the method and the program. APZ supervised the study. Both authors read and approved the final manuscript. Acknowledgements ================ One of the authors (J. Sun) greatly acknowledges the PhD Scholarship of German Academic Exchange Service (DAAD). This work is supported by grant 031U110A/031U210A from Bundesministerium für Bildung und Forschung (BMBF) of Germany. We also thank Dr. H. Ma for his help in the structure and evolution analysis for the metabolic network of *K. pneumoniae*. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Schematic illustration of methods for reconstructing the metabolic network from genome data: (A) conventional three-step method based on coding sequence (CDS) prediction, function assignment and metabolic reconstruction; (B) proposed two-step method based on a simultaneous and direct identification of coding sequences and their functions from raw genome sequences. Note that the query process has been reversed in the new method in comparison to the conventional method. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Distribution of false positive, true positive, sensitivity and specificity as functions of bits score (Fig. 2A), E-value (Fig. 2B) and identity (Fig. 2C). The bars for the distribution of false positive and true positive indicate the corresponding number of CDSs in a given interval of the statistic parameter (5 for bits score, 1 for -log(E-value) and 0.01 for identity) divided by the number of all false positive or true positive CDSs submitted for analysis. The lines show the specificity and sensitivity change along the cut-off scoring values. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Glycolysis pathways as example for demonstration of metabolic comparison among different organisms. KPN: *K. pneumoniae*MGH78578, ECO: *Escherichia coli*K-12 MG1655, STM: *Salmonella typhimurium*LT2, STY: *Salmonella typhi*, PAE: *Pseudomonas aeruginosa*PA01, YPE: *Yersinia pestis*strain CO92. Green background with enzyme EC number means that this enzyme exists in all the compared organisms. Colored rectangles under the EC number box represent links to the strain-specific annotation of the corresponding enzyme. Colored bars above the EC number link the enzyme to the corresponding entry in different public databases. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Evaluation of the method IdentiCS for identification of CDSs in the genome of *Salmonella typhimurium*. KEGG: KEGG genome database; SW: SWISS-PROT + TrEMBL + TrEMBL updates. ::: Database KEGG SW ------------------------------ ------------- ------------- True positive 4121 4339 False positive 907 987 False negative 328 110 Sensitivity %\* 92.6 (91.1) 97.7 (98.2) Specificity %\* 82.0 (94.9) 81.5 (87.2) Inconsistence rate (%) in TP 0.35 0.64 \*: Values shown in parentheses are calculated based on the nucleotide level according to Burset and Guigo\[32\] ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Effects of different scoring criteria on CDS identification in the genome of *S. typhimurium*using IdentiCS and the database SWISS-PROT and TrEMBL. Criteria 1: E-value \< = E-10; Criteria 2: E-value \< = E-10 and Bits score \> = 75; Criteria 3: E-value \< = E-10, Bits score \> = 75 and Identities \> = 25% ::: Criteria 1 Criteria 2 Criteria 3 ---------------- ------------ ------------ ------------ True positive 4339 4305 4303 False positive 987 602 555 False negative 110 144 146 Sensitivity % 97.5 96.8 96.7 Specificity % 81.5 87.7 88.6 ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Evaluation of the performance of IdentiCS for the prediction of EC number -containing CDSs (EC-CDSs) with the EC-number containing subset of the protein database SWISS-PROT and TrEMBL. ::: Compared to originally annotated EC-CDSs Compared to all originally annotated CDSs ---------------------------- ------------------------------------------ ------------------------------------------- EC-CDSs predicted 1894 1894 True positive 1204 1813 False positive 690 55 False negative 14 NA Sensitivity 98.9% NA Specificity 63.6% 97.1% Inconsistence rate in T.P. 0.08% 3.32% NA: not applicable. ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Comparison of EC numbers identified with different methods and different versions of the genome sequence of *Klebsiella pneumoniae*. WIT: WIT version of annotation by gene prediction from the 3.9-fold genome sequences; KEGG3.9 and KEGG7.9: annotations of the 3.9-fold and 7.9-fold genome sequences by applying the KEGG genome database. SW3.9 and SW7.9: annotations of the 3.9-fold and 7.9-fold genome sequences by applying SWISS-PROT and TrEMBL protein databases. ::: Annotation version Number of unique ECs identified Version-specific EC numbers\* compared to: -------------------- --------------------------------- -------------------------------------------- ----- ----- ----- ----- WIT 764 \- 162 159 152 131 KEGG3.9 646 44 \- 4 63 56 KEGG7.9 653 48 11 \- 68 57 SW3.9 693 81 110 108 \- 15 SW7.9 735 102 145 139 57 \- \*The version-X-specific EC number compared to the version Y is referred to EC numbers that are only found in the annotation Version X but not in Y. For example, the KEGG3.9 has only 4 version-specific EC numbers compared to KEGG7.9. On the other side, KEGG7.9 has 11 version-specific EC numbers compared to KEGG3.9. ::: ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Distribution of the unique EC numbers of *K. pneumoniae*in different function categories compared to other organisms (see legend of Fig. 3 for name abbreviations of organisms). The EC numbers for *K. pneumoniae*were identified from the unannotated 7.9-fold coverage genome sequences by SWISS-PROT-TrEMBL based IdentiCS. The EC numbers for other organisms are taken from the KEGG genome annotations. The total number of strain-specific ECs is shown in parenthesis under the strain name. ::: ------------------------------------------------------------------------------------ Function Categories^a^ ECs^b^ KPN\ ECO\ STM\ STY\ PAO\ YPE\ (735) (681) (655) (445) (573) (585) --------------------------- -------- ------- ------- ------- ------- ------- ------- Carbohydrate metabolism 427 151 154 147 101 105 123 Energy metabolism 167 60 62 62 46 61 58 Lipid metabolism 126 27 25 25 17 27 17 Nucleotide metabolism 166 83 82 80 45 66 72 Amino Acid metabolism 561 211 189 189 132 205 183 Other Amino Acids 146 49 45 43 26 47 41 Complex Carbohydrates 184 63 58 55 33 43 53 Complex Lipids metabolism 171 46 38 34 24 31 34 Cofactors and Vitamins 225 104 105 107 74 100 96 Sum of unique EC numbers 1899 567 537 529 349 464 476 ------------------------------------------------------------------------------------ ^a^: Functional categorization according to KEGG; ^b^: unique EC numbers of corresponding metabolism category calculated from KEGG metabolic pathway maps. ::: ::: {#T6 .table-wrap} Table 6 ::: {.caption} ###### Comparison of coding sequences (CDSs) prediction by IdentiCS and CRITICA from unfinished genome sequences of *K. pneumoniae*with different genome sequence coverage. ::: 3.9 × genome data 7.9 × genome data -------------------------------------------------- ------------------- ------------------- ------ ------ Number of all CDSs 6734 5650 5135 5261 CDSs shared by both programs 6332\* 4302 4823 4512 CDSs merely identified by the respective program 402 1348 312 749 \*CDSs predicted by IdentiCS can cover more than one smaller CDSs predicted by CRITICA. :::

PubMed Central

2024-06-05T03:55:47.187392

2004-8-16

PMC514701

Background ========== The genome projects have increased our knowledge of genomic sequences for several organisms. Taking advantage of this knowledge, the microarrays technologies allow the characterization of a whole-genome expression by showing the relative transcript levels of thousand of genes in one experiment \[[@B1]\]. Numerous applications were developed apart gene expression analysis like single nucleotide polymorphism (SNP) and genotyping \[[@B2],[@B3]\], diagnosis \[[@B4]\] and comparative genomics \[[@B5]\] analysis. Particularly, the transcriptome analysis provides insight into gene regulations and functions. To help the characterization of relevant information in microarray data, specific computational tools are needed. The identification of co-expressed genes is commonly performed with unsupervised approaches, such as clustering methods with the Hierarchical Clustering (HC) \[[@B6]\], the *k*-means \[[@B7]\] and the Self-Organizing Map (SOM) \[[@B8],[@B9]\], or with projection methods such as the Principal Component Analysis (PCA) \[[@B10]\] and Independent Component Analysis (ICA) \[[@B11]\]. Among these techniques, the HC approach is a widely held method to group genes sharing similar expression levels under different experimental conditions \[[@B12]\]. The HC is performed from the distance matrix between genes computed from the microarray data, *i.e*. the gene expression levels for various experimental conditions. Different aggregation methods can be used for the construction of the dendogram generally leading to different tree topologies and *a fortiori*to various cluster definitions \[[@B13]\]. For instance, the single-linkage algorithm is based to the concept of joining the two closest objects (*i.e*. genes) of two clusters to create a new cluster. Thus the single-linkage clusters contain numerous members and are branched in high-dimensional space. The resulting clusters are affected by the chaining phenomenon (*i.e*. the observations are added to the tail of the biggest cluster). In the complete-linkage algorithm, the distance between clusters is defined as the distance between the most distant pair of objects (*i.e*. genes). This method gives compact clusters. The average-linkage algorithm is based on the mean similarity of the observations to all the members of the cluster. Yeung and co-workers \[[@B14]\] showed that single-linkage hierarchical clustering is inappropriate to analyze microarray data. Gibbons and Roth \[[@B15]\] showed by using gene ontology that single-and average-linkage algorithms produce worse results than random. In addition these authors conclude that the complete-linkage method is the only appropriate HC method to analyze microarrays experiments. The microarrays experiments frequently contain some missing values. The missing values are part of the experimental errors due to the spotting conditions (e.g. spotting buffer, temperature, relative humidity\...) and hybridization (e.g. dust on the slide\...) \[[@B16],[@B17]\]. The users commonly discard suspicious dots during the images analysis step. Thus, the resulting data matrix contains missing values (MVs) which may disturb the gene clustering obtained by the classical clustering methods, e.g. HC, SOM, or projection methods, e.g. PCA. To limit the effects of MVs in the clustering analyses, different strategies have been proposed: (i) the genes containing MVs are removed, (ii) the MVs are replaced by a constant (usually zero), or (iii) the MVs are re-estimated on the basis of the whole gene expression data. Few estimation techniques have been applied to such data. The *k*-nearest neighbors approach (*kNN*) computes the estimated value from the *k*closest expression profiles among the dataset. Troyanskaya and co-workers showed that the weighted *kNN*approach, with *k*= 15, is the most accurate method to estimate MVs in microarray data compared to replacement by zero, row average, or Singular Value Decomposition \[[@B18]\]. A recent work proposes Bayesian principal component analysis to deal with MVs \[[@B19]\]. In the same way, Zhou and co-workers \[[@B20]\] have used a Bayesian gene selection to estimate the MVs with linear and non-linear regression. However, the *kNN*approach is the most popular approach for estimating the MVs. To explore the incidence of MVs in gene clustering, we first assessed the proportion of MVs in different sets of public data from *Saccharomyces cerevisiae*and human. Observing that MVs are widely present in the expression data, we then analyzed the effects of MVs on the results of a hierarchical clustering (HC) according to the chosen clustering algorithm. In the same way, we evaluated the impact of MVs replacement and estimation in the gene cluster definition by using a hierarchical clustering method. Results and Discussion ====================== Missing values overview ----------------------- Table [1](#T1){ref-type="table"} summarizes the proportions of MVs in eight series of microarray experiments \[[@B21]-[@B28]\]. The number of genes and experimental conditions are within the range (*552*; *16523*) and (*4*; *178*) respectively. The percentage of MVs varies from 0.8% to 10.6%. No relation had been found between the number of genes and the percentage of MVs or the percentage of genes without MVs. As expected, the percentage of genes with MVs increases in function of the number of experimental conditions. In the Sorlie \[[@B22]\], Spellman \[[@B28]\] and Gasch sets \[[@B25]\] respectively, 94.7%, 91.8% and 87.7% of the genes profiles have MVs. In these cases, it is not possible to systematically delete the genes profiles with MVs in the data analysis. Indeed, the percentage of MVs is not negligible in the microarrays data and may be a strong factor of gene clustering instability. Hence, we evaluated in this study the effects of MVs on gene group definition by using hierarchical clustering algorithms. Main steps of the analysis -------------------------- Figure [1](#F1){ref-type="fig"} describes the four steps of the analysis. (i) *Definition of a complete datafile*: We selected from literature a set of microarray data. This set was filtered to retain only genes without MVs. (ii) *Construction of reference gene clusters*: A gene clustering was performed by different HC algorithms (*algo*). Seven types of algorithms were used (see section Methods). For each analysis, we defined a number *K*^*algo*^of gene clusters, representing the reference clusters. (iii) *Generation of a datafile with MVs*: MVs were randomly inserted with a fixed rate *τ*. This rate correspond to the percentage of genes with 1 missing value. So, we defined data with MVs. (iv) *Analysis of the newly generated expression data:*We carry out a data processing similar to step (ii), with the construction of dendograms for each algorithm (*algo*) and the definition of the *K*^*algo*^new clusters in the different experiments (no replacement of MVs was carried out). Finally, we assessed the stability of the gene clustering by calculating an index measuring the percentage of conserved genes between the clusters of the reference set and the generated set. Moreover, two other experiments were (v) designed and (vi) evaluated: clustering with MVs estimated by the *kNN*approach, and clustering with MVs replaced by zero. Experimental sets ----------------- As we work on the impact of MVs in gene clustering, we need at first biological datasets without MVs. These sets have been extracted from Ogawa set \[[@B26]\] (noted OS) and Gasch set \[[@B25]\] (noted GS). OS and GS have been chosen because they contain few MVs and, after filtering the number of genes remains important, (*ca.*6000). The original Ogawa set contained 6013 genes with 230 genes having MVs. The elimination of the genes with MVs (i.e. 3.8% of the genes) leads to a set with 5783 genes. For the GS, the number of MVs is more important and some experimental conditions have more than 50% of MVs. So we have limited the final number of selected experimental conditions from 178 to 42 (see section Methods), it allows to conserve 5843 genes, i.e. only 310 genes are not analyzed, representing 5.0% of all the genes. Moreover, we have defined two smaller sets, GS~H2O2~and GS~HEAT~, from GS corresponding respectively to H~2~O~2~and heat shock experimental conditions. To assess the influence of size of the datasets and the number of observations (genes), we have generated smaller sets corresponding to a ratio 1/*n*(*n*= 2, 3, \..., 7) of the initial OS, GS, GS~H2O2~and GS~HEAT~gene content (see section Methods). Example of clustering disturbance caused by missing values introduction ----------------------------------------------------------------------- Figure [2](#F2){ref-type="fig"} gives an example of significant clustering disturbance caused by the MVs. Figures [2a](#F2){ref-type="fig"} and [2b](#F2){ref-type="fig"} show the dendograms obtained with the complete-linkage hierarchical clustering of the gene set before and after introducing 1% of MVs. Surprisingly, the genes belonging to one cluster in the reference dataset are reallocated in several clusters after this slight data transformation. Number of clusters ------------------ To perform the comparison of the gene clustering by HC, the numbers of clusters according to every type of hierarchical clustering algorithms have to be defined. Hence we defined the number of clusters,*K*^*algo*^, for each clustering algorithm (*algo*). The rule consists in determining *K*^*algo*^clusters as the 10 most important clusters correspond to 80% of the genes of the dataset (see section Methods). The results for OS are shown in Table [2](#T2){ref-type="table"} (column 1) for each clustering methods. As expected, we observe a correlation with the type of algorithm used. The number of clusters is lower for the well-balanced tree generally obtained by the Ward and complete-linkage methods, e.g. 19 and 36 clusters respectively, compared to those providing by the other methods. For instance, the single-and centroid-linkage methods lead to the definition of 175 clusters. Index \"Conserved Pairs Proportion (CPP)\" ------------------------------------------ The clusters defined from the reference data and the data with MVs are compared using *Conserved Pairs Proportion (CPP) index*which corresponds to the percentage of genes found associated in the reference clusters and found again associated in the clusters generated from the data with MVs. Figure [3](#F3){ref-type="fig"} summarizes the results about the influence of MVs on the hierarchical clustering. The given results were computed on the (1/7) subset from OS using the seven aggregative algorithms. The metric used is the Euclidean distance. We observe that (i) the single-and the centroid-linkage methods show a low *CPP*decrease, the *CPP*values are always greater than 95%, (ii) the average-and median-linkage methods are within the range \[65%; 80%\] and (iii) the mcquitty, Ward and complete-linkage methods show the most striking loss. We observe a drastic loss of the clustering stability since *τ*= 1% of MVs. For instance, with 5% of genes with MVs, *i.e*. 40 missing data, the mcquitty, Ward and complete-linkage methods have a *CPP*of 62%, 57% and 52%, respectively. Beyond a rate of *τ*equal to 10%, the decrease becomes lower. Similar results are observed for all the sets OS, GS, GS~H2O2~and GS~HEAT~and all the generated sets from 1/2 to 1/7. These last results show that the quality of the gene clustering is not disturbed by the reduction of the number of genes. It must be clearly noted that to limit the effect of the topology of each algorithm, we have fixed that the 10 most populated clusters must represent 80% of the genes. The number of the most populated clusters is fixed at 10 due to the Ward linkage method that gives a very limited number of clusters. Then, the percentage of genes belonging to the 10 most populated clusters have been tested ranging from 70% to 90% (data not shown). For example with a percentage equal to 90%, the *CPP*values of single-and the centroid-linkage methods remain too stable to observe a clear decrease as seen in Figure [3](#F3){ref-type="fig"}. The choice of 80% allows to analyze the precise decrease of the different *CPP*values and to compare the different aggregation methods. *k*-Nearest Neighbor (*kNN*) ---------------------------- The *kNN*method has been described by Troyanskaya and co-workers \[[@B18]\] for the MVs in microarray data. The *kNN*approach goal is to compute the expected value of a missing value from the *k*nearest vectors without a missing value. As no theoretical approach exists to define the optimal *k*values (*k*~*opt*~), we have assessed every value of *k*within the range 1 to 100 and, selected the *k*~*opt*~value as the *k*value which has the minimal error rate. Table [3](#T3){ref-type="table"} shows the *k*~*opt*~values obtained for the four sets used in this study and their corresponding error rates. The *k*~*opt*~values are lower in OS subsets showing a low number of genes. The *k*~*opt*~values of GS~H2O2~are within the range 11 to 17. The GS and GS~HEAT~sets exhibit more important variations within the range 8 to 28. The error rate decreases slightly according to the number of genes, but these variations are not significant. Nevertheless, this poor correlation may be due to the subsets composition. Indeed, they keep an equivalent number of clusters with a smaller number of genes per cluster. In the same way, it may simply be due to the *k*~*opt*~variation. Figure [4](#F4){ref-type="fig"} shows that the *kNN*method gives worse prediction of the extreme values than the values close to zero. The real data distribution follows approximately a normal distribution and the *kNN*approximation leads to a reduction of the standard deviation of this distribution. So, the prediction of the extreme values increases the global error rate implying a higher *k*~*opt*~to reduce this effect. For instance, in the OS sets, we observe that the values within the range \[-1.0; 1.0\] are approximated with a mean error rate within the range \[0.12; 0.14\]. Conversely the values more than 1.5 or less than -1.5 are approximated with a mean error rate superior to 1.8. The misestimating of extreme values has an impact on the clustering. One can notice that the unweighted *kNN*(mean of the *k*observations) exhibits worse results compared to the weighted *kNN*used in this study (data not shown). Improvements of CPP with *kNN*approach and zero-value replacements ------------------------------------------------------------------ The *CPP*was computed for the seven agglomeration methods. We have compared the HC results obtained with the reference sets and the generated sets without replacement of MVs, with *kNN*replacement or with *zero-value*replacement methods. Table [2a](#T2){ref-type="table"} shows that the *kNN*and *zero-value*replacements both improved the mean *CPP*whatever the clustering method used, except for the Ward method with the zero value replacement. The *kNN*approach is the most relevant method to replace MVs. In 55.2% to 66.3% of the simulations, *kNN*is better and globally gives a mean increase of the *CPP*within the range \[0.7; 2.1\]. The centroid-and single-linkage methods have better increase in 74.8% and 99.3% of the simulations respectively due to their particular topologies. The zero value replacement is clearly less efficient. Nevertheless, as a slight variation can displace one gene into a close cluster, we have characterized another index named *CPP*~*f*~to consider the *f*closest clusters of the selected cluster. This index is similar to the previous one and takes into account that the genes may be relocated in close clusters (see section Methods). It allows the evaluation of the topology conservation. We used *f*= 5. We observed that the co-associated genes in the reference sets are often displaced to close clusters in the simulated sets. As observed for the *CPP*, the *kNN*approach improves the *CPP*~*f*~for all the clustering methods in 51.9% to 60.2% of the simulations within the range \[0.2; 1.5\]. Due to their high initial *CPP*values (97.7% and 98.8%), single-and centroid-linkage methods do not have a gain as previously observed for the *CPP*. Similar results are obtained with the other sets with slight variations. For example, GS~H2O2~have *CPP*and *CPP*~*f*~close to the one of OS. Conversely, the *CPP*and *CPP*~*f*~of GS~HEAT~are better than the ones of OS and GS~H2O2~for the complete-linkage and Ward methods, but lower for the others. The GS set has higher *CPP*and *CPP*~*f*~for single-likage to mcquitty method due to a lower influence of the MVs in a vector with a higher number of experimental conditions. Nevertheless, the complete-linkage and Ward linkage still remain at a very low *CPP*(close to 50%). Moreover, we have tested the influence of the number of MVs per gene by introducing more than one missing value per gene. We obtained similar results showing less than 0.2% of variations of the *CPP*values. In addition we have tested the consequences of using *k*values different of *k*~*opt*~values in the range \[*k*~*opt*~-10 ; *k*~*opt*~+10\] and we observed a decrease of *CPP*within the range \[1%; 5%\] (data not shown). Extreme values -------------- We have followed the same methodology to analyze the extreme values, *i.e*. values superior to 1.5 and lower than -1.5. Table [2b](#T2){ref-type="table"} summarizes the results of the OS (1/7). The *CPP*values are superior to the *CPP*values obtained previously, because only the genes with important variations have MVs and are members of small clusters. We observe that MVs replacement has little effect. Indeed, the *kNN*and zero -- value replacement cannot restore a correct distribution (cf. Figure [4](#F4){ref-type="fig"}). However, the *CPP*~*f*~shows that the *kNN*is better than the replacement by zero, allowing a better topology preservation. Same results are observed for the other sets (data not shown). Conclusions =========== MVs are a common trait of microarrays experiments. Few works had been reported about MVs replacements \[[@B18]-[@B20]\] and none analyse their influence in the clustering of microarrays data. In our study, we showed that MVs significantly biased the hierarchical clustering. In addition, we observed that the effects of MVs are correlated to the chosen clustering method. The single linkage-method is the most stable due to the building of cluster of large size and numerous small clusters and singletons. At the opposite, the Ward and complete-linkage methods create well distributed population of clusters inducing a higher sensitivity to MVs. The topology of the dendogram is highly disturbed by transferring genes in distant clusters. We showed that the *kNN*replacement method was the most efficient approach to compensate the MVs effects compared to the classical replacement by zero. The *k*~*opt*~depends on the sample size. It is important to keep in mind that the MVs corresponding to extreme values are difficult to estimate with the *kNN*method. The impact of their approximation upon the clustering is significant. Hence, new approaches like the Bayesian Principal Component Analysis (BPCA) may overcome this problem. In a recent work Liu and co-workers suggest to potentially eliminate the incomplete series of data by using robust Singular Values Decomposition \[[@B29]\]. In addition, our work showed clearly the need of evaluation of the data quality and statistical measurements as noted by Tilstone \[[@B30]\]. Contrary to Yeung and co-workers \[[@B14]\] and Gibbons and Roth \[[@B15]\], we have defined for each type of hierarchical clustering algorithm (*algo*) a specific number of clusters (*K*^*algo*^). This point is one of the main difficulties noted by Yeung and co-workers \[[@B31]\] to evaluate the clustering methods as the topology generated are different. The comparison of the different aggregative clustering algorithms remains constrained by the topology (e.g. 175 clusters defined in our study for single-linkage compared to 19 clusters for Ward method). All these results are in accordance with the results of Nikkilä and co-workers \[[@B32]\] which show a hieratic problem of topology preservation in hierarchical clustering. Recent methods like SOTA \[[@B33],[@B34]\] or Growing SOM (Self-Organizing Maps) \[[@B35]\] have combined a hierarchical clustering visualization with the preservation of the topology allowed by the SOM. Our future works will address the definition of a most robust clustering method. Methods ======= Data sets --------- We used 8 public data sets from the SMD database (\[[@B36]\]; see Table [1](#T1){ref-type="table"}). Two sets were used for a thorough analysis. The first one (Ogawa set) was initially composed of *N*= 6013 genes and *n*= 8 experimental conditions about the phosphate accumulation and the polyphosphate metabolism of the yeast *Saccharomyces cerevisiae*\[[@B26]\]. The second one corresponds to various environmental stress responses in *S. cerevisiae*\[[@B25]\]. This set (Gasch set) contains *N*= 6153 genes and *n*= 178 experimental conditions. Due to the diversity of conditions in this set, we focused on two experimental subsets corresponding to heat shock and H2O2 osmotic shock respectively. Data sets refinement: missing values enumeration ------------------------------------------------ To evaluate the incidence of MVs on hierarchical clustering, we built complete datasets without MVs. All the genes containing at least one missing value were eliminated from the Ogawa set (noted OS). The resulting OS set contains *N*= 5783 genes and *n*= 8 experimental conditions. The second set without MVs was taken from Gasch *et al.*and called GS. The experimental conditions (column) containing more than 80 MVs were removed. The resulting GS matrix contains *N*= 5843 genes and *n*= 42 experimental conditions. Two subsets were generated from GS and has been noted GS~HEAT~and GS~H2O2~. They correspond to specific stress conditions as described previously. GS~HEAT~and GS~H2O2~contain respectively *N*= 3643 genes with *n*= 8 experimental conditions and *N*= 5007 genes with *n*= 10 experimental conditions. To test the influence of the matrix size, *i.e*. the number of genes, we built six smaller sets corresponding to 1/2, 1/3, 1/4, 1/5, 1/6 and 1/7 of OS, GS, GS~HEAT~and GS~H2O2~. To obtain representative subsets, we can not use a random generation which can bias the results. So, we searched for each subset the series of genes which reflect at best all the genes of the complete set. First, the distance matrices between all the genes were computed. Then, we performed an iterative process by: (i) computing the sum of the distances for each possible *t-uplets*(*t=*2 to 7) of the set, (ii) choosing the *t*genes which have the minimal distance, (iii) selecting 1 representative gene upon the *t*selected genes, this gene is chosen as the closest to the barycenter of the cluster, (iv) eliminating from the process the *t*genes. Step (i) to step (iv) are repeated until all the genes are used. All the representative genes constitute the subset (1/*t*). This procedure allows one to reduce the redundancy of similar genes and to maintain approximately a number of gene clusters constant. Missing values generation ------------------------- From the sets without MVs, we introduced a rate *τ*of genes containing one MV (*τ*= 1 to 50.0%), these MVs are randomly drawn. Each random simulation is generated at least 100 times per experiment to ensure a correct sampling. Replacement of MVs by the *kNN*method ------------------------------------- To fulfill *v*~*i*~, a missing value *i*for a given expression vector **v**(*i.e*. a gene), with the *kNN*method, the *k*vectors **w**corresponding to the *k*most nearest vectors to **v**(without taking the *i*th elements of the **w**-vectors into account) are searched. The missing value *v*~*i*~is then estimated by a weighted value of the *k*retained *w*~*i*~values. The similar vectors are identified by calculating the Euclidean distance *d*between the vector **v**and every vector **w**. The *k*minimal distances *d*(**v**, **w**) are selected and the estimated value  is computed as follow \[[@B18],[@B19]\]:  In the weighted *kNN*, *s*(**v**, **w**^*t*^) = 1 / *d*(**v**, **w**^*t*^), this similarity measure *s*(**v**, **w**^*t*^) is deduced from the distance *d*(**v**, **w**^*t*^) between **v**and its neighbors **w**. In equation (1), more a vector **w**close to **v**, more it contributes to the estimation of the missing value. The *kNN*approach has no theoretical criterion to select the optimal *k*value (*k*~*opt*~) \[[@B19]\]. We have estimated for each subset the corresponding *k*~*opt*~using the sets without MVs to ensure a minimal bias in the comparison of the hierarchical clustering results. The determined *k*~*opt*~value is associated with a minimal global error rate as defined by Troyanskaya and co-workers \[[@B18]\]. Hierarchical Clustering ----------------------- The hierarchical clustering (HC) algorithm allows the construction of a dendogram of nested clusters based on proximity information \[[@B6]\]. The HC have been performed using the *hclust*package in R software \[[@B37]\]. Seven hierarchical clustering algorithms have been tested: average-linkage, complete-linkage, median-linkage, mcquitty, centroid-linkage, single-linkage and Ward minimum variance \[[@B13]\]. The distance matrix between all the vectors (*i.e*. genes) is calculated by using an external module written in C language. We used the normalized Euclidean distance *d\**to take account of the MV:  **v**and **w**are two distinct vectors and *m*is the number of MVs between the two vectors. Thus, (*v*~*i*~- *w*~*i*~) is not computed if *v*~*i*~and/or *w*~*i*~is a missing value An index for clustering results comparison: Conserved Pairs Proportion (*CPP*) ------------------------------------------------------------------------------ To assess the influence of missing data rates and different replacement methods into clustering results, we have analysed the co-associated genes of an original dataset (without MVs) compared to these genes location in a set with MVs. Hence, we realized in a first step the clusterings with the data sets without MV by each aggregative clustering algorithm. The results obtained by these first analyses are denoted *reference clusterings (RC)*. In a second step, we generated MVs in data. The MVs are replaced by using the different replacement methods. Then we performed the hierarchical clustering for each new set. The results obtained by these second analyses are denoted *generated clusterings (GC)*. We compared the resulting clusters defined in *RC*and *GC*. We assessed the divergence by using an index named *Conserved Pair Proportions*(*CPP*). The *CPP*is the maximal proportion of genes belonging to two clusters, one from the *RC*and the other one from the *GC*(cf. Figure [1](#F1){ref-type="fig"}). The procedure for computing the index *CPP*is as follow (figure [5](#F5){ref-type="fig"} gives an example of *CPP*computation): i\) For each reference clustering based on a given clustering algorithm (*algo*), we defined *K*^*algo*^, the number of clusters. As every type of hierarchical clustering algorithm gives a particular topology, we cannot use the same number of clusters to compare each aggregative method. So, we defined *K*^*algo*^such as its 10 most important cluster must represent 80% of the genes. For this purpose, we defined *K*^*init*^, an important initial number of clusters (equals to 500), and counted the number of occurrences associated to the 10 most populated clusters. Then we diminished *K*^*init*^by one unit and counted again. We stopped the process when the 10 most important clusters represent 80% of the occurrences (*K*^*algo*^= *K*^*init*^). We denote by  the *j*^*th*^cluster for a given clustering algorithm with *j*= {1, \..., *K*^*algo*^}. The  clusters are associated with their corresponding gene list . ii\) Three hierarchical clusterings are performed after generating MVs in proportion *τ*in the data, the first one without replacing data -- in this case, the normalized Euclidean distance (Eq.2) is used -, the second one after estimating the missing data by the *kNN*method (Eq.1), and the third one after replacing the missing data by zero. For each resulting tree, *K*^*algo*^clusters are defined. The  clusters are associated with their corresponding gene list , with *j\'*= {1, \..., *K*^*algo*^}. iii\) Finally, to estimate the *CPP*index, we searched for each  cluster the closest  cluster. For each clustering algorithm (*algo*), the corresponding  cluster is selected as the maximum number of genes  from the gene list  found in . Then, the Conserved Pairs Proportion (CPP) is computed as follow for one simulation:  where . The term  is the Kronecker symbol, *i.e*. it is equal to 1 when the genes *i*and *i\'*in the two gene lists are identical, otherwise 0. *G*denotes the total number of genes. This index takes the maximal value 1 when the clusterings *RC*and *GR*are identical. In addition, a variation of the tree topology may induce a *CPP*-variation. If the remaining genes of a cluster are in the direct neighbour clusters, the use of *CPP*can bias the analysis. Thus, we characterized the *CPP*~*f*~ratio to consider the *f*closest clusters of the  cluster. The computation of *CPP*~*f*~ratio is based on the previous ratio and corresponds to the *f*clusters which are the closest to the winning cluster. From a selected , the upper node of the dendrogram is examined. If the number of clusters linked to this node is inferior to *f*, the upper node is selected. This process is performed until the number of clusters is inferior or equals to *f*. The last node tested (*i.e*. with the number of clusters inferior or equal to *f*) is used to compute the *CPP*~*f*~ratio. Authors\' contributions ======================= AdB conceived of the study and carried out the MVs generation and the hierarchical clustering results. SH worked on the statistical analysis. AM has the initial idea of evaluating the MV influence on clustering, and participated in its design and coordination. All authors read and approved the final manuscript. Acknowledgements ================ The authors want to express their appreciation to Cristina Benros for her kind help in the course of this research. The authors thank their colleagues of Equipe de Bioinformatique Génomique et Moléculaire for fruitful discussions and Eric Molle and David Polverari from Atragene Bioinformatics for their technical assistance. This work was supported by grants from the Ministry for Research, Institute for Health and Medical Care (INSERM) and Genopole^®^. AdB was supported by the *Fondation de la Recherche Médicale*. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Research of the co-associated genes in clusters.**(i) The original dataset is cleaned by removing all the genes with MVs. (ii) A hierarchical clustering is performed for a given clustering method and the  clusters are selected to characterized the original co-association of the genes, (iii) MVs are generated with a *τ*rate per gene, and replaced by zero or by an estimated value computed with the *kNN*approach. (iv) Hierarchical clusterings are done for the pondered Euclidean distance, the *kNN*and zero replacement methods.  clusters are selected to characterize the new co-association of the genes. From this last step, the Conserved Pairs Proportion (*CPP*) is computed (see Methods). :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Example of clustering disturbance caused by missing values introduction.**Hierarchical clustering dendograms from the same set for the complete-linkage method with (a) no MVs and (b) MVs randomly introduced with a rate of *τ*= 1.0%. The genes from one cluster noted by a grey bar in (a) and are shown by stars (\*) in (b). :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Variation of the Conserved Pairs Proportion (*CPP*).**The mean *CPP*values of the OS (1/7) are given in function of the percentage of MVs for the seven aggregation methods (complete-linkage, Ward minimum variance, single-linkage, average-linkage, mcquitty, centroid-linkage and median-linkage). 200 independent simulations have been done per rate (*τ*) of MVs. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Comparison of real and *kNN*predicted values distribution.**Histograms of the real values of the OS (1/7) (black bars) and the predicted values using *kNN*method with *k*~*opt*~= 14 (grey bars) are shown. The values are shown in absolute value. All the values superior to 1.0 (respectively inferior to 0.1) have been pooled in the \'\> 1.0\' group (respectively \'\< 0.1\'). :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Example of CPP computation.**a) A hierarchical clustering is performed for a given algorithm and the  clusters are selected to characterize the original co-association of the genes (with *j*= 4 in the example). b) After the introduction of the MVs,  clusters are selected to characterize the new co-association of the genes. c) The number of common genes is computed between the *reference clustering*(ref) and the *generated clustering*(new). In yellow is given the maximum number of common genes allowing the identification of the closest  cluster for each  cluster. d) The corresponding *CPP*value is equal to 76.9% :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Missing values. Examples of missing value occurrences in different datasets from the yeast *S. cerevisiae*and human microarrays experiments. The year of publication, the studied organism, the number of genes on the DNA chip, the number of experimental conditions, the percentage of MVs and the percentage of genes with missing values are indicated. ::: Authors Ref Organism Number of genes Number of experimental conditions Missing Values percentage (%) genes with missing values (%) --------------------------- -------- ---------- ----------------- ----------------------------------- ------------------------------- ------------------------------- Bohen *et al*. (2003) \[21\] human 16523 16 7.6 63.6 Sorlie *et al.*(2003) \[22\] human 552 122 6.2 94.7 Garber *et al .*(2001) \[23\] human 918 73 2.4 72.7 Alizadeh *et al*.(2000) \[24\] human 4026 96 5.3 78.8 Gasch *et al.*(2000) \[25\] yeast 6153 178 3.0 87.7 Ogawa *et al.*(2000) \[26\] yeast 6013 8 0.8 3.8 Ferrea *et al.*(1999) \[27\] yeast 5769 4 10.6 30.9 Spellman *et al*. (1998). \[28\] yeast 6178 77 5.9 91.8 ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Influence of the replacement method on the Conserved Pairs Proportion (CPP). The numbers of clusters defined per metric are indicated (*K*^*algo*^). The *CPP*and *CPP*~*f*~values are given for the (1/7) subset of Ogawa \[26\] with missing values introduced (a) in the entire subsets and (b) only in extreme values (values \> 1.5 or \< -1.5). The *CPP*was computed by using the seven algorithms on the data without replacing the MVs (*MV*), after replacing the MVs by the *kNN*method (*kNN)*with the *k*~*opt*~values (see Table 3), or, after replacing the MVs by zero (*zero*). For *kNN*and *zero*values, the *CPP*and *CPP*~*f*~(with *f*= 5) values are given in (%). The percent of gain for all the simulations is indicated (+/-), i.e. the proportion of simulations with a better *CPP*value than the *CPP*value of *MV*. The values given in this table are the average values for MV in the range \[1%; 50%\] of genes with one MV per genes. ::: (a) --------------- ------------- ------- ---------- -------- ------ -------- -------- -------- -------- ------ -------- *single* 175 96.4 98.3 *99.3* 98.2 *99.6* 98.8 *99.1* *53.7* 99.0 *47.8* *centroid* 175 95.3 96.0 *74.8* 95.7 *64.1* 97.7 *97.9* *57.6* 97.8 *53.3* *average* 83 73.2 75.3 *66.3* 74.8 *62.0* 92.0 *92.5* *54.4* 93.0 *57.6* *median* 165 70.1 71.5 *55.2* 72.2 *54.4* 85.2 *86.0* *52.8* 86.2 *52.1* *mcquitty* 48 58.2 59.1 *54.5* 58.6 *50.1* 74.9 *75.3* *51.9* 75.1 *49.5* *Ward* 19 51.0 52.4 *62.4* 50.7 *49.1* 71.8 *73.3* *60.2* 71.5 *49.1* *complete* 36 47.9 49.2 *60.4* 48.4 *51.9* 71.9 *72.7* *53.7* 72.1 *51.6* (b) *CPP* *CPP*~f~ *MV* *kNN* *zero* *MV* *kNN* *zero* **Algorithm** *K*^*algo*^ (%) (%) (+/-) (%) (+/-) (%) (%) (+/-) (%) (+/-) *single* 175 97.9 98.4 *98.5* 98.7 *98.3* 98.8 99.3 *86.9* 99.7 *76.6* *centroid* 175 96.5 96.7 *64.8* 96.8 *60.6* 98.5 98.5 *55.9* 97.7 *32.8* *average* 83 81.4 81.0 *45.1* 83.3 *66.1* 93.9 94.5 *57.0* 94.9 *61.3* *median* 165 75.0 74.2 *45.7* 74.2 *47.9* 87.0 86.5 *50.3* 85.1 *45.4* *mcquitty* 48 71.6 71.3 *48.1* 71.9 *51.0* 83.1 83.2 *49.6* 84.6 *58.2* *Ward* 19 63.8 62.4 *42.1* 60.8 *29.9* 83.0 81.1 *37.1* 79.8 *27.0* *complete* 36 56.5 55.2 *44.5* 56.0 *47.2* 75.7 74.6 *46.1* 74.6 *45.0* ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### *kNN*approach: The error rates according to the number of genes. The sets of Ogawa (OS, \[26\]) and Gasch (GS, GS~HEAT~and GS~H2O2~, \[25\]) have been cut into subsets representing from (1/2) to (1/7) of the original sets. For each of the subsets, the optimal value for *k*(*k*~*opt*~), i.e. the number of neighbours used in *kNN*approach, has been computed. It was chosen as the number which minimizes the *error rate*. In regards to each subset is given the number of genes (*nb genes*). ::: OS GS GS~HEAT~ GS~H2O2~ --------- ------------ -------------- ------------ ------------ -------------- ------------ ------------ -------------- ------------ ------------ -------------- ------------ subsets *nb genes* *error rate* *k*~*opt*~ *nb genes* *error rate* *k*~*opt*~ *nb genes* *error rate* *k*~*opt*~ *nb genes* *error rate* *k*~*opt*~ (1/7) 827 0.190 14 815 0.292 18 523 0.183 12 717 0.143 14 (1/6) 968 0.186 20 955 0.276 8 608 0.199 15 837 0.155 11 (1/5) 1159 0.187 20 1141 0.283 23 731 0.195 22 1003 0.155 17 (1/4) 1448 0.180 22 1427 0.270 26 913 0.199 28 1254 0.146 13 (1/3) 1929 0.180 23 1903 0.274 8 1215 0.187 23 1669 0.139 15 (1/2) 2892 0.174 30 2853 0.255 7 1822 0.187 8 2504 0.135 16 set 5783 0.177 39 5705 0.255 8 3643 0.186 10 5007 0.135 17 :::

PubMed Central

2024-06-05T03:55:47.191651

2004-8-23

PMC514702

Background ========== Multiple large-scale analytical techniques to assess gene expression in defined cell populations have been developed. These include microarray analysis, EST enumeration, SAGE and MPSS. Each of these techniques offers unique advantages and disadvantages. Technique selection largely depends on the expertise of the investigator, the cost, the availability of the techniques, the amount of RNA/DNA that is available, and the existence of the genome databases. The human genome dataset is the best annotated one available \[[@B1],[@B2]\]- making large scale gene expression analysis of human tissues and cells uniquely fruitful for investigators due to the increased ability to identify full length transcripts with predicted gene function instead of EST\'s. Human ES cells have been isolated relatively recently and ES cell genes are underrepresented in current databases. More importantly, recent evidence has suggested that mouse ES and human ES cells differ significantly in their fundamental biology \[[@B3],[@B4]\] and one cannot readily extrapolate from one species to another. However, comparing results between species may provide unique insights. Given the wealth of SAGE and microarray data available from rodent ES cells examining human ES cells with similar techniques as has been done recently by several investigators \[[@B3]-[@B11]\] should be very useful in furthering our understanding of this special stem cell population. Until recently however, it has been difficult to obtain RNA from a homogenous population of undifferentiated hESC for such an analysis as cells could not be grown without feeders and few unambiguous ES cell markers had been described. However, we and others have now described markers that will clearly assess the state of ES cells using a combination of immunocytochemistry and RT-PCR \[[@B3],[@B12],[@B13]\] In addition, techniques of harvesting ES cells away from feeder layers have been developed and verified (our unpublished results) and methods of growing ES cells without feeders have been described \[[@B14]\]. These techniques, have allowed us (and others) to obtain large amounts of validated RNA/cDNA samples for comparison by microarray \[[@B3]-[@B11]\], SAGE \[[@B8]\] or EST enumeration \[[@B9]\]. We selected MPSS for this analysis as it offers some unique advantages over other methods including SAGE \[[@B15],[@B16]\]. MPSS offers sufficient depth of coverage when over one million transcripts are sequenced \[[@B16]\] and is efficient, as the numbers of sequences obtained are an order of magnitude larger than with shotgun sequencing or SAGE. It is relatively rapid with a turnaround of a six to ten weeks, and if done with human tissues, more than 80% of transcripts can be mapped to the human genome with current tools. Further, independent analysis has suggested that expression at greater than 3 tpm (transcripts per million) is predictive of detectable, reliable expression, equivalent to roughly one transcript per cell -- a sensitivity that is unparalleled when compared to other large-scale analysis techniques \[[@B16]\]. Finally, MPSS libraries can be translated into SAGE libraries and compared to existing SAGE library sets using freely available tools such as digital differential display, allowing ready comparisons to existing SAGE/MPSS libraries of mouse ES cells. It is important to note that we found 14 base pair SAGE tags are generally not as specific as 17 base MPSS signatures and that SAGE sampling depth is usually insufficient. Newer technologies such as extended sequencing to 20 base pairs in MPSS, 24 base pairs in SAGE or cheaper bead alternatives such as those described by Illumina may offer additional depth of coverage and a cheaper price but these at present remain limited in availability. We have utilized MPSS using a pooled sample of three human ES cell lines grown in feeder-free culture conditions over multiple passages \[[@B17],[@B18]\] to assess the overall state of undifferentiated ES cells. Our rationale for using pooled sample rather than individual samples was based on the fact that no standardized medium and culture conditions have been established for growing and propagating ES cell lines. Variation observed by sampling single lines may be due to culture conditions rather than intrinsic differences. We reasoned therefore that a need existed to establish a reference baseline using pooled samples to enhance the similarities and provide evidence for candidate genes that should be examined for differences such as expression of HLA genes, Y chromosome and X chromosome genes, imprinted genes and genes regulating the methylation state. Our results show that MPSS provides a greater depth of coverage than EST scan or microarray and provides a comprehensive expression profile for this stem cell type. The data set generated allows us and others to identify multiple genes that were not previously known to be expressed in this population, including novel gene as well as obtain a global overview of pathways that are active during the process of self-renewal. Results ======= MPSSS analysis of pooled samples -------------------------------- A pooled sample of undifferentiated human ES cell lines H1, H7, and H9 grown in feeder-cell free conditions \[[@B19]\] was used for the preparation of mRNA as previously described \[[@B20]\]. Growth without feeders avoids complication from feeder contamination, which even with good harvesting techniques \[[@B14],[@B21]\] ranges between 1--3% (unpublished data) and is sufficient to be detected by MPSS (Dr. B. Lim-Harvard University personal communication). Under these conditions, 80--95% of the cells express SSEA-4, 91--94% express TRA-1-60, and 88--93% express TRA-1-81, previously described markers for undifferentiated hESC \[[@B19]\]. Microarray analysis of 2802 genes suggests that these cells are remarkably similar in their gene expression profiles, with only 5 genes being more than 2-fold different between the three cell lines \[[@B17],[@B18]\] (and data not shown). The undifferentiated state of the cells was also assessed by RT-PCR of known markers of undifferentiated hESC on mRNA of the pooled hESC sample (Figure [1](#F1){ref-type="fig"}). In addition absence of early markers of differentiation was assessed. No expression of GATA, Sox-1, nestin, Pdx-1 or markers of trophoectoerm were detected in samples used (Supplementary table 3a, see also 3) ::: {#F1 .fig} Figure 1 ::: {.caption} ###### RT-PCR analysis (a), cumulative tpm (b) and tpm of known ES cell markers (c) is shown. Note that MPSS identifies most known markers of huES cells and expression is at high tpm levels. \* -- signature maps to \>100 location in the genome (class 0); \*\* -- artifactual (class 5) signature :::  ::: Pooled mRNA of the three hESC lines was subjected to MPSS analysis at Lynx Therapeutics (Hayward, CA), generating 22,136 distinct and significant signature sequences from a total of 2,786,765 sequences (see Methods and [additional file 1](#S1){ref-type="supplementary-material"}). Each signature was ranked, as outlined in Methods (Table [1](#T1){ref-type="table"}), based on its position and orientation within the transcript, and the presence of a polyadenylation signal and polyA in the transcript sequence. 16,675 signatures (75%) mapped to UniGene transcripts; 40 signatures (0.2%) mapped to mitochondrial transcripts; 3,818 signatures (17%) matched genomic sequences but did not map to a UniGene cluster; 927 (4%) signatures matched sequences present at more than 100 genome locations (class 0, representing transcripts containing repetitive elements in their 3\' UTR). 676 (3%) signatures did not match to genome or UniGene sequences. Some UniGene clusters contain multiple signatures. These signatures likely represent either transcripts of alternative termination sites, or artefacts of MPSS library construction. Signature classification helps to distinguish artifactual signatures from signatures representing expressed transcripts. For example, signatures of class 1 to 3 are 3\'most signatures in mRNA or EST sequences with poly (A) signal and/or polyA tail and most likely represent transcripts with multiple polyadenylation sites. Artifactual signatures constituted 1--3% of the tpm count of the \"real\" signature, although occasionally close counts were observed (data not shown; see supplementary data tables, [additional files 2](#S2){ref-type="supplementary-material"}, [3](#S3){ref-type="supplementary-material"}). To simplify the MPSS data analysis and pair-wise comparison of ES cell data from this study to other datasets, multiple signatures mapping to the same Unigene ID (Hs build 169) were combined into one tpm count as the sum of tpm for signatures of class 1, 2, 3, 22, 23 if any found. These are 3\'most signatures close to polyA signal and/or polyA tail, most probably representing true transcripts with alternative termination. If no signatures of above classes were found, then sum of class 4 (3\'most, no polyA features) was used. If none the above, the sum of class 5 signatures was used for the tpm calculation per unigene cluster. Resulting table containing data for 8679 unigene clusters, 11 mitochondrial genes, and including 1991 signatures that did not map to unigene but uniquely matched genomic sequences (potential novel transcripts), is presented in supplementary table ([additional file 4](#S4){ref-type="supplementary-material"}) and available for download from Lynx \[[@B27]\]. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Classification of the MPSS cDNA signatures. The signature classification used for annotation is shown \* The Class 0 signatures are the signatures that hit genome more than 100 times, which is treated as a \"repeat sequence\". \*\* The polyA tail is defined as a stretch of A\'s (at least 13 out of 15 bases) that is no more than 50 bases away from the end of the source sequence. The polyA signal is either AATAAA or ATTAAA that has at least one base within the last 50 base before the end of the source sequence or the polyA tail. \*\*\* All the virtual signatures extracted from the genomic sequences are classified as class 1000 signatures. ::: **Virtual Signature Class** **MRNA Orientation** **Poly-Adenelation Features \*\*** **Position** ----------------------------- ------------------------------------------ ------------------------------------ -------------------- 0\* Either -- Repeat Warning Not applicable Not applicable 1 Forward Strand Poly-A Signal, Poly-A Tail 3\' most 2 Poly-A Signal 3\' most 3 Poly-A Tail 3\' most 4 None 3\' most 5 None Not 3\' most 6 Internal Poly-A Not 3\' most 11 Reverse Strand Poly-A Signal, Poly-A Tail 5\' most 12 Poly-A Signal 5\' most 13 Poly-A Tail 5\' most 14 None 5\' most 15 None Not 5\' most 16 Internal Poly-A Not 5\' most 22 Unknown Poly-A Signal Last before signal 23 Poly-A Tail Last before tail 24 None Last in sequence 25 None Not last 26 Internal Poly-A Not 3\' most 1000\*\*\* Unknown -- Derived from Genomic Sequence Not applicable Not applicable ::: The frequency distribution of the signatures shows that the 200 most abundant signatures represent 99% of the total number of signature counts obtained from the hESC (Figure [1](#F1){ref-type="fig"}). Most of top 200 genes (unigene clusters, [additional file 5](#S5){ref-type="supplementary-material"}) represent ribosomal genes and genes involved in protein and nucleic acid synthesis and are consistent with results obtained by EST scan and other analyses (data not shown, and \[[@B5],[@B8],[@B9]\]). We note that several ribosomal genes were identified as being overexpressed by microarray, SAGE and EST scan as well (see [additional files 16](#S16){ref-type="supplementary-material"}, [17](#S17){ref-type="supplementary-material"}, [18](#S18){ref-type="supplementary-material"}). Comparison of the pattern of gene expression with other cell types showed a very similar expression profile with housekeeping genes being the predominant population of sequences in all cell types examined (data not shown). Only three known ES cell specific genes were present in the top 200 genes ([additional file 5](#S5){ref-type="supplementary-material"} and Figure [1](#F1){ref-type="fig"}). These included SOX-2, DNMT3β, and Oct-4. As in other cells cell type specific genes, transcription factors and cytokines were present at much lower abundance (\<50 tpm on average). These low tpm level genes were often not detected by other methods (discussed below). The expression level of cell surface receptors for fibronectin are high (ITGB1 -- 578 tpm) and their presence was confirmed by immunocytochemistry and RT-PCR, suggesting that feeder-free clones may grow well on this substrate (data not shown, see also Figure [2](#F2){ref-type="fig"} and \[[@B14],[@B21]\]). The major signaling pathways represented in the top 200 most abundant genes are the FGF signaling pathway, with FGFR1 being most abundant (673 tpm, Figure [2](#F2){ref-type="fig"}), and the ras activated pathway, with two members of the ras family (NRAS-related and ran) being present in the top 200. This is consistent with data that E-Ras is critical for rodent ES cell self-renewal \[[@B22]\]. No transcripts for HRASP (Homologue of ERAS pseudogene) were detected however (Figure [2](#F2){ref-type="fig"}), suggesting that these other ras family members may subserve this critical role of self-renewal \[[@B9]\]. The absence of E-Ras was confirmed by RT-PCR (data not shown), as was the presence of FGFR1 (Figure [2](#F2){ref-type="fig"}, \[[@B22]\], and data not shown). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### RT-PCR for E-ras/RASP, FGFR1 and novel genes identified as enriched in undifferentiated ES cells is shown in Panel A and B. Localization of E-cadherin and β-catenin in undifferentiated ES cell is shown in Panel C. All of the genes identified by MPSS and tested were present in undifferentiated ES cells and most were significantly downregulated as cells differentiated. Note the high expression at the cell surface and low or undetectable levels of β-catenin in the nucleus. :::  ::: Major pathways present at detectable levels by MPSS --------------------------------------------------- To gain a broad overview of the properties of hESC, we mapped the genes found in the hESC cells to the human genome to get an overview of the chromosomal distribution of genes expressed in hESC (Figure [3](#F3){ref-type="fig"} and [additional files 6](#S6){ref-type="supplementary-material"}, [7](#S7){ref-type="supplementary-material"}, [8](#S8){ref-type="supplementary-material"}, [9](#S9){ref-type="supplementary-material"}, [10](#S10){ref-type="supplementary-material"}, [11](#S11){ref-type="supplementary-material"}). Overall, MPSS detected gene expression in most of the previously identified zones of transcriptional activity within chromosomes. Two chromosomal regions contained more genes expressed in hESC -- than expected, and several regions where fewer genes were expressed, compared to the total number of genes located within a particular chromosomal region. No bias to chromosome 17, 12 or X was seen either in overall gene expression or in a particular cytoband. The failure to detect a bias was confirmed by mapping EST scan data \[[@B8]\] as well. The overall distribution patterns were similar and did not show any bias at this level of resolution. Interestingly, gene expression from both X and Y chromosomes was observed. Unlike rodent ES lines both male and female ES lines have been obtained with roughly equal frequency \[[@B20]\] suggesting that when individual cell lines are examined differences between levels of expression between male and female will be present and detectable. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Cytoband mapping of ES cell expressed genes and regions of relatively high and low transcription relative to the refseq database is shown. More detailed mapping information is presented in supplementary tables. :::  ::: Likewise, MPSS detected expression of several MHC Class I and II genes, suggesting that MPSS can identify differences between ES cell samples when HLA gene expression is used to type cells \[[@B17],[@B18]\]. We also note that both H19 and Igf2 were expressed at detectable levels. H19 and Igf2 are located adjacent to each other on chromosome 11p15.5 and are reciprocally regulated by imprinting, H19 being paternally imprinted, and IGF2 being maternally imprinted \[[@B23],[@B24]\]. It is therefore likely that their ratio of expression is likely to differ between cell populations and may represent a simple assessment of the imprinting status of cells. We classified genes expressed into ECM related, homeobox containing, zinc finger proteins, novel genes as well as genes which could assigned to major signaling pathways such as wnt, BMP/TGFβ, LIF, receptors, etc. This data is provided in excel files in the supplementary information provided ([additional files 12](#S12){ref-type="supplementary-material"}, [13](#S13){ref-type="supplementary-material"}). Overall certain general themes emerged when genes were classified into such a fashion. We find that: A) hESC express markers characteristic of ES cells in general and few markers characteristic of differentiated cells confirming the initial purity of ES cells used for this analysis and the fidelity of the analysis B) Ribosomal protein transcripts, and mitochondrial genes are highly expressed in ES cells (relative to other transcripts) and constitute more than 50% of the total transcripts analyzed (Figure [1](#F1){ref-type="fig"}, [additional files 5](#S5){ref-type="supplementary-material"}, [16](#S16){ref-type="supplementary-material"}, [17](#S17){ref-type="supplementary-material"}, [18](#S18){ref-type="supplementary-material"}). And this is similar to other samples analyzed \[[@B3]-[@B11]\], (Lynx Inc. data not shown) C) Positive regulators of the cell cycle, TERT and antisenescence related genes and DNA repair pathway regulators are expressed at high levels while proapototic genes, Rb and p53 pathways regulators are expressed at low levels (see table [2](#T2){ref-type="table"} for an example of TERT related gene expression, see supplementary tables ([additional files 12](#S12){ref-type="supplementary-material"}, [13](#S13){ref-type="supplementary-material"}) for cell cycle, apoptosis and other pathways) D) The number of novel genes or genes of unknown function is high (2600/11,000) and constitutes approximately 25% of the unique signatures (see [additional file 13](#S13){ref-type="supplementary-material"} for a listing of genes of unknown function, their chromosomal mapping, and UniGene identity). Comparison with other samples suggest that the number of novel genes or genes of unknown function seen are higher in ES cells (25% versus 20%). E) Components of most major signaling pathways are present but so are negative regulators (including zinc finger proteins), suggesting that inhibition plays an important role in maintaining cells in an undifferentiated state (see [additional file 13](#S13){ref-type="supplementary-material"}). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Senesence and Aging related genes A subset of genes related to senescence and aging that may regulate the lack of senescense in ES cells is shown. Note that the telomerase, morf\'s, nortalins and sirtuins are all expressed in ES cells. \*The TERT gene has a signature uniquely mapping to an intron (cryptic exon?), which was present in all runs of the ES cell analysis and was not found in other human samples (not shown). ::: **HuES\_TPM** **Gene Bank** **Hs169** **Gene** **chr** --------------- --------------- ----------- ---------- --------- 802 BC029378 Hs.442707 TERF1 8q13 56 AF289599 Hs.274428 TERF2IP 16q23.1 38 AI742882 Hs.409194 TNKS 8p23.1 15 AF002999 Hs.63335 TERF2 16q22.1 10 AW271065 Hs.9645 TNKS1BP1 11q12.1 9 BC005030 Hs.7797 TINF2 14q11.2 7 AF264912 Hs.280776 TNKS2 10q23.3 10\* NM\_003219 Hs.439911 TERT 5p15.33 321 NM\_004134 Hs.184233 HSPA9B 5q31.1 94 AF070664 Hs.374503 MORF4L1 15q24 80 BC017305 Hs.528641 SIRT7 17q25 42 AF100620 Hs.411358 MORF4L2 Xq22 27 NM\_012238 Hs.31176 SIRT1 10q22.1 10 BM803485 Hs.511950 SIRT3 11p15.5 16 AL579291 Hs.282331 SIRT5 6p23 ::: Examination of signaling pathways suggest that wnt, TGFβ and FGF signaling pathways are likely important in regulating the ES cell state while LIF/gp130 signaling is not as important. These conclusions are based on examining the expression of the positive and negative regulators of a particular pathway by MPSS, and EST scan. When critical components are low or absent we have tentatively assumed that the pathway is unlikely to be active. An example of the Igf/PTEN pathway is shown to illustrate the logic (Table [3](#T3){ref-type="table"}) and other pathways along with verification with EST scan are summarized in the supplementary tables ([additional files 12](#S12){ref-type="supplementary-material"}, [13](#S13){ref-type="supplementary-material"}). Note the high levels of soluble frizzled receptors and the expression of E-cadherin (negatively regulating β-catenin translocation). The expression of cadherin and β-catenin was confirmed by immunocytochemistry (Figure [2](#F2){ref-type="fig"}). The relatively fidelity of the conclusion was confirmed by examining the expression of E-cadherin by immunocytochemistry and localizing β-catenin expression. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### IGF-/PTEN/Akt and Ras/Raf/MAP pathway A subset of genes related to Igf/PTEN pathway that are expressed in undifferentiated ES cells is shown. Note that the overall pattern of expression suggest that this pathway is active in undifferentiated ES cells. ::: **Tpm ES** **Tpm EB** **Unigene ID** **Locus ID** **Description** ------------ ------------ ---------------- -------------- ----------------- 14 32 Hs.239176 3480 IGF-1 receptor N.D. N.D. Hs.390242 3667 IRS-1 0 0 Hs.253309 5728 PTEN N.D. N.D. Hs.32942 5294 PI3K 11 8 Hs.433611 5163 PDK1 0 15 Hs.92261 5164 PDK2 N.D. N.D. Hs.6196 3611 ILK 75 157 Hs.368861 207 AKT1 15 82 Hs.170133 2308 FKHR (FoxO1A) 78 54 Hs.14845 2309 FKHRL1 (FoxO3A) 15 88 Hs.282359 2932 GSK3beta 39 14 Hs.238990 1027 p27 280 240 Hs.371468 595 Cyclin D1 0 594 Hs.370771 1026 p21 1 10 Hs.329502 842 Caspase 9 0 39 Hs.76366 572 Bad 98 193 Hs.260523 4893 N-Ras 2 0 Hs.37003 3265 H-Ras 35 128 Hs.257266 5894 Raf1 N.D. N.D. Hs.132311 5604 MEK1 128 218 Hs.366546 5606 MEK2 37 75 Hs.324473 5594 ERK (p42 MAPK) ::: We compared the signature sequences detected in the hESC to an MPSS database of 36 human tissues and cell lines to look for genes that are unique to, or highly overexpressed in hESC. A list of several hundred was generated when a cutoff of 30 tpm or higher (ten fold above detection level) that were elevated in ES cells when compared to neural stem cells examined in a similar manner was used. This list is provided in supplementary materials ([additional file 14](#S14){ref-type="supplementary-material"}). A list of 13 highly enriched genes of unknown function is shown in Table [4](#T4){ref-type="table"}, and the tpm values for the corresponding signatures in each of 36 tissues or cell lines is provided in the supporting information ([additional file 15](#S15){ref-type="supplementary-material"}). The expression in ES cells, of these 13 genes was confirmed by designing PCR primers to different regions and examining gene expression (Figure [2](#F2){ref-type="fig"}). Several of these genes are highly expressed in hESC and absent in most other tissues tested (Table [4](#T4){ref-type="table"}, additional file 15, and data not shown), are downregulated as ES cells differentiate (Figure [2](#F2){ref-type="fig"}), and are good novel, candidate markers for undifferentiated hESC. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Novel genes enriched in hESC as assessed by MPSS A short list of genes of unknown function that are highly enriched in three ES cell lines comparing to 36 different tissues and cells are shown. A complete list of unknown genes expressed in pooled hESC cells is presented in supplementary tables. \* NS-neural stem cells, TH-thymus, HY-hypothalamus, PG-pituitary gland, TE-testis \*\* this gene (Hs.507833 in the unigene Hs.169) is transcribed in antisense to HDCMA18P (Hs.278635) ::: **SIGNATURE** **HuES,TPM** **Chr** **GB:description** **Other 36, TPM\*** ------------------- -------------- --------- --------------------------------------------------- ----------------------- GATCTCCAGTAGACTTA 1646 4 CD250365:Homo sapiens transcribed sequence \*\* **NS-10** GATCTGTTAACAAAGGA 967 16 BC008934:claudin 6 **ND** GATCTAGAAGTTGCAAC 489 1 NM\_019079:hypothetical protein FLJ10884 **ND** GATCTTTTTTTTTGCCC 455 3 NM\_018189:hypothetical protein FLJ10713 **TH-47, HY-3, PG-3** GATCCCCATCCAAAAGA 366 7 AI636928:Homo sapiens transcribed sequences **MCF7-2** GATCCACCTAGGACCTC 244 X CD174249:Homo sapiens transcribed sequence **ND** GATCCGCCTCCTTGGCC 240 4 AK092578:Sapiens cDNA FLJ35259 fis **ND** GATCCTAGCCAAGCCCC 169 3 BF223023:Homo sapiens transcribed sequences **ND** GATCTGGCCCGCCACCA 150 16 NM\_032805:hypothetical protein FLJ14549 (ZNF206) **ND** GATCGTTGTGGTGGACT 146 3 XM\_067369:similar to Heterochronic gene LIN-41 **ND** GATCCACCACATGGCGA 92 11 CD176172:Homo sapiens transcribed sequence **ND** GATCCAACAATTCTACT 78 U CD173198:Homo sapiens transcribed sequences **TE-33** GATCTTCTAAACCCATC 75 12 BU608353:Homo sapiens transcribed sequence **ND** ::: Comparing with other data sets ------------------------------ Recently we and others have begun examining hESC with EST scan \[[@B10]\] and microarray analysis to develop a characteristic profile of this unique population \[[@B3]-[@B10]\]. We used this data to compare the sensitivity of MPSS with EST scan and microarray analysis. We have previously reported a set of 90 genes reported common to 6 different hESC lines \[[@B10]\]. Of these, eighty-five were detected by MPSS showing a high degree of concordance (\>90%). Of the five genes missing from the MPSS hESC data set, four of the genes had valid MPSS signatures (Table [5](#T5){ref-type="table"}) and were readily detected in other human samples (data not shown). One gene (SNRPF) lacked a DpnII (GATC) site making it non-detectable by MPSS. GDF3 was detected at non-significant level in the hESC, though was detected by MPSS at higher level (10--30 tpm) in other ES cells tested (Dr. B. Lim-Harvard University personal communication, and [additional file 17](#S17){ref-type="supplementary-material"}). Sperger et al., also used microarray to examine gene expression in undifferentiated cell lines \[[@B11]\]. They compared expression in undifferentiated cells with expression in EC carcinoma lines and with microarray data from several other cell lines. They have identified 895 genes (GenBank accession numbers) which reduce to 718 number of unigene identities when mapped to the unigene build Hs161. We have compared this data with the MPSS data and see that MPSS identified the large majority of these genes as well ([additional file 16](#S16){ref-type="supplementary-material"}). Similar results were obtained when data was compared with that reported by Sato et al., and Abetya et al., \[[@B6],[@B7]\] and a similar concordance in gene expression was observed (data not shown). Thus, MPSS provides an independent verification of the microarray results and in addition identifies other genes that may not be present on the arrays or detectable by current microarray techniques. ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### MPSS tpm of genes reported as enriched by microarray in hESC Table 5 Tpm of genes identified as overexpressed microarray analysis of six pooled human ES cell lines. Note that most of them have high tpm values and are detected by MPSS. \* -- PSIP2 and PSIP1 have 3\' alternate termination and distinguished by MPSS (but not by microarray); \*\* -- PODXL: TPM for signature of class 5; 3\' most signature has double palindrome and underrepresented. \*\*\* -- higher expression of GDF3 was detected in other ES cells (suppl.table for BG02 and not shown). \*\*\*\* -- expression detected in other human samples (not shown). ::: **GB\_accession** **Gene Symbol** **HuES\_TPM** ------------------- ----------------- --------------- X85372 SNRPF No GATC NM\_002295 LAMR1 6135 D23660 RPL4 5269 NM\_001002 RPLP0 4656 NM\_002520 NPM1 3207 X69391 RPL6 3745 M31520 RPS24 3183 AF070600 OK/SW-cl.56 2702 X57958 RPL7 1923 NM\_024674 LIN-28/ 1692 NM\_145899 HMGIY 1618 NM\_018407 LAPTM4B 1326 M94314 RPL24 1279 X62534 HMGB2 989 D13748 EIF4A1 1070 NM\_006086 TUBB4 809 J04164 IFITM1 788 X69804 SSB 874 M93651 SET 1323 D00760 PSMA2 673 AL162079 SLC16A1 991 AF225425 SEMA6A 742 U28386 KPNA2 542 X74929 KRT8 543 NM\_002300 LDHB 527 M97856 NASP 536 AF311912 SFRP2 457 AF020038 IDH1 450 D83174 SERPINH1 477 S74445 CRABP1 437 NM\_000165 GJA1 392 AB040903 TD-60 524 AF063020 PSIP2\* 389 U76713 HNRPAB 166 NM\_000224 KRT18 302 NM\_021144 PSIP1\* 389 M94856 FABP5 257 NM\_016304 Ribo 60S L30 247 AK094423 HNPRA1 like 214 AF055270 HSSG1 (SFRS7) 201 M77140 GAL 199 AF257659 CALU 100 AF098158 C20orf1 338 U41387 DDX21 179 AD001528 SMS 175 NM\_006548 IMP-2 177 AJ223953 PTTG1 154 X54326 EPRS 210 D13627 CCT8 167 NM\_012247 SEPHS1 306 D00762 PSMA3 123 AF005418 CYP26A1 121 M25753 CCNB1 168 NM\_000884 IMPDH2 174 X16396 MTHFD2 113 NM\_005159 ACTC 98 U31814 HDAC2 112 J04031 MTHFD1 104 NM\_006341 MAD2L2 95 J03746 MGST1 88 NM\_020997 LEFTB 62 M74091 CCNC 86 AK001962 BRIX 66 M36981 NME2 93 AL133611 Novel 63 X05360 CDC2 62 AB040930 LRRN1 46 AF071592 KIF4A 71 AF015254 STK12 41 X14253 TDGF1 37 AB023420 HSPA4 42 M19309 TNNT1 54 BC004200 PPAT 34 NM\_024090 ELOVL6 23 NM\_014366 NS 30 U97519 PODXL 26\*\* AF048722 PITX2 25 NM\_024498 ZNF117 32 NM\_001878 CRABP2 24 X59244 ZNF43 13 BC001068 C20orf129 17 NM\_024865 Nanog 15 NM\_024900 Jade-1 11 AB046793 KIAA1573 11 Z26317 DSG2 18 NM\_020634 GDF3 1\*\*\* AF070651 ZNF257 0\*\*\*\* NM\_016448 RAMP 0\*\*\*\* U88573 NBR2 0\*\*\*\* AB044157 GSH1 0\*\*\*\* ::: Comparison with an EST scan analysis of 37,081 EST sequenced from a similar pooled sample of hESC \[[@B9],[@B10]\] also showed a high degree of concordance. The EST scan analysis detected 8,801 distinct UniGene clusters in hESC versus 9,996 distinct UniGene clusters expressed at 4 tpm or higher in the MPSS dataset. Of the 8,801 UniGene clusters identified by the EST scan, 1,139 are singletons, i.e. identified by only one EST out of the 37,081 total EST\'s. 5,286 UniGene clusters have 5 or more ESTs as evidence, and only 118 UniGene clusters have more than 100 EST\'s as evidence. In contrast, all 9,996 UniGene clusters identified by MPSS were detected at 4 or more tpm and identified in multiple sequencing runs. More than 8,000 have at least 10 tpm, and over 1,000 have more than 100 tpm. Thus, although the EST\'s are longer in length and thus easier to assign to a particular gene, MPSS appears more sensitive than EST scan. MPSS for example identified almost twice as many genes as EST scan consistent with the difference in the depth of analysis (No of sequences MPSS/EST). Richards et al \[[@B8]\] have used SAGE analysis to two ES cell lines. Their analysis revealed expression of approximately four thousand genes which was significantly fewer than that identified by MPSS consistent with the fewer number of gene tags sequenced. Comparison of the data sets however showed good concordance particularly for genes expressed at higher tpm levels. The entire comparison is presented in supplementary table ([additional file 18](#S18){ref-type="supplementary-material"}) and is available for download from Lynx \[[@B27]\]. Overall MPSS could identify genes that other methods identified with an average concordance rate of 70%. The depth of analysis with MPSS at 2.4 million signatures however was significantly greater. MPSS in general identified many more genes than microarray or EST scan or SAGE (see above). The most direct comparison is with EST scan or SAGE, which do not rely on comparative gene expression to establish significance of gene expression. Overall our comparison suggests that MPSS results provide a complementary global overview of the transcriptome of the ES cell. The data supplement and extend the microarray, SAGE and EST scan data sets and provide an independent verification of the same. MPSS in addition identifies additional genes expressed particularly at lower tpm, that are either not present on microarrays or not detected with a lower resolution analysis. Discussion ========== Our results provide a global overview of the gene expression pattern of undifferentiated human ES cells and allow comparisons with other data sets. These results suggest the hESC are an actively dividing population of cells that exhibit high metabolic activity. Our analysis detected expression of approximately 10,600 unique transcripts, a figure that about a third of the total number of mapped genes. Unlike other cell types, however, a much larger fraction of unknown or novel genes was present. This high ratio likely represents the paucity of information available in existing libraries on this relatively newly characterized cell population rather than the possibility that ES cells use radically different pathways for self-renewal, survival, proliferation and differentiation. Our results confirm the reported differences between rodent and human ES cells. We confirm the absence of expression of ERAS, Ehox and the orthologs PEPP1 and 2. The apparent lack of LIF requirement of hESC is reflected by the absence or low tpm levels for genes of the LIF pathway and high tpm for suppressors of LIF mediated signalling (see supporting information). The high level of expression of genes in the FGF pathway likely reflects the requirement of hESC for bFGF. The high level of FGFR1 expression suggests that FGFR1 is an important signal transducer and that FGF\'s other than FGF4 are important in hESC self-renewal. The high tpm of the fibronectin receptor also suggest that fibronectin or vitronectin are likely useful substitutes for matrigel and that activation of ras mediated signalling is likely critical, as has been described in the rodent ES cell analysis \[[@B20]\]. Comparing data from the MPSS analysis with microarray, SAGE and EST scan analyses suggest that MPSS is a powerful alternative to these techniques. MPSS identified virtually all of the genes highlighted as genes common between six different human ES cell lines surveyed by microarray. We noted that most genes detected by microarray were expressed at high tpm indicating that MPSS is more sensitive than microarray analysis. MPSS however appeared to be able to identify genes detected by microarray. Analysing an additional 400 markers detected by MPSS using focused microarray or RT-PCR confirmed their expression \[[@B3]\], (data not shown). Likewise, MPSS analysis showed good concordance with the EST scan data at a fraction of the price. In contrast to the EST scan, tpm levels determined by MPSS are highly correlated to the mRNA levels present in the cells, even at low tpm values \[[@B25]\], and (Lynx unpublished results). Due to the low sampling number of most EST scans, this is not true for relatively low number of EST\'s found for a particular gene, and can be used only as a rough estimate of gene expression. Unlike other in depth analyses, the absence of markers in MPSS runs is also a powerful control provided that the marker possesses a GATC site. The chromosomal distribution of the genes expressed in hESC did not reveal any bias for a particular chromosome or chromosomal region. While a couple of \"hotspots\" and several \"cold spots\" were identified, in no case was any region comprised of all transcribed or all silent genes. Another important conclusion from our analysis is that selection of input RNA is critical. In our case we tested samples repeatedly to assess their purity and made considerable efforts to establish subclones that did not require feeder cells that could be potentially contribute transcripts to the analysis. Given the range of tpm of biologically relevant molecules (5 to 32,000 in this experiment) we predict that even a 5% contamination can confound results or detailed comparisons across different laboratories. We note also that gene transcription from both the X and Y chromosome is observed indicating that at least subtle differences will exist between male and female lines even in the undifferentiated state. Sex-based gene expression, along with MHC gene expression and ratio of expression of imprinted genes could serve to distinguish between different ES cell populations. The present results further suggest that analysing embryoid bodies that differentiate stochastically or analysing tissue samples (with variable proportions of cells) by MPSS will prove more difficult and that results will be variable. We suggest that variability can be reduced by pooling samples, normalizing by careful testing for known markers of differentiation, by semi quantitative PCR, or by focused microarray analysis. While MPSS is cost-effective and sensitive, it is by no means perfect. MPSS is limited by the requirement that DpnII sites (GATC) be present in a gene and be present in a unique locus such that the signature obtained is unique. For example, SNRF expression could not be assessed directly, as no GATC site is present. The signatures for ZFP42 are ambiguous and map to multiple transcripts. Although MPSS can distinguish between alternate transcript termination sites, MPSS cannot distinguish between alternative splicing events and possible incomplete digestion during the sample preparation process. Signature lengths are relatively short and it is possible to have to select between multiple genome hits (reviewed in \[[@B16]\]. Sequencing is performed four bases at a time and transcripts that contain palindromic sequences (in particular double palindromes) are often undetected because of self-hybridization of single DNA strands on the bead. A survey of the genome suggests that this is a rare event (approximately 3% of all virtual signatures in human MGC database have double palindromes). The NODAL gene is an example for such an event, where the class 1 signature was lost and NODAL expression is detected only by a signature resulting from incomplete digestion during library construction (see results). The success of MPSS analyses also depends to a large extent on the quality of genomic information available and, in our opinion, currently is best utilized to analyse human cells. Furthermore, MPSS itself may not be the best method for routine, lower throughput analyses, given price per sample, sample processing time and the large amount of data generated, which requires considerable analysis. However, the database, once developed, is extremely valuable provided it is freely available to make comparisons and to select subsets of genes for further analysis. MPSS information can be effectively utilized by establishing a common database of markers expressed at a defined stage in the differentiation of cells. Additional data sets from sampling of cells at well-controlled stages of differentiation that can be readily accessed and compared to existing datasets will provide the most information while still being cost effective. The genome database is an example of such sharing that has proven to be an invaluable resource for our experiments. Such a strategy requires cooperative pooling of information and free sharing such that individual results can be readily compared against validated datasets. Our future experiments will be directed and developing additional data sets of ES cell differentiation, which can be shared in a manner similar to the present set. Conclusions =========== Our results provide a comprehensive data set that can be effectively utilized to analyse expression patterns of known and unknown genes. Comparison with other data sets provides independent confirmation of results and shows a high level of concordance. The caveats to all such large-scale comparisons are discussed and the importance of pooling data and comparing across multiple data sets is demonstrated. Methods ======= Cell culture ------------ The human ES cell lines H1, H7, and H9 were maintained under feeder-free conditions in MEF-conditioned medium supplemented with bFGF as described previously \[[@B19],[@B26]\]. MPSS ---- MPSS was performed using RNA from three pooled ES cell lines (H1, H7, and H9) that had been maintained in feeder free culture conditions and evaluated for the presence of ES cell markers and absence of markers of differentiation. The mRNA was converted to cDNA and digested with DpnII. The last DpnII site and the downstream 16 bases were cloned into Megaclone vectors and their sequences determined according to the MPSS protocol \[[@B15],[@B16],[@B25]\]. A total of 2.786.765 sequences were read from four different runs and 48,388 unique signatures were identified. The abundance for each signature was converted to transcripts per million (tpm) for the purpose of comparison between samples. Signatures at an abundance of less than 4 tpm or those that were not detected in at least two runs were removed and a total of 22,136 sequences were analyzed further. All data is available for download from Lynx \[[@B27]\] MPSS signature classification and annotation -------------------------------------------- To generate a complete, annotated human signature database, we extracted all the possible signatures (\"*virtual signatures*\") from the human genome sequence, the human Unigene sequences, and human mitochondrion. Each virtual signature was ranked, as outlined in the table [1a](#T1){ref-type="table"}, based on its position and orientation in the original sequence. Unigene, genomic, and mitochondrial hits were combined and grouped by signature. The annotation was then assigned to the signature in following order of preference: repeat warnings (signature hits more than 100 genome locations); mitochondrial hits; Unigene hits; genome hits (if no transcript match found). If a signature matched only one Unigene cluster, the MPSS signature class is the lowest class of the member sequences of the cluster. If a signature hits multiple Unigene clusters, the best cluster hit is selected based on the lowest MPSS signature class or the largest number of member sequences. The resulting signature database was used to annotate the data from the experiments Initially the signatures were annotated using genome version hg15 (April 2003, Golden Path, UCSC,) and Unigene build \#161 ([additional file 2](#S2){ref-type="supplementary-material"}). Recently we re-annotated all signatures using genome version hg16 (July 2003, Golden Path, UCSC) and Unigene build \#169 ([additional file 3](#S3){ref-type="supplementary-material"}). Both annotations are available for download in supplemental tables \[[@B27]\]. Microarray ---------- Analysis was performed as described in Bhattacharya et al., \[[@B9]\] using six different samples. These included two lines from Bresagen (01 and 02), the pooled sample from Geron comprising feeder free subclones of (H1, H7, H9), H1, grown in our laboratory on feeders and H9 and I6 from Dr. Itskovitz-Eldor grown following their published protocols. EST-enumeration --------------- EST frequency counts of genes expressed in human ES cells were done as described (\[[@B8]\]). Statistical significance was determined using the Fisher Exact Test \[[@B28]\]. Chromosomal mapping of MPSS signatures and UniGene clusters to the human genome ------------------------------------------------------------------------------- MPSS signatures with a hit to a UniGene cluster were mapped to the Giemsa staining cytobands of the hg15 release of the human genome (April 10, 2003 freeze, \[[@B29]\]). By this method, 7731 MPSS signatures were mapped to the cytobands of the human genome. Similar mapping was done for all UniGene clusters for which the chromosomal mapping is known. In order to achieve a gene-based rather than a transcript (i.e. splice variant) based distribution of genes splice variants the UniGene clusters were filtered using LocusLink data \[[@B30]\], since LocusLink captures all characterized splice variants of a particular gene. 23,828 UniGene clusters were identified by this method and mapped to the cytobands of the human genome. To discover differences in the number of genes mapped to each cytoband, the number of genes mapped to each cytoband was compared to the total number of genes analyzed, for both the MPSS signatures as well as for the UniGene clusters. The Fisher test \[[@B28]\] was used to determine the statistical significance, using a p-value = 0.05 as cutoff. Gene detection by RT-PCR ------------------------ Total RNA was isolated from cell pellets using RNAeasy Qiagen mini protocol and kit. cDNA was synthesized using 100 ng of total RNA in a 20-μl reaction. Superscript II (Gibco-BRL), a modified Maloney murine leukemia virus RT, and Oligo (dT)12--18 primers were used according to the manufacturer\'s instructions (Gibco-BRL). The list of primers used for RT-PCR and annealing conditions are described previously \[[@B3]\]). Authors\' contributions ======================= RB, IK and MR were primarily responsible for the data analysis and writing and editing the manuscript. ST generated the ES cell samples and verified their quality, TM performed the RT-PCR and JC performed the immunocytochemistry. RP generated the microarray data and TV and JL provided support and supervision for RB, IK, ST. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 The document describing details of MPSS analysis of the HuES cells performed at Lynx, and algorithm for initial MPSS signature annotation and classification. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 The file contains MPSS data for 22,136 significant and reliable signatures, annotated with genome version hg15 (April 2003, Golden Path, UCSC) and the human Unigene build Hs.161. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 3 The file contains MPSS data for 22,136 significant and reliable signatures, annotated with genome version hg16 (July 2003, Golden Path, UCSC) and the human Unigene build Hs.169. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 4 table containing data for 8679 unigene clusters, 11 mitochondrial genes, and including 1991 signatures that did not map to unigene but uniquely matched genomic sequences (potential novel transcripts). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 5 Top 200 rows from the table HuES17\_onetpmHs169hg16.xls. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 16 List of tpm of genes identified by Sperger et al \[11\] present in the MPSS dataset. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 17 List of tpm of genes identified by Richards et al \[8\] present in the MPSS dataset. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 18 List of tpm of genes identified by MPSS in the BG02 dataset. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 6 This file lists the genes located on the X and Y chromosome for which a MPSS signature sequence was found in the HuES cells. The X and Y chromosome genes are listed in separate worksheets. Chromosome: X or Y chromosome Cytoband: Giemsa staining cytoband Signature: MPSS signature sequence Tpm: mean abundance for a signature derived from all MPSS runs for the sample, in transcripts per million Seq\_id: Repeat if hit genome more than 100 locations; UniGene cluster ID if hit one or more UniGene clusters; Chromosome number if hit one genome location; MultiGenome if hit genome multiple times; Description: description of annotation (see Appendix B for more details). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 7 List of all UniGene clusters (release 161) located on the X and Y chromosome. The X and Y chromosome genes are listed in separate worksheets. Chromosome: X or Y chromosome Cytoband: Giemsa staining cytoband Locusid: Locus Link identifier UniGene cluster: UniGene cluster ID Description: description derived from UniGene ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 8 Graphical representation of hotspots and coldspots for all human chromosomes. Cytobands with significantly higher number of genes in HuES cells are marked with \'+\' and cytobands with lower number of genes in HuES cells are marked with \'-\'. The number of \'+\' or \'-\' signs is proportional to the difference, with one \'+\' or \'-\' representing the fold difference. For example \'++\' represents a 2-fold higher number of genes expressed in HuES cells compared to the expected number based on the number of genes known to be located in the same cytoband. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 9 List of all chromosal regions that appeared statistically distinct. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 10 List of genes expressed in HuES cells and located in the hotspots and coldspots. Chromosome: X or Y chromosome Cytoband: Giemsa staining cytoband Signature: MPSS signature sequence Tpm: mean abundance for a signature derived from all MPSS runs for the sample, in transcripts per million Seq\_id: Repeat if hit genome more than 100 locations; UniGene cluster ID if hit one or more UniGene clusters; Chromosome number if hit one genome location; MultiGenome if hit genome multiple times; Description: description of annotation. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 11 List of all UniGene clusters located in the hotspot and coldspot regions. Chromosome: X or Y chromosome Cytoband: Giemsa staining cytoband UniGene ID: UniGene cluster ID Locusid: Locus Link identifier Description: description derived from UniGene; ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 12 This file contains the mapping information of MPSS signatures to the EST scan genes for major signaling pathways. Each pathway is presented in a separate worksheet Columns: Locusid: Locus Link identifier Pathway: Pathway the gene has been assigned to. Signature: MPSS signature sequence Tpm: mean abundance for a signature derived from all MPSS runs for the sample, in transcripts per million Stdev: standard deviation of the mean abundance from multiple MPSS runs Hit genome: HitGenome -- Numbers of genomic locations a signature maps to HitUniGene: Numbers of UniGene Clusters a signature maps to; Seq\_id: UniGene cluster ID (\'Hs.\' Has been omitted). NULL if there is no MPPS signature but an EST from the EST scan. Class: signature class (see Appendix B for more details); Title: description, derived from UniGene ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 13 This file contains additional lists of potential homeodomain proteins, genes of unknown function, zinc finger proteins in separate worksheets. Tpm: mean abundance for a signature derived from all MPSS runs for the sample, in transcripts per million Stdev: standard deviation of the mean abundance from multiple MPSS runs Hit genome: HitGenome -- Numbers of genomic locations a signature maps to HitUniGene: Numbers of UniGene Clusters a signature maps to; Seq\_id: UniGene cluster ID. Class: signature class (see Appendix B for more details); Title: description, derived from UniGene. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 14 List of tpm of potential 400 ES enriched genes in other cell lines examined. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 15 List of all tpm of potential 13 ES novel genes in all tissues and cell lines examined. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ This work was supported by the NIA and an ALS center grant. JC, TM, MR were supported by the NIA. RP was supported by the FDA. ST, RB, JL are employees of Geron Inc. IK and TV are employees of Lynx Inc. We thank Drs. Ginis and Limke for careful manuscript reading and all members of our laboratories for constant stimulating discussions. MSR acknowledges the contributions of Dr. S. Rao that made undertaking this project possible.

PubMed Central

2024-06-05T03:55:47.196163

2004-8-10

PMC514703

Background ========== Aromatic L-amino acid decarboxylase (E.C. 4.1.1.28, AADC) catalyzes the second enzymatic step in synthesis of the neurotransmitters dopamine and serotonin, which are found in neurons of all animals (Figure [1](#F1){ref-type="fig"}). Alteration in the normal expression of these transmitters is associated with human neurological disorders such as Parkinson\'s disease and depression \[[@B1],[@B2]\]. In mammals, AADC is expressed in many tissues beside the nervous system, associated with additional regulatory roles of dopamine and serotonin in a wide range of tissues \[[@B3]\]. In insects, AADC is further required to produce amines for cuticle synthesis and pigmentation \[[@B4]\]. Because of its role in the synthesis of both transmitters, by decarboxylation of L-dopa and 5-hydroxytryptophan, AADC is also known as dopa decarboxylase or 5-hydroxytryptophan decarboxylase (reviewed in \[[@B3]\]). AADC belongs to the α family (subgroup II) of pyridoxal-5\'-phosphate (PLP) dependent enzymes. Other subgroup II enzymes include histidine, tyrosine, tryptophan and glutamate decarboxylases \[[@B5]\]; in animals some of these enzymes mediate synthesis of other biogenic amines (e.g., histamine, tyramine, octopamine) and GABA. In mammals and in *Drosophila*, a single gene encodes the serotonin- and dopamine-synthetic AADC \[[@B6],[@B7]\], although tissue-specific isoforms of the protein are generated by alternative splicing \[[@B8],[@B9]\]. Different genes encode PLP-dependent decarboxylase enzymes for histamine, octopamine and GABA synthesis. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Serotonin and dopamine biosynthetic pathways.**Serotonin and dopamine are synthesized from the aromatic amino acids tryptophan and tyrosine, respectively. The first and rate-limiting step in synthesis is carried out by a neurotransmitter-specific aromatic amino acid hydroxylase enzyme, either tryptophan or tyrosine hydroxylase. In *C. elegans*, these genes are encoded by the *tph-1*and *cat-2*genes, respectively \[19, 25\]. The second synthetic step for both neurotransmitters shares the aromatic L-amino acid decarboxylase (AADC) enzyme, which has a relatively broad substrate specificity, and is also known as 5-hydroxytryptophan decarboxylase or dopa decarboxylase. :::  ::: In the nematode *Caenorhabditis elegans*, serotonin is expressed in at least nine neurons in the hermaphrodite and nineteen in the male; dopamine is found in eight neurons in the hermaphrodite and fourteen in the male \[[@B10]\]. By examining the behavior of worms in which specific neurons have been ablated and examining mutants lacking serotonin and/or dopamine, we have learned that serotonin is involved in behaviors including egg laying \[[@B11]-[@B13]\], pharyngeal pumping \[[@B14],[@B15]\], male mating \[[@B16]\], and experience-dependent regulation of locomotion \[[@B17],[@B18]\]. Serotonin-deficient mutants also display abnormalities in entry into the diapause-like dauer stage and in fat storage, mediated via an insulin-related signaling pathway \[[@B19],[@B20]\]. Dopamine plays roles in male mating \[[@B21]\], in regulating locomotion via mechanosensation \[[@B17],[@B22]\], and in foraging behavior \[[@B23]\]. Identification of genes involved in neurotransmitter synthesis and related aspects of signaling in *C. elegans*was greatly accelerated by genomic sequencing, which was essentially completed in 1998 \[[@B23],[@B24]\]. For genes identified originally by mutants via a traditional genetic approach, a candidate gene approach often allowed rapid confirmation of a gene\'s identity; for predicted genes identified from the genomic sequence by homology, a reverse genetic approach has been taken. Many components of the serotonin and dopamine synthesis and transport pathways in *C. elegans*have now been identified by these traditional and reverse genetic approaches, including tyrosine hydroxylase (*cat-2*; \[[@B25]\]), tryptophan hydroxylase (*tph-1*; \[[@B19]\]), serotonin reuptake transporter (*mod-5*; \[[@B26]\]), dopamine reuptake transporter (*dat-1*; \[[@B27]\]) and vesicular monoamine transporter (*cat-1*; \[[@B28]\]). Postsynaptic components have also been identified, including various receptors \[[@B29]-[@B32]\] and intracellular G protein signaling components \[[@B33]-[@B36]\]. Further analysis of gene function, regulation and evolution in *C. elegans*is being facilitated by genomic sequencing of related nematodes. A whole genome shotgun sequence of *Caenorhabditis briggsae*was recently completed; the sequence is estimated to be 98% complete \[[@B37]\]. The divergence of *C. briggsae*and *C. elegans*is estimated between 80 -- 110 million years ago \[[@B37],[@B38]\], although it should be noted that these estimates lack a fossil record to anchor the dates \[[@B39]\]. This is considered to be a favorable evolutionary distance to identify conserved non-coding regulatory sequences, although the sequences from only two orthologous genes from related species is often inadequate to identify such sequences unambiguously. Genomic sequencing is planned or underway of three additional congeneric relatives of *C. elegans*that are more closely related than *C. briggsae*, which will enhance our ability to analyze the genes of *C. elegans*. We have used genomic sequences of both *C. elegans*and *C. briggsae*to help identify and characterize another component of the serotonin and dopamine signaling systems -- the *bas-1*gene -- and to examine the evolution of this and related genes. The *bas-1*\[*b*iogenic *a*mine *s*ynthesis abnormal\] mutant is serotonin- and dopamine-deficient, and displays several behavioral abnormalities \[[@B12],[@B16],[@B17]\]. Unlike wildtype and other serotonin-deficient mutant worms, *bas-1*mutants are unable to convert exogenous 5-hydroxytryptophan (5HTP) into serotonin (5-hydroxytryptamine, 5HT), as assessed by serotonin antiserum staining. Because of this phenotype, we have previously proposed that the *bas-1*gene likely encoded the AADC enzyme of *C. elegans*\[[@B16]\]. Results ======= Rescue of the *bas-1*mutant with an AADC-homologous sequence ------------------------------------------------------------ The *bas-1*gene maps to chromosome III, between *dpy-17*and *unc-32*. When this region was sequenced by the *C. elegans*Genome Sequencing Consortium, two AADC-homologous predicted genes, designated C05D2.4 and C05D2.3, were found to be located close together on a single cosmid, C05D2 (Fig. [2A](#F2){ref-type="fig"}). This suggested that one (or both) of these sequences comprised the gene mutated in *bas-1*mutants. To test this hypothesis, we injected *bas-1*mutants with the cosmid C05D2 plus *rol-6 (dom)*plasmid DNA as a co-injection marker. We isolated transgenic Roller progeny (expressing the *rol-6 (dom)*marker phenotype) of the injected worm and propagated strains that transmitted the marker, then tested these worm strains using serotonin antibody staining. We found that 3 of 3 independent Roller transgenic lines were rescued for serotonin immunoreactivity, confirming that the *bas-1*gene was located within this 46 kb of genomic DNA (Fig. [2B](#F2){ref-type="fig"}). We then injected plasmid subclones of C05D2, each of which still contained both the predicted C05D2.4 and C05D2.3 genes. A 15.1 kb plasmid subclone (C05D2XN) also rescued *bas-1*mutants (n = 4/4), as did smaller subclones of C05D2XN, including an 11.3 kb subclone (pCL3001, n = 11/11) and an 8.8 kb subclone (pCL7001, n = 1/1). These results confirm that at least one of AADC-homologous genes likely corresponds to the *bas-1*gene. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Molecular genetics and transformation rescue of *bas-1.***(A) Genetic and physical map of *bas-1*region and *bas-1*mutant alleles. Locations and extent of mutations for each *bas-1*allele are shown to scale with respect to C05D2.4 and C05D2.3 coding sequences, based on known splicing patterns or Genefinder predictions. Exons are indicated by rectangular bars; an alternatively spliced 27 bp exon is also indicated after exon 2 in C05D2.4. Four of five *bas-1*mutants affect only C05D2.4; *ad446*is a larger deletion removing most coding sequence of both C05D2.4 and C05D2.3. (B) Genomic DNA constructs that rescue or fail to rescue *bas-1*mutants. Constructs are shown to scale (top), based on the 15.1 kbp insert of the plasmid clone C05D2XN; construct names are indicated in the box on the left. The cosmid C05D2 is larger, as indicated by the arrows. Clones below are subclones or modifications of C05D2XN. Coding regions for the two predicted AADC genes C05D2.4 and C05D2.3 are indicated by the blue boxes; intergenic regions are shown in yellow. C05D2XN upstream of C05D2.4 contains two other predicted genes, one complete (C05D2.8) and one partial (C05D2.5). There are no predicted genes in the 3 kb downstream of C05D2.3. In the constructs with the least upstream sequence, only a portion of C05D2.8 remains. Constructs mutated to introduce premature stop codons are indicated with a X in the coding sequence, and red downstream of the introduced stop codon. The construct pCL8001 has a GFP gene inserted in a manner that would inactivate the C05D2.3 gene, so is comparable to the pCL7991 construct. \[No GFP expression was seen in the C05D2.3::GFP reporter construct lines.\] :::  ::: To determine which of the two predicted AADC sequences was needed to rescue *bas-1*mutants, we prepared two constructs from C05D2XN, one mutated in C05D2.4, the other in C05D2.3 (Fig. [2B](#F2){ref-type="fig"}). In each case, a mutation was created by eliminating a unique restriction site early in the predicted coding region, creating a frameshift resulting in premature stop codons. We found that constructs mutated in C05D2.3 when injected rescued serotonin immunoreactivity in *bas-1*mutants (n = 3/3), whereas the construct mutated in C05D2.4 failed to rescue (n = 0/5 rescued). A construct containing a GFP gene inserted into the C05D2.3 coding sequence (and disrupting the gene) also rescued *bas-1*mutants (n = 2/2). In Roller transgenic lines lacking rescue, we confirmed the presence of the injected construct by PCR. Therefore, an intact C05D2.4 gene is necessary to rescue *bas-1*mutants, whereas the C05D2.3 gene is not. In all rescued transgenic lines, we saw the complete set of known serotonergic neurons, although not necessarily all cells in every animal -- mosaicism from somatic loss of extrachromosomal DNA is expected in these transgenics. This result suggests that no critical cell-specific regulatory sequences were missing from even the smallest construct we injected. To confirm further that C05D2.4 is the *bas-1*gene, we identified the mutations in four *bas-1*mutant alleles; we also examined the phenotypes of deletion mutants in C05D2.4 and C05D2.3 generated by the *C. elegans*Gene Knockout Consortium (GKC). We found that the *bas-1*alleles *pa4*, *n2948*, and *n3008*contained point mutations in C05D2.4 coding sequence resulting in premature stop codons (Fig. [2A](#F2){ref-type="fig"}). We found that the original *bas-1*allele (*ad446*) had a 4268 bp deletion from the second exon of C05D2.4 to the final intron of C05D2.3; therefore, *ad446*is a knockout of both predicted genes. We examined the phenotypes of GKC-generated deletion mutants in each predicted gene. The C05D2.4 knockout (*tm351*) removes the entire predicted second exon. We found that both *tm351*homozygotes and *tm351/ad446*worms were deficient in serotonin immunoreactivity. On the other hand, a knockout of C05D2.3 (*ok703*) is wildtype for serotonin staining. Therefore, *tm351*is a fifth mutant allele of the *bas-1*gene, and C05D2.4 corresponds to the gene *bas-1*. Transcripts from the *bas-1*gene and the predicted BAS-1 protein ---------------------------------------------------------------- To continue our characterization of the *bas-1*gene, we isolated cDNAs using RT-PCR; we also obtained cDNA clones from the *C. elegans*EST/Transcriptome project (courtesy of Yuki Kohara) and the ORFeome project \[[@B40]\]. We found that C05D2.4/*bas-1*cDNAs are trans-spliced to SL1 just 15 nucleotides upstream of the predicted translation start site. The consensus sequence from our clones and others we examined predicts a 514 amino acid, 58 kDa protein product (Fig. [3A](#F3){ref-type="fig"}). This is similar in size to other known AADC/dopa decarboxylase proteins such as those of *Drosophila*(510 aa) and human (480 aa). The predicted protein possesses a conserved lysine PLP binding site at residue 343, and has other amino acids identical to those shown to be essential for rat AADC function \[[@B5],[@B41],[@B42]\]. A number of possible phosphorylation sites can be predicted, including three serines and one tyrosine that are conserved in all known AADCs and HisDCs (Fig. [3A](#F3){ref-type="fig"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### ***C. elegans bas-1*cDNA sequences.**(A) Consensus cDNA sequence and translation for C05D2.4/*bas-1*, based on the most common splice form. Nucleotide numbering is shown on the left side and amino acid numbering on the right side of the sequence. SL1 spliced leader sequence is overlined in blue in the top line. Intron locations are indicated with blue arrowheads; the phase at all intron locations is 0 (between codons; see also Fig 6). The conserved lysine (K) pyridoxal 5-phosphate binding site at amino acid 343 is boxed in black. Red amino acids in the predicted Bas-1 protein (T286, D292, H309, D311, S336, K343, K357, V378, R393, and W401) are identical to those shown to be essential for rat AADC function \[5, 41, 42\]. Possible phosphorylation sites that are absolutely conserved in known DDC and HisDC proteins are boxed in green (Y37, S149, S229, S230). The polyadenylation signal in the final line is underlined. Mutations found in bas-1 alleles are indicated with the allele designation and the changed base over the wildtype sequence. The allele tm351 deletion, which removes the entire second exon, is indicated by a red line over the missing sequence. The wildtype cDNA sequence shown is consistent with our RT-PCR clones (primers SL1B, C05D2-B), those we sequenced from the ORFeome project (from predicted translation start to stop), and *C. elegans*EST project \'YK clones\' ends (used to determine the 3\' end, including the site of polyadenylation). (B) Alternatively spliced 27 bp exon and surrounding genomic sequence in *C. elegans*and *C. briggsae*. The additional exon is found in a fraction of *C. elegans bas-1*transcripts, and the sequence is conserved in genomic sequence from *C. briggsae*as shown. \[We did not isolate a cDNA containing this exon among our *C. briggsae bas-1*cDNAs, S. DePaul & C. Loer, unpublished results.\] Predicted translation of the exon is shown above or below the nucleotide sequence. Consensus splice signals are overlined in blue, and identical nucleotides are indicated by vertical lines between the two nucleotide sequences. :::  ::: We found two splice variants different from the Genefinder-predicted cDNA described above, which was the predominant form. About 20% of clones we sequenced had a 27 bp microexon inserted between the predicted exons 2 and 3 (Fig. [3B](#F3){ref-type="fig"}). The 27 bp microexon is found within what is the second intron in the more commmon splice form. This intron is not conserved among other AADCs, and is inserted within a region of the BAS-1 protein that is not conserved among AADC proteins. Modeling of BAS-1 protein structure, based on a recent crystal structure of porcine DDC \[[@B43]\], indicates that this region is located at the surface of the protein where it would not interfere with the conserved enzymatic function of the protein (data not shown). We observed that this additional exon is conserved in the *C. briggsae*ortholog of *bas-1*in genomic sequence (Fig. [3B](#F3){ref-type="fig"}), although we did not isolate any splice variants with this exon among *C. briggsae bas-1*cDNAs we sequenced (see also below). We found a single clone that used an alternative splice acceptor 60 bp upstream of the usual splice site for exon 3; this alternate splice introduces a premature stop codon in the coding sequence. This transcript may be a rare, aberrant splice form without functional significance. Expression of a *bas-1*::GFP reporter fusion in transgenic worms ---------------------------------------------------------------- We examined the pattern of expression of a GFP reporter construct with \~4500 bp upstream of the predicted *bas-1*translation start site and an in-frame fusion with the 2nd exon, injected with *rol-6(dom)*plasmid into wild-type worms (kindly provided by Ian Hope). Two independent transgenic Roller lines with extrachromosomal arrays had the same pattern of expression. The reporter was reliably expressed in several easily identified cells including the paired serotonergic neurons NSM and HSN and the dopaminergic PDE postdeirid sensory neurons (Fig. [4](#F4){ref-type="fig"}). NSM processes studded with varicosities were apparent in the isthmus of the pharynx labeled with GFP (Fig. [4A,4D](#F4){ref-type="fig"}). The egg-laying neuron HSN normally expresses serotonin only in adulthood, and we found the reporter to be expressed in adult hermaphrodites and sometimes late L4 larvae. Often the HSN processes were apparent extending to vulval muscles and anteriorly within the ventral nerve cord (Fig. [4C,4F](#F4){ref-type="fig"}). We saw a cell we identified as PDE, which is born during L2, only after this stage. In some worms, we saw a PDE process and dendrite, confirming our identification (Fig. [4F](#F4){ref-type="fig"}). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Expression pattern of a *bas-1*::GFP reporter fusion in transgenic Roller worms.**(Panels A-C are from the same adult hermaphrodite. Ventral is down and anterior to the right.) A. Ventral, slightly oblique view of the head, showing NSMs, CEPDs, ADEL and likely AIMs. B. Same head, higher (more dorsal) focal plane, showing CEPDs and ADER. C. Photomontage showing ventral oblique view of HSNs and their processes in the ventral nerve cord; note also apparent labeling of muscles associated with the vulva. A second worm is immediately adjacent above, obscuring the edge of the worm shown. (Panels D-F: Anterior is to the left.) D. Adult hermaphrodite head, ventral view, chosen to show the characteristic highly varicose processes of the NSM cells within the isthmus of the pharynx. E. Larval head, ventral view with fluorescence and brightfield. This clearly shows the location of the NSM somata in the ventral pharynx, anterior bulb; it also shows the serotonergic ADF neurons not seen in A, B. CEPDs would be seen in a dorsal focal plane in this worm. F. Adult hermaphrodite lateral view of body wall. Ventral is down. Shows HSN and PDE; note PDE process extending ventrally toward the ventral nerve cord and dendrite extending dorsally into postdeirid sensillum. Twisting of the body axis associated with Roller phenotype makes HSN and PDE somata appear at the same lateral level when HSN is actually located sublateral and PDE subdorsal; twisting also takes ventral nerve cord out of plane of focus in the right of the panel. (Panels G -- I are from males; anterior is to the right.) G. Late L4 male tail showing ray neurons (RNs) with processes extending into the rays. In some males we saw spicule cell staining likely belonging to spicule socket cells (SpSo). Ventral, slightly oblique view. H. Adult male tail showing RNs and their neurites in rays 7 and 9 on the right side, view ventral, slightly oblique. I. Male-specific ventral nerve cord motoneurons CP5 and CP6, the CP neurons most commonly expressing the transgene. The PDE soma in the lateral body wall is out of the plane of focus. :::  ::: The *bas-1*::GFP reporter was also expressed in other neurons in the head, around the nerve ring. We believe that all of these cells are known serotonergic and dopaminergic neurons. It was somewhat more difficult, however, to be certain about these identifications since we saw few processes, and even when present we could not always unambiguously associate a process with a particular neuronal soma. Nevertheless, the reporter was expressed in probable dorsal and ventral cephalic sensilla neurons CEPD and CEPV; we sometimes observed as many as four processes extending to the tip of the nose (Fig [4A,4B,4E](#F4){ref-type="fig"}). We also saw expression in the anterior deirid sensory neurons ADE (Fig. [4A,4B,4E](#F4){ref-type="fig"}). Less frequently we saw expression in probable ADF and AIM neurons (Fig [4A,4E](#F4){ref-type="fig"}). We saw as many as 12 neurons (6 bilateral pairs) expressing the reporter in the head of young larvae. This includes all the identified serotonergic (NSM, ADF, AIM) and dopaminergic (CEPD, CEPV, ADE) head neurons excepting the unpaired RIH neuron \[[@B10]\]. In a small number of males examined, we saw expression in male-specific serotonergic and dopaminergic neurons, including up to 6 pairs of ray sensory neurons (RNs) in both adults and late L4 larvae (Fig. [4G,4H](#F4){ref-type="fig"}). (There are three pairs of serotonergic, and three pairs of dopaminergic RNs among the 18 RNs.) Expression in CP neurons, male-specific ventral cord motoneurons controlling tail curling during mating, was limited and usually weak in the male worms we examined. Six CP neurons are strongly serotonin-immunoreactive in males \[[@B16]\]. At most we saw three posterior cells staining, and usually only one or two posterior cells (CP5, CP6) weakly stained, when expression was present at all (Fig. [4I](#F4){ref-type="fig"}). We never saw CP staining in L4 animals, and often none even in male worms expressing GFP strongly in the RNs. C05D2.4 (*bas-1*) and its downstream homolog C05D2.3 ---------------------------------------------------- Just downstream of the *bas-1*/C05D2.4 gene is C05D2.3, the product of an ancient tandem duplication event. The two genes have diverged considerably -- being only 59% identical at the amino acid level (Table [1](#T1){ref-type="table"}). The genomic structures of the two genes have also diverged. The two genes share four introns, but C05D2.4 has one and C05D2.3 has three introns not found in the other (Fig. [7](#F7){ref-type="fig"}). Nevertheless, comparisons with other AADC proteins showed that *bas-1*/C05D2.4 is most similar to C05D2.3 and the predicted gene F12A10.3 (Fig. [5](#F5){ref-type="fig"}, Table [1](#T1){ref-type="table"}). The predicted amino acid sequence of C05D2.3 contains one noteworthy gap: it is missing six amino acids from a highly conserved region found in all other PLP-dependent decarboxylases. This sequence, the consensus of which is VDAAYA, contains an aspartate (D) residue that is absolutely essential for function of Rat DDC. Substitution of an alanine or asparagine completely abolishes enzymatic activity, and even the conservative substitution of a glutamate at this site reduces activity to 2% of wildtype \[[@B5]\]. It is therefore unlikely that a C05D2.3 protein could function enzymatically as a typical AADC. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Pairwise BLAST comparisons with *C. elegans*AADCs. ::: -------------------------------------------------------------------------------------------------------------- C05D2.4 C05D2.3 F12A10.3\* K01C8.3 ZK829.2 C09G9.4 *Ce*GAD ------------ ------------- ------------- ------------- ------------- ------------- ------------- ------------- C05D2.4\ Score\ \- \- \- \- \- \- (*bas-1*) %Id / %Sim C05D2.3 1674\ \- \- \- \- \- \- 59 / 75 F12A10.3\* 1756\ 1793\ \- \- \- \- \- 60 / 77 63 / 78 K01C8.3\ 970\ 844\ 733\ \- \- \- \- (*tdc-1*) 37 / 57 34 / 54 33 / 51 ZK829.2 583\ 541\ 559\ 1147\ \- \- \- 29 / 47 27 / 46 27 / 49 44 / 66 C09G9.4 224\ 150\ 190\ 272\ 275\ \- \- 22 / 40 18 / 38 19 / 40 22 / 42 23 / 44 *Ce*GAD\ 210\ 164\ 216\ 299\ 250\ 129\ \- (*unc-25*) 22 / 38 20 / 36 23 / 37 25 / 43 24 / 41 20 / 40 *Dm*DDC **1095**\ 909\ **984**\ 1390\ 883\ 256\ 328\ **41 / 60** 36 / 56 **38 / 58** 50 / 69 37 / 57 22 / 42 24 / 40 *Hs*DDC 1067\ **912**\ 949\ 1458\ 935\ 233\ 322\ 41 / 60 **36 / 55** 38 / 58 55 / 73 39 / 59 20 / 42 26 / 44 *Dm*HisDC 996\ 816\ 807\ 1348\ 910\ **278**\ 339\ 39 / 57 33 / 53 32 / 54 52 / 70 40 / 59 **23 / 42** 26 / 44 *Hs*HisDC 988\ 802\ 876\ 1290\ 920\ 231\ 306\ 38 / 56 33 / 53 34 / 56 48 / 69 39 / 59 21 / 42 26 / 42 *Dm*G30446 971\ 845\ 686\ **1671**\ **1008**\ 254\ n.s. 38 / 57 34 / 53 30 / 50 **64 / 77** **40 / 62** 20 / 44 *Dm*AMD 961\ 799\ 668\ 1124\ 741\ 201\ 278\ 38 / 57 35 / 53 36 / 53 44 / 63 33 / 51 20 / 40 25 / 41 *Dm*G30445 837\ 797\ 814\ 1407\ 905\ 259\ n.s. 34 / 52 35 / 53 33 / 53 55 / 71 39 / 57 22 / 42 Cr TrpDC 743\ 674\ 735\ 916\ 269\ 272\ 333\ 30 / 51 28 / 50 30 / 52 37 / 59 31 / 51 24 / 43 25 / 42 *Dm*GAD 226\ 207\ 232\ 374\ 226\ 99\ 1146\ 22 / 39 21 / 38 22 / 41 27 / 43 22 / 41 25 / 55 44 / 64 *Hs*GAD67 215\ 173\ 192\ 324\ 228\ 86\ **1535**\ 22 / 36 19 / 36 21 / 36 24 / 44 24 / 41 16 / 41 **56 / 73** -------------------------------------------------------------------------------------------------------------- Comparisons of *bas-1*/C05D2.4 and other pyridoxal-phosphate dependent decarboxylase amino acid sequences were made using \"BLAST 2 Sequences\" \[version 2.2.6, <http://www.ncbi.nlm.nih.gov/blast/bl2seq/bl2.html> \[73\]; Settings (largely default): Matrix -- BLOSUM62, Open gap penalty -- 11, extension gap penalty -- 1, low complexity filtering -- OFF). As shown in the table above, on the top line, each comparison shows the blast score; below is the percent identity and percent similarity for the \'alignable\' sequence. The highest scoring match (excluding among *C. elegans*AADCs) is indicated in bold. Sequences in the left column are arranged in order of blast score in comparison to C05D2.4 *C. elegans*AADCs are indicated by their predicted gene designation: C05D2.4, C05D2.3, F12A10.3, K01C8.3, ZK829.2 and C09G9.4. Abbreviations: *Ce*-- *C. elegans*, *Cr*-- *Caranthus roseus*(periwinkle plant), *Dm*-- *Drosophila melanogaster*, *Hs*-- *Homo sapiens*, DDC -- dopa decarboxylase, HisDC -- histidine decarboxylase, AMD -- Alpha-methyl dopa hypersensitive protein, TrpDC -- tryptophan decarboxylase, GAD -- glutamate decarboxylase. \*An amino acid sequence for F12A10.3 was generated from cDNA sequence by introducing 2 frameshifts to preserve AADC homology in the predicted aa sequence merely for sake of comparison to other AADC\'s (see text); this sequence is different from predicted sequences in found in Genbank and Wormbase which are based on incorrect cDNA predictions. ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Alignments of AADC protein sequences with *C. elegans*BAS-1 predicted protein.**Gaps are indicated with a *dash*(-); at the beginning or end of a sequence, *periods*indicate additional sequence upstream or downstream that is not shown. Alignments were performed with CLUSTALW. Abbreviations for species and gene names are the same as listed in the legend for Table 1. For genes with multiple splice forms, the most readily aligned sequence was chosen. Red shading indicates amino acids are identical in ≥ 90% of the aligned sequences. Yellow shading indicates similar amino acids found in that position in ≥ 90% of the aligned sequences. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Phylogenetic trees of AADC protein and nucleotide sequences.**Trees were made from sequences aligned with CLUSTALW. Species and gene names are abbreviated as listed in the legend for Table 1. (A) The single minimum-length tree resulting from a heuristic search using parsimony from alignments of core protein sequences (531 characters) of selected *C. elegans*, *C. briggsae*, Human and *Drosophila*AADCs. *C. roseus*(periwinkle plant) TrpDC was used as an outgroup. Branch lengths are indicated, with bootstrap values using the same search conditions (1000 replicates) in parentheses. The search used the tree-bisection-reconnection (TBR) branch-swapping algorithm; characters were equally weighted. An identical tree topologically was obtained by a branch-and-bound search. *C. elegans*F12A10.3 was excluded from this analysis since it lacks a functional protein sequence (see Fig. 7 and text). Trees determined by distance methods were similar, but rearranged some of branches with low bootstrap values in the tree shown. (B) The single minimum-length tree resulting from a heuristic search using parsimony (same settings as above) of nucleotide sequence alignments (1608 characters) from a subset of AADCs above, with the addition of *C. elegans*F12A10.3. *Dm*DDC was used as an outgroup. Branch lengths and bootstrap values using the same search conditions (1000 replicates) are shown as in A. An identical tree topologically was obtained by a branch-and-bound search. :::  ::: Because the *bas-1*and C05D2.3 genes are so close together -- only 369 bp from predicted translation stop to predicted translation start -- we considered whether they might be expressed as an operon. In *C. elegans*and other nematodes, genes that are very close together (and often functionally related) may be expressed from a single promoter initially as a single primary transcript \[[@B44]\]. Operon transcripts are subsequently processed to yield separate mRNAs. The first gene in an operon is trans-spliced to the leader sequence SL1; downstream genes are typically spliced to a slightly different leader sequence termed SL2. We would expect to find C05D2.3 transcripts trans-spliced to SL2 if it is a downstream gene in an operon with *bas-1*. We were unable to isolate either SL1 or SL2-spliced transcripts from C05D2.3 by RT-PCR, although we did isolate a partial cDNA using internal primers. DNA microarray experiments suggest the gene is not expressed above background levels, unlike C05D2.4/*bas-1*(Table [2](#T2){ref-type="table"}). Furthermore, a global analysis of expression specifically designed to identify operons did not select C05D2.4 and C05D2.3 as likely members of an operon \[[@B45]\]. Since genes comprising an operon should be expressed at similar levels, these data provide no support for the idea that *bas-1*and C05D2.3 constitute an operon. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### *C. elegans*AADC genes expression and *C. briggsae*orthologs. ::: ***C. elegans*AADC** ***C.e.*cDNAs** ***C.e.*Microarray** ***C. briggsae*ortholog** ***C.b.*cDNAs** ---------------------- ----------------- ---------------------- -------------------------------- ----------------- C05D2.4 (*bas-1*) +(3) ^a,b,c,d^ \+ FPC2187 (84,978 / - strand) \+ ^a^ C05D2.3 \+ ^a,c^ \- none NA C09G9.4 \+ ^d^ \- FPC4079 (\~28,330 / + strand) \- F12A10.3 +(2) ^b,d^ \- none NA K01C8.3 (*tdc-1*) +(2) ^c,d^ \+ FPC0011 (663,153 / + strand) \+ Y37D8A.23 (*unc-25*) +(3) ^c,d^ \+ FPC4030 (765,572 / - strand) \- ZK289.2 \+ ^c,d^ \- FPC0143 (1,747,392 / - strand) \- *C. e.*cDNAs: Parentheses indicate number of splice forms found. ^a^cDNAs found via our RT-PCR experiments and ^b^our sequencing of ORFeome project clones. ^c^*C. elegans*EST project. ^d^Worm ORFeome project. Microarray: Expression levels at all developmental stages as shown by *C. elegans*microarray experiments found in Wormbase. \"+\" indicates significant expression at some stage; \"-\" indicates no expression above background detected at any stage. *C. briggsae*AADC orthologs: We did TBLASTN searches of the *C. briggsae*whole genome shotgun assembly (cb25.agp8) on the Sanger Centre *C. briggsae*blast server using complete predicted amino acid sequences for each *C. elegans*AADC gene. *C. briggsae*genes are designated by contig location, first nucleotide of predicted coding sequence, and strand, based on predicted *C. elegans*sequence. In each case, alignments showed extended regions of 100% or near 100% amino acid identity beginning at the site indicated. (We did not locate the beginning of the *C.b.*C09G9.4 ortholog; alignments did not identify a matching site for the first 11 *C. elegans*amino acids). For C05D2.3 and F12A10.3, the best matches in *C. briggsae*were the C05D2.4 ortholog (FPC2187), followed by the K01C8.3 ortholog (FPC0011); these two genes appear to be absent from *C. briggsae*. We have isolated an SL1-spliced *C. briggsae bas-1*cDNA by RT-PCR (S. DePaul & C. Loer, unpublished); *C. briggsae tdc-1*ESTs are in GenBank. NA -- not applicable. ::: The *bas-1*-AADC and other AADC genes in *C. elegans* ----------------------------------------------------- We compared the predicted amino acid sequences of five other *C. elegans*AADC-like genes revealed by deletion mapping \[[@B46]\] and by whole genomic sequencing \[[@B24]\], along with a previously identified *C. elegans*glutamate decarboxylase (GAD) gene, *unc-25*\[[@B47]\] to related PLP-dependent decarboxylases from other organisms. Some of the *C. elegans*genes are clearly closely related to other AADCs, whereas others are more divergent (Fig [5](#F5){ref-type="fig"}, Table [1](#T1){ref-type="table"}). All contain the core conserved domain (PFAM 00282) defining this group of PLP-dependent decarboxylases. None of the AADC or GAD predicted proteins in *C. elegans*appears to have a signal sequence. The protein predicted from K01C8.3 is now believed to encode a tyrosine decarboxylase (*tdc-1*) used for tyramine and octopamine synthesis, which both appear to be used as neurotransmitters in *C. elegans*\[[@B48],[@B49]\]. The best match to K01C8.3/*tdc-1*is a predicted *Drosophila*AADC-homologous protein of unknown function (G30446). Interestingly, K01C8.3/*tdc-1*shows a stronger match to known DDCs than any of the other *C. elegans*AADCs, including C05D2.4 (Table [1](#T1){ref-type="table"}), although it is equally similar to known histidine decarboxylases (HisDCs). The strongest match of C05D2.4/*bas-1*(outside of nematodes) is to insect and mammalian DDCs, but again the match is only slightly better than to HisDCs. The predicted genes C05D2.3, F12A10.3 and ZK829.2 also have about the same level of identity and similarity to known AADCs and HisDCs. The ZK829.2 predicted protein, however, is much larger (830 AA) than a typical AADC, having extended N- and C-terminal domains not found in other PLP-dependent DCs. Most of ZK829.2 predicted coding sequence is confirmed by cDNA sequences, suggesting that the predicted protein \'extensions\' likely are real. The predicted gene C09G9.4 is the most divergent from known AADC\'s with only 20 -- 24% amino acid identity; it is even more divergent than *C.e.*GAD/*unc-25*. It also appears to lack the absolutely conserved Lys of PLP-dependent decarboxylases, although it otherwise retains considerable homology with the conserved domain of this family of proteins. There are no similar proteins among other organisms to provide clues about a possible function for this gene; C09G9.4 is a truly novel member of the group II PLP-dependent enzyme family. Proteins with a similar level of divergence with AADC (\~20% identity over a few hundred amino acids) include other group II PLP-dependent enzymes such as sphingosine-1-phosphate lyase and cysteine sulfinic acid decarboxylase. C09G9.4, however, has very little or no significant similarity to these other enzymes. The *C. elegans*GAD/*unc-25*predicted protein has a strong match to identified *Drosophila*and mammalian GADs (Table [1](#T1){ref-type="table"}), and is found as a single copy. There are no other GAD-like genes in *C. elegans*such as cysteine sulfinate decarboxylase, which is the rate-limiting enzyme in taurine synthesis, and the closest non-AADC relative to GAD in the vertebrates \[[@B50]\]. Comparison of *C. elegans*and *C. briggsae*AADC genes ----------------------------------------------------- We performed BLAST searches of a *C. briggsae*whole genome shotgun assembly using predicted protein sequences of all six *C. elegans*AADC genes and the *unc-25*/GAD gene. We found five orthologous genes in *C. briggsae*-- four AADC homologs and one GAD homolog (Table [2](#T2){ref-type="table"}). All of these matches included 100% or near 100% identity over extended regions of aligned predicted amino acid sequences, and were paired with high confidence in phylogenies (Fig. [6](#F6){ref-type="fig"}). Using a core AADC sequence for alignments and tree building, we found that the *bas-1*orthologs have evolved more quickly than some of the other AADC\'s. The *C. elegans*gene K01C8.3 and its ortholog, for example, are 98% identical in this core region (vs. 91% identity for *bas-1*orthologs). Most of the divergence between K01C8.3 and its ortholog is in N- and C-terminal extensions that are not found in other AADC\'s. The *C. elegans*C09G9.4 and *C. briggsae*ortholog are even less similar to one another than are the *bas-1*orthologs. ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Genomic structure of *C. elegans*AADC genes compared with Human DDC.**Rectangular blocks represent coding exons of the genes indicated (relative size of exons is approximate). Red triangles indicate ancient conserved introns found in both Human DDC and at least one of the *C. elegans*AADC genes; blue triangles indicate introns conserved among *C. elegans*AADC genes; and open triangles indicate non-conserved splice sites (comparing only among the genes shown). Roman numerals above triangles indicate the phase of the intron. Vertical dashed lines between solid triangles indicate splice sites conserved between at least two genes. Diagonal dashed lines indicate probable conserved sites that are shifted by 2--3 amino acids relative to the other splice site. Alignments of homologous splice sites are based on amino acid multiple alignments of the predicted proteins; insertions and deletions are ignored in the drawing. No alternative splicing is indicated; the most readily \"alignable\" version of each gene was used in cases with multiple splice variants. Dashed boxes at the ends of genes indicate non-AADC homologous extensions unique to the given gene. The most divergent AADC, C09G9.4, is more difficult to align; assignments of splice sites on either side of exon 9 as conserved are more tentative (indicated by question marks). In a few cases where gaps occur in the protein sequence alignments at intron-exon boundaries, introns marked as homologous only begin or end at an homologous location. The splicing pattern shown is fully supported by cDNA sequences for C05D2.4/*bas-1*, F12A10.3, K01C8.3, ZK829.2 ; the pattern is supported by partial cDNA sequences for C05D2.3 and C09G9.4. The extent of supporting cDNA sequence is shown by the heavy black line beneath the colored blocks. F12A10.3 is a special case in that frameshifts (indicated by \'fs\') occur in the cDNA relative to the AADC homologous sequence. The first frameshift occurs at a site where many AADCs are spliced (and where a splice is incorrectly predicted by gene prediction programs), and the second at a splice junction. :::  ::: Our most striking observation is that *C. briggsae*appears to lack orthologs for the *C. elegans*predicted genes C05D2.3 and F12A10.3. This suggests that gene duplications giving rise to these two genes, which are most closely related to *bas-1*/C05D2.4, occurred either in the *C. elegans*lineage after its split with the *C. briggsae*lineage, or that *C. briggsae*lost both C05D2.3 and F12A10.3 orthologs (or their common ancestor) following the split. Using phylogenetic analysis of aligned amino acid and nucleotide sequences, we found that C05D2.3 and F12A10.3 share a common ancestor and that the gene duplication giving rise to *bas-1*and C05D2.3/F12A10.3 likely occurred prior to the *C. elegans/C. briggsae*divergence. This is also suggested by the pattern of introns in the genes. \[We have confirmed the splicing pattern of *C. briggsae bas-1*by isolating a cDNA (S. DePaul & C. Loer, unpublished data).\] The *C. elegans*and *C. briggsae bas-1*genes have identical genomic structure which differs from that of C05D2.3 and F12A10.3, which are more similar to one another (Fig. [7](#F7){ref-type="fig"}). Therefore C05D2.3 and F12A10.3 (or their common ancestor) were retained in the line leading to *C. elegans*but lost in the *C. briggsae*line. The original duplication event giving rise to the tandem copies of C05D2.4 and C05D2.3 on chromosome III probably occurred via an unequal crossing-over or similar event. The duplication creating F12A10.3, which is found on chromosome II, presumably occurred subsequently. We noted no homology of other predicted genes downstream of F12A10.3 and C05D2.3 that might suggest an event duplicating more than the AADC gene. The retention of the genes and their expression in *C. elegans*suggests that they may have acquired a new function that is under selection, retain a subfunction of the AADC, or instead that they are still in the process of being lost. After sequencing F12A10.3 cDNAs (courtesy of the ORFeome Project), we found that current splicing predictions for the gene were incorrect. We sequenced six F12A10.3 clones and found two slightly different splicing patterns, both different from Genefinder and Intronerator predictions. The two types of clones differed only in whether a final intron was removed or not. We found 4 clones with 9 exons, and 2 clones with 8 exons. The failure of the gene prediction programs in this case is likely to be due to their preference for creating functional transcripts. All F12A10.3 cDNAs instead appear to be non-functional: they have frameshifts relative to AADC-homologous reading frames. The first frameshift occurs in the second exon and quickly leads to a premature stop codon. At best F12A10.3 transcripts would result in a 158 amino acid protein that could not function as an AADC. F12A10.3 therefore appears to be an expressed pseudogene. DNA microarray experiments and representation in cDNA sequencing projects suggest that F12A10.3, like C05D2.3, is likely expressed at a low level (Table [2](#T2){ref-type="table"}). In order to assess whether the *bas-1*-like genes C05D2.3 and F12A10.3 might be under reduced selection pressure, we calculated the ratio of non-synonymous to synonymous substitutions (K~A~/K~S~) comparing the *bas-1*orthologs and *bas-1*-like genes. We also calculated these values for the other AADC ortholog pairs from *C. elegans*and *C. briggsae*(Table [3](#T3){ref-type="table"}). K~A~/K~S~\< 1 indicates purifying (negative) selection, K~A~/K~S~= 1 indicates no selection (as in true pseudogenes), and K~A~/K~S~\< 1 indicates Darwinian (positive) selection. K~A~/K~S~for 8179 *C. elegans*and *C. briggsae*ortholog pairs had a mean value of 0.06, indicating most genes are under purifying selection \[[@B37]\]. We found that the *bas-1*genes are under purifying selection (K~A~/K~S~= 0.039), but the *bas-1*-like genes appear to be under reduced selective pressure; the average K~A~/K~S~for comparisons with *bas-1*-like genes was 0.148, more than three times the value of the *bas-1*ortholog comparison. The proportion of observed to potential non-synonymous substitutions (pN) among the *bas-1*-like gene comparisons was similarly much higher than for the *bas-1*orthologs. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Synonymous vs. non-synonymous codon substitution between *C. elegans*and *C. briggsae*AADC orthologs and *bas-1*-like paralogs. ::: ***Ce*, *Cb*AADCs compared** **Codons** **pS** **pN** **K~A~/K~S~** ------------------------------ ------------ ------------- ------------- --------------- *bas-1* 521 0.68 0.07 0.039 C05D2.3, F12A10.3, *bas-1*\* 521 0.71 ± 0.03 0.26 ± 0.02 0.148 ± 0.044 ZK289.2 833 0.71 0.09 0.043 C09G9.4 507 0.74 0.12 0.037 *tdc-1*full length 626 0.79 0.03 NA *tdc-1*core 474 0.82 0.02 NA *tdc-1*N, C terminals 152 0.71 0.07 0.035 Nucleotide alignments of *C. elegans*and *C. briggsae*genes were analyzed by SNAP software (see Methods). Only the *C. elegans*member of the ortholog pair is named. pS = proportion of observed/potential synonymous substitutions; pN = proportion of observed/potential nonsynonymous substitutions. NA -- not applicable (cannot be calculated when pS \> 0.75). \*Includes all pairwise comparisons (n = 5) except *C.e.*vs. *C.b. bas-1*. Values are mean ± SD (strict statistical comparison with other values is not intended, as K~A~/K~S~values are not distributed normally). ::: Two other AADC ortholog pairs showed strong purifying selection at work, with values like that calculated for the *bas-1*orthologs (Table [3](#T3){ref-type="table"}), but a value could not readily be calculated for the *tdc-1*orthologs. In all the AADC ortholog comparisons, the proportion of observed to potential synonymous substitutions (pS) was near mutational saturation (pS \> 0.75); K~A~/K~S~cannot be calculated when pS \> 0.75. Thus, a value could not be calculated either for full-length *tdc-1*alignments, or using a *tdc-1*core sequence. We were able to calculate a value by aligning the N- and C-terminals sequence of the *tdc-1*orthologs (Table [3](#T3){ref-type="table"}). These regions of the protein are under levels of selection like the other AADCs, whereas the core has a very low rate of non-synonymous substitution, consistent with the high level of amino acid conservation in this region of the protein. Discussion ========== Our experiments demonstrate that the predicted gene C05D2.4, which encodes an aromatic L-amino acid decarboxylase (AADC), corresponds to the genetically-defined *bas-1*gene. Serotonin immunoreactivity is restored in *bas-1*mutants by DNA containing an intact C05D2.4 gene, but not with DNA mutated in C05D2.4. The adjacent AADC-homologous gene, C05D2.3, is not needed to rescue *bas-1*mutants. The *bas-1*gene is therefore likely to encode the serotonin- and dopamine-synthetic AADC of *C. elegans*. Although we did not test for rescue of dopamine expression, it is likely that *bas-1*encodes the same AADC required for DA synthesis. Mutants with point mutations in C05D2.4 -- *bas-1*alleles *n2948*and *n3008*-- have been shown previously to be DA-deficient \[[@B17]\], and neither of these appears to contain mutations in the C05D2.3 gene. Furthermore, AADC proteins from other animals have been consistently shown to catalyze both 5HTP and L-dopa decarboxylation reactions \[[@B3]\]. Finally, a *bas-1*reporter construct is expressed both in identified serotonergic and dopaminergic cells. The *bas-1*gene is expressed in at least two alternatively spliced forms, one of which appears to be less common and contains a small additional 27 nucleotide exon. The short segment of protein encoded by the additional exon, and the surrounding region are not found in other AADC proteins, suggesting a novel function for this region of the AADC protein. In other organisms, the serotonin- and dopamine-synthetic AADC genes have alternative splicing that result in tissue-specific protein isoforms. Currently we have no indication that *bas-1*is expressed in any cells other than serotonergic and dopaminergic neurons, and no information about the functional significance of this alternative splicing. AADC has received somewhat less attention with respect to the regulation of serotonin and dopamine synthesis than the specific, rate-limiting synthetic enzymes tryptophan hydroxylase and tyrosine hydroxylase \[[@B51]\]. This is in part due to the view that AADC activity is not limiting, and that its activity is not regulated. Regulation of AADC activity by protein kinase A-dependent phosphorylation has recently been proposed based on *in vitro*experiments \[[@B52]\], although its functional significance has been questioned \[[@B53]\]. Our examination of the predicted BAS-1 protein revealed several potential phosphorylation sites that are highly conserved, although none fit the consensus sequence for PKA phosphorylation. Any possible regulation of *C. elegans*AADCs by phosphorylation remains speculation. Possible functions of other AADC homologous genes in C. elegans --------------------------------------------------------------- We compared the protein sequences of other predicted AADCs in *C. elegans*with those of other organisms in order to guess about their possible functions. This is particularly relevant because all *bas-1*mutants retain weak, residual serotonin immunoreactivity (\[[@B13]\]; C. Loer, unpublished) suggesting that other enzymes may be able to carry out the same reaction. This would not be surprising since animal AADCs tend to have broad specificity \[[@B3]\]. Based purely on sequence homology, it seems that predicted genes K01C8.3 and ZK829.2 could act as AADCs or as HisDCs. In fact, the predicted gene K01C8.3 is now believed to be a tyrosine decarboxylase and has been named *tdc-1*\[[@B48]\]. If correct, then its best match in *Drosophila*(G30446), an uncharacterized AADC homolog, is likely to encode the fly\'s octopamine-synthetic tyrosine decarboxylase. It has long been known that a separate gene encoded this enzymatic activity in flies, since the activity is still detectable in *Ddc*deletion mutants \[[@B4]\]. It will be interesting to see whether such tyrosine decarboxylases in animals have more restricted substrate specificity, such as the tyrosine and tryptophan decarboxylases in plants \[[@B54]\], or are more similar to typical animal AADCs with a broad specificity. Tighter substrate specificity of a *tdc-1*protein could be reflected in the much slower rate of amino acid substitution seen in its *C. elegans*&*briggsae*orthologs than in the *bas-1*orthologs which encode more \'promiscuous\' enzymes. Whether *C. elegans*or other nematodes make the neurotransmitter histamine, and therefore need a HisDC enzyme, is unclear. Although histamine has been reportedly isolated from *C. elegans*\[[@B55]\], this observation is unique among nematodes, and has not subsequently been confirmed. There is no particularly good candidate for a HisDC in *C. elegans*. The ZK829.2 predicted protein may be most closely related to *tdc-1*in its core sequence, although its long N- and C-terminal extensions are perhaps suggestive of a new function. Unfortunately, transgenics with reporter fusions of this gene to date have shown no expression, the pattern of which might suggest a function (C. Loer, unpublished; M. Alkema, personal communication). As with *tdc-1*, *C. elegans*ZK829.2 and its *C. briggsae*ortholog have also evolved more slowly than the *bas-1*orthologs. A recent analysis of eukaryotic AADC sequences that includes the *C. elegans*ZK829.2 and its *C. briggsae*ortholog as the only nematode representatives clearly demonstrates that AADC genes can evolve at very different rates, and that a constant \"molecular clock\" cannot be assumed in phylogenetic analyses \[[@B56]\]. Finally, since the C09G9.4 predicted protein is so highly divergent from the typical AADC, and lacks a critical lysine residue that binds the PLP cofactor, it is unlikely to be an AADC enzyme. It has a similar level of divergence from genuine AADCs as do other group II PLP-dependent enzymes such as cysteine sulfinic acid decarboxylase, to which it has little or no similarity. Whatever the function of a C09G9.4-encoded protein, it appears to represent a new PLP-DC-related protein; sequencing of more genomes may yet reveal additional members. Duplicate gene retention and loss in *Caenorhabditis* ----------------------------------------------------- We found that the closest relatives of C05D2.4/*bas-1*in *C. elegans*, the genes C05D2.3 and F12A10.3, are missing from *C. briggsae*. Furthermore, phylogenetic analysis indicates the two extra genes did not arise in the *C. elegans*line, but were present (or their commmon ancestor was present) in the species that gave rise to both the *C. elegans*and *C. briggsae*lines. Finally, careful examination of the cDNAs and predicted protein sequences of C05D2.3 and F12A10.3 reveals that neither is likely to be functional as an AADC: the former lacks critical amino acids and the latter can encode only a truncated protein. Both are expressed, based on the presence of cDNAs, but probably at a very low level, which is not above background in microarray experiments. It is possible that the duplicate genes are functionally \'lost\' in *C. elegans*as well. The features of C05D2.3 and F12A10.3 raise a number of interesting questions about the fate of duplicate genes, and the true nature of many \'predicted genes\' in *C. elegans*. Taking a random sampling of predicted genes and generating transgenics with reporter fusion constructs (in order to determine a pattern of expression), Mounsey and colleagues \[[@B57]\] found that a much higher percentage of recently duplicated genes than conserved or unique genes failed to show expression. Assuming that failure of expression was no more likely among recently duplicated genes for technical reasons, this meant that many more of these are in reality not expressed. The numbers suggested that up to 20% of annotated, predicted genes in *C. elegans*may be pseudogenes. In fact, careful inspection of recently duplicated genes showed that many were actually pseudogenes, like we found to be the case for F12A10.3. Overall, close inspection of predicted genes revealed at least 4% were pseudogenes. So, why are C05D2.3 and F12A10.3 still present in *C. elegans*if they lack a function? *C. briggsae*and *C. elegans*may have diverged 80 -- 110 million years ago \[[@B37],[@B38]\]. Since the *bas-1*-like gene or genes were likely present in the common ancestor of *C. elegans*and *C. briggsae*, then there seems to have been ample time for loss in the *C. elegans*line. Under a simple model of gene loss following duplication, only a few million generations would be the mean time to fix a null allele of the gene duplicate \[[@B58]\]. In *Caenorhabditis*, a million generations could be completed in 10,000 years or less. This seems to suggest that the downstream duplicate of *bas-1*(ancestor of C05D2.3) may have continued to function for a considerable time after the duplication, perhaps by gene conversion which might have continued until sufficient divergence from *bas-1*\[[@B59]\]. Loss of the critical six amino acids occurred after the second duplication giving rise to the ancestor of F12A10.3, since the appropriate sequence is still present there (although frame-shifted). It is also possible that the C05D2.3 gene retains some function. The gene still encodes a respectable protein, albeit one that seems unable to function as an AADC. It has diverged considerably from *bas-1*, but has not accumulated stop codons and frameshifts expected for a pseudogene. Walsh \[[@B60]\] has proposed that fixation of an allele with an advantageous new function, vs. becoming a pseudogene, may be the fate of many duplicate genes even when such mutations are rare, given a population that is sufficiently large. C05D2.3 and F12A10. 3 seem to have been retained longer than expected. Lynch and Force \[[@B61]\] proposed that the unexpectedly high rate of gene duplicate retention in eukaryotic genomes is due to \'subfunctionalization\' -- the retention of a portion of the original single gene\'s function by each of the duplicates, which then complement one another. Although this was suggested to occur primarily by regulatory mutations that partition expression of the genes spatially, other forms of subfunctionalization could also occur. Another possible reason for retaining such genes is the presence of non-coding regulatory functions associated with transcription and splicing of these sub-functional transcripts that affect the transcription of other nearby genes, although a *bas-1*::GFP construct is expressed well without such sequences *in cis*. Our analysis of synonymous vs. non-synonymous substitutions indicates that the *bas-1*-like genes C05D2.3 and F12A10.3 are under relaxed selection relative to *bas-1*and other AADCs. It should be noted that precise quantitative comparisons cannot be made with the results presented in the *C. briggsae*whole genome analysis \[[@B37]\], since we used a different method of calculating K~A~/K~S~; however our calculations indicate that *bas-1*and the other AADC\'s, like most genes in the *Caenorhabditis*genomes, are under strong purifying selection. Even if both C05D2.3 and F12A10.3 are now pseudogenes, some significant period of time during which they functioned and were under purifying selection could act to obscure this fact in an analysis of K~A~/K~S~. Even if C05D2.3 has acquired a new, adaptive function, such a new function might result from changes in only a few sites in the protein, and so again this could be obscured by a majority of sites under purifying selection. With the sequencing of three related *Caenorhabditis*species it will be interesting to learn of the fates of *bas-1*and the *bas-1*-like genes in other lines. Conclusions =========== The *bas-1*gene encodes a serotonin- and dopamine-synthetic AADC enzyme in *C. elegans*. The *C. elegans*genome possesses five other AADC-homologous genes, two of which are closely related to *bas-1*. These *bas-1*-like genes are missing, however, from the congeneric *C. briggsae*, and evidence suggests that, despite their persistence in *C. elegans*, the genes do not encode functional AADC proteins. Since one or more of the *bas-1*-like genes was likely present in the common ancestor of *C. elegans*and *C. briggsae-*which may have diverged over 80 million years ago -- it is unclear why the *bas-1*-like genes have been retained in the *C. elegans*line. This is another example of unexpected retention of duplicate genes in eukaryotic genomes. Methods ======= Routine culturing of *Caenorhabditis elegans*was performed as described by Brenner \[[@B62]\]. Nomenclature used here for *C. elegans*genetics conforms to the conventions set forth by Horvitz *et al*. \[[@B63]\]. Strains used include N2 (wild type); CB1490: *him-5(e1490)V*; MT7988: *bas-1(ad446)III*; MT7990: *bas-1(n2948)III*; MT8002: *bas-1(n3008)III*; LC7: *bas-1(pa4)III*; LC33: *bas-1(tm351)III*. The *him-5(e1490)*strain generates approximately 30% males by increased X chromosome non-disjunction \[[@B64]\], but is otherwise essentially wild-type. Mutant allele sequencing ------------------------ Mutations were identified by PCR-amplifying small regions from genomic DNA, then sequencing purified PCR product. We isolated genomic DNA (Puregene Kit, Gentra Systems) from four *bas-1*mutant strains (alleles *ad446, n2948, n3008, pa4*) and then PCR-amplified small segments of C05D2.4 and C05D2.3 protein-coding sequence. Five pairs of primers were used to survey the C05D2.4 gene (A+B, E+F, G+H, P+Q, R+S; see below for primer sequences) and 4 pairs for C05D2.3 (C+D, K+L, M+N, P+Q). Bands of the predicted size were excised from 2% agarose gels and purified with GeneClean (Bio101/Qbiogene), then sequenced. All mutations were confirmed by sequencing both strands for two independent PCR reactions. All PCR and sequencing primers were designed using the program Primer3 (<http://frodo.wi.mit.edu>; \[[@B65]\]). Primer sequences were as follows. Within C05D2.4: C05D2-A: gaggaaactcaaggcgacac; C05D2-B: tgttgatggaaccaagtgga; C05D2-E: cgtccttttctctttgcgac; C05D2-F: tggctccgacttgattctct ; C05D2-G: ttacaattaggccgcaaacc; C05D2-H: ccacctgaactgtggtgatg; C05D2-P: ggactcacatgtttccgattg; C05D2-R: ttagacgttggttgcacgag; C05D2-S: attggcgagcagtcaaagtt ; C05D2-T: tcttatgggattaccagaac; C05D2-U: ctacataaagctggaatggt; C05D2-V: gtttcctaaaaatccacgtg; C05D2-W: atgatcgattgatagctgag. Within C05D2.3: C05D2-C: ctaggtgcctttgccttctg; C05D2-D: caagagacgctcgttgtcag; C05D2-K: gccatctaatcctccaacca; C05D2-L: acattgctcccttttcaacg \[note that primer L can also prime within C05D2.4\]; C05D2-M: ccatcaactttccaatggct; C05D2-N: tctcgacgcccatatttctc; C05D2-Q: ccaattccagcggagaagta. Microinjection to create transgenics ------------------------------------ All DNAs for microinjection were purified with Qiagen tip20 columns. Experimental DNAs were co-injected with the dominant marker plasmid pRF4 containing the mutant gene *rol-6(su1006)*\[[@B66]\] into N2 (wildtype), CB1490, or *bas-1*mutant worms. Progeny of injectees that express the *rol-6*dominant plasmid have the easy-to-identify Roller phenotype which results in worms with a helically-twisted body along the anterior-posterior axis. Roller transgenic worms typically carry co-injected DNAs. Transgenic Roller progeny were isolated and propagated; rescue of *bas-1*mutants was scored by staining with serotonin antiserum as described previously\[[@B16],[@B67]\]. Mutant rescue with subclones of cosmid C05D2 and plasmid C05D2XN; GFP Reporter Construct ---------------------------------------------------------------------------------------- The rescuing plasmid C05D2XN contains a 15.8 kbp genomic DNA insert (XhoI to NheI) derived from the cosmid C05D2, and contains both C05D2.4 and C05D2.3 predicted genes. A number of deletions of C05D2XN were made to test for rescue of *bas-1*mutants. Clones were analyzed by restriction digests. We also made C05D2XN derivatives in which either C05D2.4 and C05D2.3 was mutated to introduce a premature stop in coding sequence. Clones were sequenced to determine the nature of the introduced mutation. Clone pCL6991 had a 4 bp deletion in the second exon of C05D2.4; pCL7991 had a 2 bp insertion in the first exon of C05D2.3; each causing a frameshift mutation. In the clone pCL7003, derived from the rescuing plasmid pCL3001, almost the entire C05D2.3 coding sequence is deleted. The *bas-1*GFP reporter construct (within transgenics) was kindly provided by Ian Hope, and consists of a PCR-generated fragment of C05D2 with 4595 bp upstream of the predicted translation start site and 403 bp protein coding region to make an in-frame protein fusion in the 2nd exon. The construct was created by a multi-site recombination reaction with the C05D2 fragment, PCR-generated GFP, and vector using the Invitrogen Gateway cloning system (I. Hope, personal communication) as continuation of a project to determine expression patterns for *C. elegans*genes through reporter gene technology \[[@B57]\]. RT-PCR and cDNA clones ---------------------- We isolated RNA for RT-PCR from mixed stage CB1490 worms. (The CB1490 strain has \~30% males, whereas wildtype N2 has \~0.2% males. Since adult males have 13 more serotonergic neurons and 6 more dopaminergic neurons than hermaphrodites, we reasoned that these worms might express more *bas-1*mRNA.) Worms were isolated from six 100 mm NGM plates, washed several times with M9 buffer, and pooled to form a \~100 μl pellet of worms. The pellet was mixed with 175 μl RNA lysis buffer (SV Total RNA Isolation System, Promega), frozen and ground to a fine powder with a pestle and mortar cooled with liquid nitrogen. The powder was recovered in a fresh microfuge tube, mixed with 350 μl SV RNA dilution buffer and centrifuged to remove debris. The cleared lysate was transferred to a new tube and precipitated with 200 μl 95% ethanol and applied to the SV spin column assembly. The remaining RNA purification was performed exactly as described for the SV System \"RNA Purification by Centrifugation.\" Purified RNA was eluted from the spin column with 100 μl nuclease-free water. RNA was converted to cDNA in a 80 μl reaction containing 800 Units of M-MLV Reverse Transcriptase (Life Technologies/BRL), 16 μl 5X RT buffer, 25 mM dNTPs, 80 Units RNAsin (Promega) 2.0 μg random hexamer primers (Life Tech), and 50 μl purified RNA (from above). *Bas-1*(C05D2.4) and C05D2.3 cDNAs were amplified by PCR from *him-5*cDNA produced as described above, typically using Failsafe PCR mixes (Epicentre Technologies). Conditions for PCR were 94°C (1 min.), 50°C (1 min.), 72°C (3 min.) for 40 cycles, then 72°C (10 min.). Primers used to amplify C05D2.4 cDNAs were SL1-B + C05D2-L and SL1-B + C05D2-B; a partial C05D2.3 cDNA was amplified with primers C05D2-M and -N. Spliced leader primer sequences were as follows: SL1-B: AAAGGATCCTTTAATTACCCAAGTTTGAG; SL2-B: AAAGGATCCTTTTAACCCAGTTACTCAAG. Appropriately-sized bands were isolated with GeneClean or Ultrafree-DA (Millipore), then cloned directly into the pCRII-TOPO vector using the TOPO-TA Cloning kit (Invitrogen). From seven of our RT-PCR derived clones we sequenced, we found two different splice variants different form the Genefinder prediction (see results). We also obtained cDNA clones from the ORFeome project \[[@B40]\] and the *C. elegans*EST project (courtesy Yuki Kohara). DNA we received from the ORFeome project is purified from a pool of transformants derived from ligation and transformation of their original RT-PCR, thereby allowing the isolation of internal splice variants from the mix. We transformed with this DNA and isolated several clones. We sequenced two ORFeome clones completely; six more clones were partially sequenced. Interestingly, each of the 8 clones appeared to have at least one mutation (compared to known genomic sequence). Since we found one splice variant containing the 27 bp microexon from among the eight clones that we sequenced, we analyzed 29 additional ORFeome project-derived clones by PCR from single isolated bacterial colonies. Amplification of the region between primers C05D2-T and -L allowed us to distinguish between clones with or without the 27 bp microexon based on product size (464 bp versus 437 bp). Seven of 29 clones were the larger size, so likely contained the 27 bp microexon. None of the ORFeome clones analyzed by sequence or PCR appeared to use the alternative exon 3 splice acceptor noted above. Finally, the two \'YK\' cDNA clones from the *C. elegans*EST project that we examined are from a \'full-length\' capped library, and each has an SL1 leader and a poly-A tail. We found that each *bas-1*\'YK\' clone, however, contained a different internal deletion (overall abnormality of these clones is reported at \~5% -- J. Theirry-Meig, personal communication). Each of the deletions was adjacent to a short repeated sequence. Sequence Analyses ----------------- *C. elegans*genomic and predicted cDNA sequences were retrieved from ACeDB, WormBase, and/or GenBank. Blast searches and Blast2 comparisons were performed using the NCBI Blast server. *C. briggsae*genomic sequence was searched using TBLASTN of the 7/12/02 shotgun assembly <http://www.sanger.ac.uk/Projects/C_briggsae/blast_server.shtml>. Assembly and consensus sequence determination of our own cDNA sequences was done using the program SeqMan (DNAStar, Madison, WI). Some information about ORFeome clones was retrieved from WorfDB (<http://worfdb.dfci.harvard.edu>; \[[@B68]\]). Phosphorylation predictions were made with NetPhos 2.0 <http://www.cbs.dtu.dk/services/NetPhos/> and PhosphoBase <http://www.cbs.dtu.dk/databases/PhosphoBase/predict/predict.html>. Signal sequence predictions were performed with SignalP v.1.1 (<http://www.cbs.dtu.dk/services/SignalP/>, \[[@B69]\]). Multiple sequence alignments were performed primarily with CLUSTALW <http://www.ebi.ac.uk/clustalw/>, with some manual adjustments. Phylogenetic analyses were performed with PAUP\* version 4.0b10 (Sinauer Associates, \[[@B70]\]). Some alignments of cDNAs with genomic DNA to determine intron locations were performed with SIM4 \[[@B71]\]. Estimates of rates of synonymous and non-synonymous substitutions were made with SNAP (Synonymous/Non-synonymous Analysis Program -- <http://www.hiv.lanl.gov/content/hiv-db/SNAP/WEBSNAP/SNAP.html> using pairwise or multiple sequence nucleotide alignments generated with CLUSTALW and adjusted manually. This program uses the method of Nei and Gojobori \[[@B72]\]. List of Abbreviations ===================== AADC -- aromatic amino acid decarboxylase DDC -- dopa decarboxylase HisDC -- histidine decarboxylase GAD -- glutamic acid decarboxylase TrpDC -- tryptophan decarboxylase PLP -- pyridoxal 5\'-phosphate NSM -- neurosecretory motoneuron HSN -- hermaphrodite-specific neuron PDE -- postdeirid sensory neuron ADE -- anterior deirid sensory neuron RN -- ray sensory neuron CEPD, CEPV -- dorsal or ventral cephalic sensory neuron ADF -- amphid sensory neuron, dual cilia, designation F AIM -- ring interneuron, designation M RIH -- ring interneuron (unpaired), designation H CP -- posterior daughter of designation C cell, male-specific ventral cord motoneuron Authors\' contributions ======================= EH prepared deletion and expression constructs, cloned and sequenced bas-1 cDNAs and mutant alleles, participated in anti-serotonin staining and genetics, and drafted the manuscript. CL conceived and directed the study, generated transgenics by microinjection, performed sequence and phylogenetic analyses, and completed the manuscript. Both authors read and approved the final manuscript. Acknowledgements ================ We wish to thank the following persons: Willy Christian and Patrick Merritt for technical assistance; Raffi Aroian for advice regarding RT-PCR; Fred Wolf and Gian Garriga for the C05D2XN plasmid and sharing unpublished information; Beth Sawin, Mark Alkema and H. Robert Horvitz for sharing unpublished information; and other members of the Loer lab for various contributions to this project, including Susan Anderton (isolated the *bas-1(pa4)*mutant), Scott DePaul (sequenced the *C. briggsae bas-1*cDNA), Tonya Lebet (isolated and sequenced F12A10.3 cDNAs), Nicola Marsden-Haug, Nailah Thompson, Cristina Ramirez, and Laura Rivard (for comments on the manuscript). *C. briggsae bas-1*cDNAs and genomic clones were originally isolated by members of the Spring 2002 Biology 182 (Molecular Biology) class at USD. Ian Hope generously provided two *bas-1*::GFP reporter transgenic strains made in his lab. Michael Mayer provided significant help and guidance with phylogenetic analyses using PAUP\*. Some of the strains used were obtained from the Caenorhabditis Genetics Center, which is funded by the NIH National Centers for Research Resources. The C05D2.4 and C05D2.3 deletion mutants were generated by the *C. elegans*Gene Knockout Consortium (*tm351*was specifically provided by the labs of Shohei Mitani and Yuji Kohara; *ok704*by the Barstead lab). Two *bas-1*cDNAs were provided by the Kohara lab (*C. elegans*EST/transcriptome project); *bas-1*and F12A10.3 cDNAs by J. Reboul and the M. Vidal lab (*C. elegans*ORFeome project). CL was supported by a National Science Foundation RUI grant (Developmental Neuroscience, IBN-9796217), an NIHGMS Area Grant (R15 GM60203), start-up funds from the University of San Diego, and the Fletcher Jones Foundation. Some of this work was performed when CL was at Lafayette College, and so was also supported by start-up funds from that institution, some of which came from a Howard Hughes Medical Institute grant.

PubMed Central

2024-06-05T03:55:47.203534

2004-8-2

PMC514704

Background ========== Biliary atresia (BA) is a cholestatic disease of infancy characterized by the destruction of the biliary tree. \[[@B1]-[@B3]\] Both the intra-and extra-hepatic biliary ducts demonstrate evidence of a progressive destruction. This results in cholestasis, hepatic fibrosis and eventually cirrhosis. BA is associated with significant morbidity and mortality. Although the incidence of BA is 1 in 10,000 to 14,000 live births, \[[@B4],[@B5]\] it accounts for over 40% of the neonatal cholestatic liver disease in Europe and the United States. \[[@B6]\] Prior to the hepatic portoenterostomy and liver transplantation, infants with BA had less than a 10% survival at 3 years of life and almost 100% mortality at 7 years of life. \[[@B7]\] In the United States, it is the leading cause of pediatric liver transplantation. \[[@B8]\] The etiology of biliary atresia remains unknown. Many hypotheses on the etiology of BA exist. Two leading hypothesis are that BA occurs as a result of ductal plate malformations occurring during development of the liver, or as a result of an immune mediated process triggered by yet to be determined stimulus. \[[@B9]\] As more knowledge accrues about the histology and immunologic characteristics of BA, the more it appears to be a progressive immune mediated process. Graft versus host disease (GVHD), seen after allogeneic hematopoietic stem cell transplantation, is an immune mediated process triggered by the transfer of alloimmune cells from a donor to a host. GVHD shares many similarities with BA. Both BA and GVHD are characterized by a lymphocytic infiltration around the portal triad and damage to the biliary ducts. \[[@B10]\] The predominant lymphocytes in both disease are CD4^+^T -- helper (Th) 1 cells. \[[@B11]-[@B14]\] As with BA, in GVHD there is also an increase in cell adhesion molecules and human leukocyte antigen (HLA) class II markers. \[[@B11]\] Given the similarities between BA and GVHD, we hypothesized that a contributing etiology of BA could be an alloimmune reaction such as in GVHD triggered by maternal microchimerism. Maternal microchimerism occurs when a small number of maternal cells are transferred to the offspring during pregnancy. This is known to occur in up to 40% of normal pregnancies. \[[@B15],[@B16]\] In addition, a clinical precedent of maternal microchimerism causing hepatic GVHD in children with severe combined immune deficiency (SCID) exists. \[[@B17]\] The aim of this study was to determine the presence and extent of maternal microchimerism in the livers of infants with BA. Methods ======= This study was performed with the approval of the Committee of Human Research at USCF (H9048-20247-01). Cases of BA were identified through a data search in the Anatomic Pathology CoPath system by diagnosis or during patient medical visits to the UCSF Division of Pediatric Gastroenterology, Hepatology, and Nutrition. An explanation of the study was given to each study candidate. Written consent was obtained from each study participant. Fluorescent in situ hybridization (FISH) of male BA livers ---------------------------------------------------------- We modified a previously reported method of FISH, \[[@B18]\] using X and Y-chromosomes probes. Y-chromosomes were stained with a green fluorescent dye, fluorescein isothiocyanate (FITC). X chromosomes were stained with a red fluorescent dye, cyanine 3 (Cy-3). Nuclear material was stained with a blue fluorescent dye 4\", 6\"-diamidino-2-phenylindole (DAPI). After initial digestion and staining of the liver specimens, we examined the slides for evidence of female cells, depicted by two red signals, i.e. X chromosomes, with no green signal within the blue nuclear material. Slides were analyzed in a blinded fashion. HLA typing ---------- Patients with BA and their mothers were HLA typed. The HLA typing HLA-A, -B, and -DRB1 alleles of the child and mother were determined by sequence specific PCR (SSP) (Pel-Freez Clinical Systems, LLC^®^, Brown Deer, WI, USA). Immunohistochemistry -------------------- Frozen tissues sections of BA patient\'s livers (5 μm) were fixed in acetone for 10 minutes at 4°C then washed in PBS (5 minutes centrifugation × 3). Sections were blocked with Protein Block (Dako, Carpinteria, CA, USA) for thirty minutes at room temperature. Sections were then incubated with anti-HLA-B14 mouse monoclonal antibody (US Biological, Swampscott, Mass. USA) then washed in PBS. Sections were incubated in goat anti-mouse conjugated with FITC for 30 minutes at room temperature. Sections were rewashed and then counterstained with DAPI. Slides were then examined using a fluorescent microscope. Kinetic Polymerase Chain Reaction (kPCR) ---------------------------------------- Using the HLA types of the patients with BA and their mothers, and a previously reported method of kPCR, evidence of maternal DNA was explored within the liver biopsy specimens of explanted BA livers. \[[@B18]\] The minor modifications to the kinetic PCR protocol included use of iCyler from BioRad (Hercules, Ca, USA) for amplification. Results ======= Detection of female cells in male BA liver biopsies --------------------------------------------------- Using FISH, we examined 11 male BA liver biopsies and 4 male NH liver biopsies. After initial digestion and staining of the liver specimens, we were able to interpret 8 of the BA and 2 NH liver biopsy specimens (Table [1](#T1){ref-type="table"}). Three BA and 2 NH biopsies underwent poor digestion and had no discernable signal. Figure [1](#F1){ref-type="fig"} shows a liver from a male infant with BA. A male cell, depicted by a red signal, the X chromosome, and green signal, a Y chromosome within the blue nuclear material, clearly can be seen (A). In the same specimen, one can also see a female cell, depicted by two red signals, both X-chromosomes with no green signal within the blue nuclear material (B). 7/8 BA specimens had evidence of maternal microchimerism. The number of maternal cells, per biopsy slide, was 2 -- 4 cells. Neither of the NH specimens had evidence of maternal microchimerism. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Overview of maternal microchimerism in patients with BA and NH ::: Diagnosis Chimerism Method \# of Maternal cells/DNA per slide (FISH) or per 25,000 copies of patient\'s DNA (kPCR) ----------- ------------------- --------------------------- ----------------------------------------------------------------------------------------- BA Present FISH 2 maternal cells per slide BA Present FISH 2 maternal cells per slide BA Present FISH 4 maternal cells per slide BA Present FISH 2 maternal cells per slide BA Present FISH+immunokistochemistry 3 maternal cells per slide BA Not interpretable FISH \-\-- BA Present FISH 2 maternal cells per slide BA Present FISH 3 maternal cells per slide BA Not interpretable FISH \-\-- BA Not interpretable FISH \-\-- BA None detected FISH 0 maternal cells per slide BA Present kPCR 3 copies of maternal DNA per 25,000 copies of patient\'s DNA BA Present kPCR 1 copy of maternal DNA per 25,000 copies of patient\'s DNA BA Present kPCR 10 copies of maternal DNA per 25,000 copies of patient\'s DNA BA Present kPCR 142 copies of maternal DNA per 25,000 copies of patient\'s DNA NH None detected FISH 0 maternal cells per slide NH None detected FISH 0 maternal cells per slide NH Not interpretable FISH \-\-- NH Not interpretable FISH \-\-- ::: ::: {#F1 .fig} Figure 1 ::: {.caption} ###### A male cell, depicted by a red signal, the X chromosome, and green signal, a Y chromosome within the blue nuclear material, can be seen in both male BA liver specimens (A&B). In the same specimen, one can also see a female cell, depicted by two red signals, both X-chromosomes with no green signal within the blue nuclear material (arrow). :::  ::: Detection of maternal HLA antigen among BA liver samples -------------------------------------------------------- To further investigate our findings, three patients with BA and their mothers were HLA typed. Using the HLA marker B14, we examined the BA livers for evidence of maternal microchimerism using immunohistochemistry techniques. An HLA-B14 positive patient served as our positive control (Figure [2](#F2){ref-type="fig"}). Specimen B, our negative control, was a patient who was, along with his mother, negative for HLA-B14. Specimen C, our test specimen, was HLA-B14^-^, with an HLA-B14^+^mother. Specimen A (positive control) has a bright signal for HLA-B14 while there is only background staining for HLA-B14 in specimen B (negative control). Specimen C, the test specimen, has isolated scattered bright signals suggesting maternal microchimerism. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Using the HLA marker B14, we examined the BA livers for evidence of maternal microchimerism using immunohistochemistry techniques. An HLA-B14 positive patient served as our positive control (A). Specimen B, our negative control, was a patient who was, along with his mother, negative for HLA-B14. Specimen C, our test specimen, was HLA-B14 negative, with an HLA-B14+ mother. Specimen A (positive control) has a bright signal for HLA-B14 while there is only background staining for HLA-B14 in specimen B (negative control). Specimen C, the test specimen, has isolated scattered bright signals suggesting maternal microchimerism. :::  ::: Detection of maternal DNA by kPCR --------------------------------- Finally, using polymerase chain reaction, in conjunction with the HLA typing of patients and mothers, we were able to affirm that maternal microchimerism occurs in the BA livers. An additional four fresh frozen liver specimens of BA infants, obtained at the time of transplantation, were examined for maternal DNA. All the livers had evidence of maternal microchimerism, with the range of 1 to 142 copies of maternal DNA detected per 25000 copies of patients\' DNA. Figure [3](#F3){ref-type="fig"} is a representative kPCR for maternal HLA-B40 in the second liver sample. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Representative kinetic PCR (kPCR) data for maternal HLA-B40 in BA liver sample. SYBR Green fluorescence is plotted on the y-axis as a function of amplification cycle number on the x-axis. The upper panels depicts results for standard dilutions of HLA-B40 \[left to right curves: 25000, 10000, 1000, 100,10, and 1 genomic equivalence (gEq)\]. A single copy of template produces fluorescent signal after approximately 36 cycles of amplification. The lower panel shows results from three liver replicates and two negative (no-template) controls. The three liver replicates have a positive signals at 2--5 gEq, while the negative controls remained negative through 50 cycles. :::  ::: In all eleven of twelve patients with BA had evidence of maternal microchimerism by FISH or kinetic PCR, while neither of the two neonatal hepatitis patients had evidence of maternal microchimerism (p = 0.03). Discussion ========== BA is a progressive cholestatic disease of infancy, which is characterized by a mild to moderate lymphocytic infiltration within the extra and intrahepatic bile ducts. It is plausible, that maternal microchimerism may play a role in the etio-pathogenesis of BA by causing an alloimmune reaction given the similarities between BA and GVHD. We have shown that maternal microchimerism is present within the livers of patients with BA. Maternal microchimerism occurs when maternal cells reside in the body of offspring. Maternal-fetal lymphocytic transfer is known to occur during pregnancy starting as early as the tenth week of gestation and continuing up to delivery. \[[@B19]\] The number of cells that traverse the placenta increases throughout this period. Schroder et al \[[@B15]\] demonstrated that approximately 1/10 infants has about 0.07% maternal lymphocytes at the time of birth and more recent findings by Lo et al. \[[@B16]\] have indicated the presence of maternal cells in over 40% of fetal blood samples. Disease has been shown to occur in mothers who have had fetal microchimerism, and children/adults who have had maternal microchimerism. Children who have immunodeficiencies are at greater risk for graft verses host disease owing to engraftment of maternal lymphocytes. In both SCID and DiGeorge syndrome, maternal microchimerism and GVHD have been well described. In a study by Susanne Müller et al., 121 patients with SCID were evaluated for maternal microchimerism using HLA typing. Maternal cells were found in 48 patients, with 19 patients showing signs of GVHD. GVHD manifested itself in the skin and in the liver. \[[@B20]\] There have been numerous case reports showing GVHD in infancy of patients with DiGeorge. Manifestations of this disease include skin, gastrointestinal tract and liver involvement. \[[@B21]\] The histologic and immunologic similarities of BA and GVHD are striking. The site of damage in BA and GVHD after bone marrow transplantation is the same. In each, lymphocytes congregate around the bile ducts. The damage occurs to both the intra-and extrahepatic biliary tracts. In a mouse model of acute GVHD, Nonomura et al. showed that transfer of allogeneic cells set along minor HLA mismatch can cause damage to both the intra and extrahepatic biliary ducts. Interestingly, the timeframe in this mouse model of acute GVHD, in terms of lymphocytic infiltration and fibrosis, appear similar to that of BA. \[[@B14]\] Initially, a peak of lymphocytes around the bile ducts occurs about 2 weeks after transfer of allogeneic cells from donor to host mouse; as the lymphocytic infiltration subsides, liver fibrosis increases. This correlates with disease progression in BA, where initial diagnostic biopsies of BA livers usually show a larger lymphocytic infiltration, and less fibrosis, as compared to biopsies done later, i.e. at the time of liver transplantation. \[[@B22]\] The lymphocytic infiltrates in both BA and GVHD are predominantly CD4^+^T lymphocytes. The genes in BA livers showed differential lymphocytic function, with activation of osteopontin, a regulator of cell-mediated immunity in Th 1 cells, and interferon gamma. \[[@B23]\] Similarly, in GVHD, the predominant leukocyte is the CD4^+^T lymphocyte. Although a variety of complex mechanisms are involved in causing the end organ damage in GVHD, evidence exists showing similarities in effector mechanisms. In a SCID mouse model of acute GVHD, an increase in interferon-gamma secretion with a synchronized increase in activated Th cells was seen during acute GVHD. \[[@B24]\] The inflammatory responses in both BA and GVHD are also associated with increased expression of adhesion molecules such as CD54, and increased class II HLA markers, such as HLA-DR. HLA-DR and CD54, in both BA and GVHD, are seen predominantly around the bile ducts. \[[@B11],[@B25],[@B26]\] Maternal microchimerism is not a rare occurrence in healthy individuals, but is usually not associated with disease. The high occurrence of maternal microchimerism in healthy individuals has been suggested to have a tolergenic effect that may contribute to long-term microchimerism. \[[@B27]\]Additionally, in utero transplantation of haploidentical cells does not always lead to immune tolerance \[28\] and may lead to immune sensitization. Thus, the immunological consequences of the migration of maternal cells to the fetus appear to be variable. In most cases the maternal cells are likely to be cleared by the host\'s immune system or they may escape destruction by the immune system leading to engraftment. Engraftment of maternal immune cells in the fetal biliary tract may result in an immune reaction against the host cholangiocytes. This is what we believe may occur in infants with BA. Alternatively, engrafted maternal cells in the fetus may subsequently be rejected by the immune system of the offspring, leading to destruction of the biliary tree. Maternal microchimerism occurs in the livers of patients with BA. Given the previously well-described relationship between microchimerism and GVHD, and the similarities between BA and GVHD, these findings indicate a potential etio-pathogenisis for BA. Further investigations into the role of maternal microchimerism in BA are warranted. Competing interests =================== None declared. Author\'s contributions ======================= DLS conceived of the hypothesis and contributed to the design and coordination of the study; he also drafted the manuscript. PR and MBH participated in the design of the study and manuscript preparation. DK carried out the kPCR and HLA typing analyses. GM performed the FISH analyses. LBL carried out the kPCR and HLA typing analyses and participated in the study design. MOM participated on the study design and coordination, immunohistochemistry stains and the manuscript preparation. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-230X/4/14/prepub> Acknowledgements ================ The authors wish to thank Drs. Yuet Wai Kan and Robert Suskind for their support and helpful suggestions. The invaluable support of Rebecca PH Davis and Leslie Lewinter is also much appreciated. Grant Support: This work was supported in part by NIH grants DK060617, 5T32 DK07762 and DK59301.

PubMed Central

2024-06-05T03:55:47.209859

2004-7-31

PMC514705

Background ========== Neo-adjuvant chemotherapy (NACT) is an integral part of multi-modality approach in the management of locally advanced breast cancer (LABC). It is required both for the local control (to ensure microscopically free margins during surgery) and distant or systemic control \[[@B1]-[@B5]\]. In the past few years, considerable research has been done on the molecular aspects of breast cancer. The recognition that tumor growth rate is a product of the proliferative activity and the rate of cell death has lead to a reappraisal of traditional views of tumor response and resistance to cytotoxic Drugs \[[@B2]\]. Apoptosis is a closely regulated form of active cell death defined by characteristic biochemical and morphological criteria. A large number of anti-cancer agents with widely differing modes of action have been demonstrated to induce apoptosis in vitro, suggesting this as a significant final common pathway for exerting their clinical effects. Mechanisms that suppress apoptosis may be important in the development of intrinsic and acquired resistance to cytotoxic drugs \[[@B3]\]. It was suggested more than 20 years ago that apoptosis might account for much of the spontaneous cell loss (known from kinetic studies) to occur in many tumors and its extent often is enhanced by well-established modalities such as chemotherapy, irradiation and hormone ablation. However, during the past few years, advances in the understanding of the control of apoptosis at the molecular level has extended its potential oncologic significance far beyond the mere provision of a mechanistic explanation for tumor cell deletion. In particular, the discovery that the products of certain proto-oncogenes can regulate apoptosis has opened up exciting avenues for future research \[[@B4]\]. The protean effects of various neoplastic agents on synthesis of DNA, RNA, and inhibition of synthesis which may or may not lead to cell death, requires only that some critical concentrations of active drug or metabolite be present in a cell. Proliferating normal cells may therefore be subject to the same detrimental effects of chemotherapy experienced by neoplastic tissue and successful chemotherapy dictates that recuperative abilities of normal tissues are greater than those of cancer. The two tissues generally most adversely affected by antineoplastics are the hemopoetic cells of the bone marrow and the epithelium of the aero digestive tract as a result of high growth fractions and short cycling times of the cells. The ability of the cancer patients to perform normal activities and function is also recognized both as a determinant of how well the patient may respond to chemotherapy and an index of the general toxicities of the anti cancer agents. A variety of anti-cancer drugs have been shown to induce extensive apoptosis in rapidly proliferating \"normal\" cell population and the tumors alike. Thus enhanced apoptosis is also likely to be responsible for many of the adverse effects observed following chemotherapy \[[@B5]\]. Apoptosis being the common final pathway both for tumoricidal effect and also for the toxic side effects, a significant correlation should therefore exist between the effects and the toxicity produced. Toxic effects could thus serve as reliable indicators of apoptosis Apoptosis is a regulated phenomenon capable of being inhibited and activated. Indeed there is evidence that stimulation of some cells by trophic cytokines or increase in their levels of expression of Bcl-2 proto-ontogeny can greatly increase their resistance to the apoptosis-inducing effects of anticancer drugs. Thus Bcl-2 proto-ontogeny expression may be implicated in the development of resistance of tumors to therapeutic agents and may contribute to tumor growth and perhaps to ontogenesis by allowing the inappropriate survival of cells with DNA abnormalities \[[@B6]\]. Deregulated expression of the Bcl-2 protein represents the best-known example of a potent blocker of apoptosis. Over expression of Bcl-2 has now been shown to protect a wide variety of cell types from induction of apoptosis by many different anticancer agents. Several homologues of Bcl-2 protein have also been shown to act as inhibitors of apoptosis, including Bcl-Xl and others as apoptotic proteins such as Bax. In vitro data suggest that it is the relative ratios of anti-apoptotic and pro-apoptotic proteins that determine the likelihood of cells to undergo apoptosis in response to chemotherapeutic drugs \[[@B2],[@B7]\] The increasing use of pre-operative chemotherapy (PCT) in breast cancer offers an *in vivo*test bed to further confirm the clinical relevance of these observations. The clinical response or the absence of response along with the toxic effects observed could well help predict the outcome with a particular chemotherapeutic regime and facilitate planning of an alternate regime for better response \[[@B5]-[@B8]\]. It is vital to assess the response to NACT in order to tailor the regime for a particular patient to predict the intrinsic or acquired chemo resistance. DNA-damaging agents such as chemotherapeutic drugs can induce apoptosis and increased resistance to chemotherapeutic agents, which has been found to be associated with decreased capacity to undergo apoptosis \[[@B5]-[@B9]\]. Central to this are proteins that modulate apoptosis, including bcl-2 and bax products. The change in expression of apoptotic markers brought about by various chemotherapeutic regimens is being used to identify drug resistance in the tumor cells. Various other biological markers like p-glycoprotein expression have also been used to predict the response to neoadjuvant chemotherapy \[[@B9]-[@B11]\]. The clinical response along with complete pathological response (CPR) is still considered a surrogate marker of response against which all other predictive markers are compared. The need to have a reliable and inexpensive predictor of response in a third world scenario can not be over emphasized since majority of patients present relatively late and the resources are limited and scarce \[[@B2],[@B3]\]. Since both the response and drug related toxicity due to chemotherapeutic agents are the result of the same common final pathway of apoptosis the toxicity due to chemotherapy along with clinical response may be utilized as a cost-effective and reliable indicator (predictor) of response to neo-adjuvant chemotherapy. Against this background a prospective study was contemplated with the following aims and objectives: 1\. To assess the clinical and immunohistochemical response to NACT in patients with LABC. 2\. To correlate immuno-histochemical (apoptotic markers i.e. Bcl-2/Bax ratio) and clinical response to the drug induced toxicity. 3\. To ascertain whether the drug induced toxicity could be utilized as a reliable indicator and predictor of response to neoadjuvant chemotherapy. Methods ======= 50 FNAC proven cases of locally advanced breast carcinoma according to AJCC (American Joint Committee On Cancer) classification were included in the study A thorough clinical and ultrasonographic examination (USG) of all the patients including the opposite breast was performed to stage the disease accurately. A core biopsy using a tru-cut needle was performed for immuno-histochemical estimation of the apoptotic markers i.e. base-line Bcl-2/Bax ratio before initiating the chemotherapy. Routine and metastatic work up was done including complete blood examination (total blood count, platelet count), chest radiograph, ECG (Echocardiography when ECG had a positive finding), liver function tests, Bone Scan, USG abdomen, KFT (Kidney function tests). Patients were subjected to three cycles of FAC regime (cyclophosphamide 600 mg/m^2^, adriamycin -50 mg/m2, 5-fluorouracil-600 mg/m2) at an interval of three weeks. Before each cycle the patient was clinically and sonologically examined for the breast tumor size, axillary lymph node status & appearance of systemic metastasis. All patients were given the same antitoxicity treatment according to a standardized unit protocol including adequate hydration and inject able antiemetics before initiating the chemotherapy. The patients were observed for three main toxicities of the FAC regime i.e. vomiting, alopecia and leucopenia. Vomiting with minimum of four episodes on day one and two after chemotherapy was taken as significant. Total alopecia was considered significant. Leucopenia was assessed in terms of the WHO grading of hematological toxicity. All patients were subjected to Patey\'s modified radical mastectomy three weeks after the last cycle and the specimen were again subjected to immuno-histochemistry to evaluate for any change in the Bcl-2/Bax ratio and for the histological tumor size, margins. Objective clinical response was defined as \>50% reduction in the tumor size after completion of three cycles of NACT, as assessed clinically, sonologically and histologically. Immunohistochemical response was taken as decrease in the Bcl-2/Bax ratio. Any increase or no change in this ratio was considered as no response. Immuno-histochemical methods ---------------------------- Biopsy specimen was preserved in buffered formalin solution. Five-micron sections were prepared from paraffin embedded blocks on poly-l-lysine coated glass slides. Sections were deparaffinized in xylene for 15 min. and hydrated in alcohol for 15 minutes. Further, incubation was done in 0.3% hydrogen peroxide in methanol solution for 45 min. The slides were washed with citrate buffer and kept in a water bath at 90--95°C for 45 min. for antigen retrieval. Sections were washed with Tris buffer saline (TBS) solution and incubated with blocking antibodies (DAKO monoclonal mouse antihuman Bcl-2 oncoprotein for Bcl-2 expression and polyclonal rabbit antihuman for Bax expression) at 37°C. Sections were washed with TBS solution. Incubation with avidin-biotin complex (ABC) was done at 37°C for one hour and washed with TBS solution. 3,3 Diaminobenzidine tetra hydrochloride solution applied for 3--5 min. Counter-staining with haemotoxylin solution done for 3--5 min. Sections were washed with distilled water, air dried and mounted using DPX mountant. For Bax, positive controls were taken as germinal centers of the lymphoid follicles and normal breast tissue and negative control was taken as the test slide without primary antibody. For Bcl-2, positive controls were the mantle zone of lymphoid follicles and the negative controls were the test slides without primary antibody. The pattern of positive staining for bcl-2 and bax was cytoplasmic, granular. The primary antibodies for bcl-2 and bax were procured from DAKO. Bax-Rabbit Anti-Human code no. A 3533. Bcl-2 Monoclonal Mouse Anti-Human code no. M 0887. Dilution for both was 1: 40. The results were interpreted on the basis of two criteria: \(1) Percentage of cells showing immune bodies; \<5%: score = 0, 5--25%: score = 1, 25--75%: score = 2, \>75%: score = 3 \(2) Intensity of staining; mild: score = 1, moderate: score = 2, intense: score = 3. \"*Since there was a strong correlation between the intensity of staining and percentage of cells showing immune bodies, the percentage of cells showing immune antibodies alone was considered for calculating the bcl-2/bax ratio*\". Statistics ---------- Descriptive studies were performed with SPSS version 10. The significance of response assessed using paired t-test. Significance of correlation between various variables assessed using chi-square test and coefficient of correlation was calculated by Pearson correlation coefficient. Results ======= 50 cases of locally advanced breast cancer were included in this study. Mean age of the patients was 45.5 years (range: 28 to 71 years) and 53.3% patients were pre-menopausal. Size of the tumor was measured clinically as well as by ultrasound and the patients were subdivided into four groups: \<5 cm(0%), 5--8 cm(56.6%), 8--10 cm(26.6%), \>10 cm(16.6%). According to the axillary lymph node status the patients were divided into three groups: N0 (0%), N1 (33%), N2 (67%). Objective clinical response was defined as more than 50% reduction in the tumor size after three cycles of neoadjuvant chemotherapy. Immuno-histochemical response was defined as decrease in the Bcl-2/Bax ratio. Clinical response including the reduction in the tumor size and axillary lymph node status was observed in 70% of patients and was found to be statistically significant (p \< 0.0001). There were no patients in the No group and 29.4%of the N1 patients were down staged to N0 while70.6% remained N1. In patients with N2 disease 7.7% were down staged to N0 status while 46.2% were downstaged to N1 status and 46.2% did not show any response. Immuno-histochemical response was observed in 60% and was also found to be statistically significant (p = 0.008). Correlation between immuno-histochemical and clinical response was also found to be statistically significant (p \< 0.0001) \[Table [1](#T1){ref-type="table"}\]. Acute vomiting was observed in 63.3% patients. 81% clinical responders had vomiting (p = 0.002) and 78% immunohistochemical responders also had vomiting which was statistically significant (p = 0.04). Alopecia was observed in 86% clinical responders (p = 0.000) and 94% immuno-histochemical responders (p = 0.000), which was also significant. Leucopoenia was observed in only 14% and 17% of clinical and immuno-histochemical responders respectively and was found to be an insignificant predictor of response in the present study. When multiple toxicities were correlated with the clinical and immuno-histochemical response, 46.7% of patients had both acute vomiting and alopecia. 67% clinical responders (p = 0.001) had both vomiting and alopecia.72% immunohistochemical responders (p = 0.001) had both vomiting and alopecia. A significant positive correlation was observed between the presence of vomiting (r = +0.558), alopecia(r = +0.802) and response to neoadjuvant chemotherapy. A significant negative correlation was observed between the absence of side effects and poor response to neoadjuvant chemotherapy (Table [2](#T2){ref-type="table"}). Discussion ========== Carcinoma of the breast is the leading cause of cancer in women all over the world and the second most common malignancy in India after carcinoma of the uterine cervix \[[@B1]\]. No other common epithelial cancer has been so carefully studied and so extensively characterized biologically \[[@B1],[@B2]\]. In developing countries like India rate of locally advanced breast cancer at first diagnosis is estimated to be as high as 25%--30%\[[@B2],[@B5]\]. The treatment of locally advanced breast carcinoma (LABC) has also evolved from primarily local modalities to treatment regimens that combine both systemic and local therapy. The realization that patients with LABC are likely to have undetectable micro metastases at diagnosis has lead to systemic treatment assuming major focus of the multi-modality approach as the studies have confirmed that surgery alone is an inadequate treatment in the management of these patients. Even aggressive surgical techniques have been observed to have a higher incidence of local recurrence in these patients \[[@B10],[@B11]\]. Most importantly surgery does not change the pattern of distant failure in patients who probably have micrometastatic disease at the time of diagnosis \[[@B10]-[@B13]\]. Multi-modality therapy that included surgery, radiation therapy, chemotherapy, hormonal therapy has had the greatest impact on survival in patients with LABC \[[@B10]-[@B13]\]. Neoadjuvant chemotherapy (NACT) ------------------------------- A new approach in the form of neoadjuvant chemotherapy was first reported in the 1970s and was initially utilized to convert unresectable tumors to smaller tumors making them more amenable to local control with either surgery or radiotherapy. An added advantage of this approach was the ability to assess patient\'s response to treatment both clinically after a defined number of courses of chemotherapy and pathologically after surgical resection. Perez and colleagues reported their results of a pilot study by the South-Eastern Cancer Study Group in 1979 that the primary tumor showed partial regression (\>50%)in 65% of patients after two courses of FAC \[[@B16]\]. NACT has also shown benefits in the operable breast cancers by increasing the chances of breast conservation by up to 90% in some trials \[[@B10]-[@B13]\]. The other important advantage of NACT is that it represents an *in vivo*chemo sensitivity test for assessment of tumor response from which prognostic information can be obtained. It provides an early treatment of the micrometastatic disease, counteracting the increased growth rate possibly determined by the shrinkage of the tumor. The down staging converts an inoperable case amenable to curative resection \[[@B10]-[@B13]\]. Apoptosis --------- Introduced by Kerr et al (1972) to describe characteristic morphological changes seen during programmed cell death \[[@B3]\]. It is defined as a closely regulated form of active cell death defined by characteristic biochemical and morphological criteria \[[@B3],[@B14],[@B15]\]. A wide range of anticancer drugs with widely differing modes of action have been demonstrated to induce apoptosis in vitro, suggesting this as a significant common final pathway through which they exert their clinical effect. Further more the mechanisms that suppress apoptosis may be important in the development of acquired resistance to cytotoxic drugs. Apoptosis or programmed cell death plays an important role in the regulation of tissue development, differentiation and homeostasis. It is therefore possible that deregulation of apoptosis contributes to the pathogenesis of cancer \[[@B3],[@B13]-[@B15]\]. Apoptosis can be differentiated biochemically and morphologically from necrosis by the following criteria \[[@B16]\]: \(1) Chromophin condensation \(2) Membrane blebbing \(3) Appearance of apoptotic bodies \(4) Fragmentation of genomic DNA Certain biochemical and genetic events have been identified that are associated with multiple cell types including mammary epithelium. These include the DNA fragmentation via end nuclease activation and cleavage of intracellular proteins, expression of bcl-2 family members, tumor suppressor gene p-53 directed events, proto-oncogene activation and activation of transmembrane receptor signaling pathways such as tumour necrosis factor \[[@B4],[@B17]-[@B22]\]. Although little is known about the mechanisms, which regulate apoptosis in epithelial cells, it is conceivable that defects in apoptosis related genes are involved in the pathogenesis of human cancers. The hypothesis is supported by the fact that the tumor suppressor gene product p-53, which is frequently mutated or deleted in breast cancer, is involved in regulating apoptosis \[[@B23]\]. The heterogeneous nature of breast cancer has resulted in overwhelming interest in search for prognostic markers to identify patients who might benefit most from the therapeutic modalities available. Assessment of apoptosis and individual components of apoptotic pathway are therefore relevant in determining prognosis in a particular patient \[[@B24]\]. DNA damaging agents such as ionizing radiations and chemotherapeutic drugs also induce apoptosis. Sakakura et al have shown an association between increased resistance to chemotherapeutic agents and decreased capacity to undergo apoptosis \[[@B25]\]. Central to this response are proteins that modulate apoptosis, including bcl-2 and bax gene products. Bcl-2 is anti-apoptotic in function, whereas bax is proapoptotic and it is the interaction between the two that determines the likelihood of a tumor to undergo cytotoxic drug mediated regression. Therefore any increase in bcl-2 or decrease in bax will push the balance towards chemo resistance and an increase in bax or decrease in bcl-2 will result in increased apoptosis \[[@B26]-[@B30]\]. It was observed in a study conducted by Kymionis et al \[[@B15]\] that increase in the ratio of anti apoptotic protein bcl-2 to pro-apoptotic protein i.e. bax results in markedly enhanced resistance of tumor cell lines to the cytotoxic effects of essentially all currently available chemotherapeutic drugs. In the present study the clinical response in terms of reduction of tumor size and immuno-histochemical response in terms of change of bcl-2/bax ratio correlated significantly with the drug induced toxicity following NACT. Toxicity related to chemotherapeutic agents \[[@B19]-[@B23]\] ------------------------------------------------------------- The time course of various toxic manipulations depends on the drug, its dose and frequency of administration, intrinsic characteristics of the tissue of interest and any local circumstances (e.g. radiation therapy, infection, trauma). There are few general rules however like mucosal toxicities of pain, erythema, ulceration etc. occur 3--10 days after the administration of most offending drugs. Bone marrow effects can be manifestated a few days later averaging 7--14 days. The recovery of normal functioning tissues in both cases is well under way 4--5 days after the zenith of the toxic effects. Alopecia may involve the scalp or the whole body can present within 2 weeks of the drug dose or be a progressive cumulative event. Other cumulative toxicities seen only after administration of a certain quantity of drug over some length of time include cardiomyopathy from anthracyclins. Cyclophosphamide ---------------- An alkylating agent belonging to the nitrogen mustard subgroup. It is inactive as such produces few acute effects and is not locally damaging. It is transformed in to active metabolites like aldophosphamide and phospharimide mustard in the liver, which then produces the wide variety of antitumour action. The prominent side effects are alopecia and cystitis, which are caused by another metabolite acrolein. 5-Fluorouracil -------------- It is a pyrimidine antagonist and is converted in the body to the corresponding nucleotide, 5-fluro-2-deoxy-uridine monophosphate, which inhibits thymidine synthetase and blocks the conversion of deoxyuridilic acid to deoxythymidilic acid. Selective failure of DNA synthesis occurs due to non-availability of thymidylate. Flourouracil may itself get incorporated in to nucleic acids and this may contribute to its toxicity. Even resting cells are affected, though multiplying cells are more susceptible like the cells in the gastrointestinal tract (GIT) and bone marrow. Doxorubicin ----------- It is an anti tumor antibiotic and is capable of causing breaks in the DNA strands by activating topoisomerase II and generating quinone type free radicals. They have mutagenic potential. Maximum action is exerted at S phase, but toxicity is usually exerted in G2 phase. Cardiotoxicity is a unique side effect. Rapidly multiplying cells are more susceptible therefore it also acts on cells of GIT, bone marrow in addition to the tumor cells. Leucopenia has not been observed to be a frequently encountered chemotherapy induced toxicity using commonly used regimen in most of the studies \[[@B19]-[@B24]\]. This was observed in the present study also. Conclusions =========== The rapidly proliferating normal and tumor cells are more susceptible to the action of chemotherapeutic agents, which could explain the significant correlation observed between the effects and the toxicity in the present study. There was a strong correlation observed between the immunohistochemical response (bcl-2/bax ratio), clinical response and drug toxicity. This indirectly indicates a correlation between chemotherapy induced apoptosis and the toxicity and therefore like apoptotic markers, chemotherapy induced toxic effects along with objective clinical response could serve as reliable and cost-effective indicators or predictors of response to NACT in patients with LABC. While many biological markers are in use and many are under trial to tailor the chemotherapy for a particular patient, most of these markers including apoptotic markers or p-glycoprotein etc. are not very frequently available and are expensive for a third world cancer set up. Thus the chemotherapy-induced toxicity along with clinical response may be utilized as a cost effective and reliable predictor of response to NACT in patients with LABC. This would also serve as an intermediate end point in determining drug sensitivity for adjuvant treatment, especially when adjuvant therapy is planned with the same regimen as induction chemotherapy. This can also help in planning an alternative regime in non-responders. Competing interests =================== None declared. Authors\' contribution ====================== • **CM**, the principal and the corresponding author was the Supervisor and the Chief surgeon who performed and standardized surgery on the patients and designed the study. • **VS**participated in the designing of the study, performed the statistical analysis and was the first surgical assistant and Senior Postgraduate in charge of the cases in the study. • **JP**, Postgraduate surgical resident was the second assistant in charge of the cases and participated in the sequence alignment. • **AL,**Resident surgery participated in the data processing and statistical analysis. • **SS**was in charge of the molecular genetic studies at the Tumor biology lab ICMR. • **AB**participated in the genetic studies and data processing. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2407/4/48/prepub> Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Immuno-histochemistry showing positive (moderate) staining of the tumor cells for Bax antibody. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Immuno-histochemistry showing positive (intense) staining of the tumor cells for Bax antibody. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Immunohistochemistry showing positive (moderate) cytoplasmic staining of the tumor cells for the Bcl-2 antibody. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Immuno-histochemistry showing positive (intense) cytoplasmic staining of the tumor cells for Bcl-2 antibody. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### n = 50 ::: Clinical response Clinical no response Immuno-histochemical response Immuno-histochemical no response ---------------------- ------------ --------------------- ---------------------- ------------------------------- ---------------------------------- Mean age 45.5 years Range 28--71 yrs Menopausal status Pre-menopausal 53.3% 62.5% 37.5% 56.2% 43.8% Post-menopausal 46.7% 71.4% 28.6% 64.2% 35.8% Tumor size \<5 cm 0% \- \- \- \- 5--8 cm 56.6% 68.7% 31.3% 60% 40% 8--10 cm 26.6% 75% 25% 60% 40% \>10 cm 16.6% 66.6% 33.3% 60% 40% Axillary lymph nodes N0 0% \- \- \- \- N1 33% 29.4%(toN0) 70.6% 60% 40% N2 67% 7.7%(N0), 46.2%(N1) 46.2% 60% 40% Total 70% (p \< 0.0001) 30% 60%(p = 0.008) 40% ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### n = 50 ::: Total N = 50 Clinical responders N = 35 Clinical non-responders N = 15 Immuno-histochemical responders N = 30 Immuno-histochemical non-responders N = 20 ------------------------- -------------- ---------------------------- -------------------------------- ---------------------------------------- -------------------------------------------- Acute vomiting 63.3% 81%(p = 0.002) 22.2% 78%(p = 0.04) 41.7% Alopecia 60% 86%(p = 0.000) 0% 94%(p = 0.000) 8.3% Leucopenia 10% 14%(NS) 0% 17%(NS) 0% Alopecia+Acute vomiting 46.7% 67%(p = 0.001) 0% 72%(p = 0.001) 8.3% :::

PubMed Central

2024-06-05T03:55:47.211273

2004-8-13

PMC514706

Background ========== The successful treatment and prevention of estrogen-dependent breast cancer in women by tamoxifen is attributed to its estrogen receptor (ER) occupancy \[reviewed in \[[@B1],[@B2]\]\]. In the N-nitrosomethylurea (NMU) induced breast cancer model in rats, tumor growth is estrogen dependent and tamoxifen is considerably more effective than raloxifene \[[@B3]\]. In the dimethylbenzanthracene (DMBA)-induced model in rats, in which tumor growth is predominantly dependent on prolactin for growth, tamoxifen and raloxifene show effective anti-tumor action. Tamoxifen and raloxifene have several properties in common; e.g. prevention of tumors in the DMBA induced rat mammary model, maintenance of bone density in the ovariectomized rat and reduction of low density lipoprotein cholesterol. The partial estrogen agonist activity of tamoxifen on uterine tissue, however, increases the risk of developing endometrial cancer. This does not appear to occur with raloxifene. Tamoxifen stimulates arachidonic acid release from rat liver cells \[[@B4]\]. In this report, I have compared tamoxifen and the raloxifene analog LY117018 for effectiveness at releasing arachidonic acid (AA) from rat liver, rat glial, human colon carcinoma and human breast carcinoma cells and their effects on prostaglandin (PG) I~2~production by the rat liver cells. Although both compounds release AA from these cells, LY117018 is less effective. Only tamoxifen stimulates both basal and PGI~2~production induced by incubation of rat liver cells with lactacystin in the presence of 12-O-tetradecanoyl-phorbol-13-acetate (TPA). LY117018, however, inhibits the PGI~2~production stimulated by tamoxifen. The intracellular release of AA and/or the cellular reactions that accompany that release may initiate pathways that prevent tumor growth. The tissue specific effects of tamoxifen and LY117018 may be associated with the AA or with cyclooxygenase (COX) activity and/or one of the many bioactivities resulting from oxygenation and metabolism of the released AA. Methods ======= The C-9 rat liver and BT-20 human breast carcinoma cells were purchased from the American Type Culture Collection (Manassas, VA, USA) and maintained in MEM supplemented with 10% fetal bovine serum. The C-6 rat glial cell line was obtained from Dr. Elaine Y. Lai in the Department of Biology, Brandeis University and maintained in medium 199. The human colon carcinoma cells (HT-29) were obtained from Dr. Basil Rigas, American Health Foundation, Valhalla, NY and maintained in McCoy\'s medium. \[^3^H\]AA (91.8 Ci/mmol) was purchased from NEN Life Science Products, Inc. (Boston, MA, USA); ICI-182,780 from Tocris Cookson, Inc. (Ballwin, MO, USA); tamoxifen and 4-OH-tamoxifen were from Sigma Chemical Co. (St. Louis, MO, USA). LY117018 was obtained from Dr. David A. Cox, Eli Lilly and Co. (Indianapolis, IN, USA). Raloxifene was extracted from EVISTA^®^tablets with dimethylsulfoxide. Two days prior to experiments, the cells were treated with 0.25% trypsin-EDTA and, after addition of minimum essential medium (MEM), medium 199 or McCoy\'s medium containing 10% fetal bovine serum, the floating cells were seeded onto 35 mm culture dishes. The plating densities varied from 0.1 to 0.5 × 10^5^cells/35 mm dish. The freshly seeded cultures were incubated for 24-h to allow for cell attachment. After decantation of incubating media, 1.0 ml fresh media (MEM for the rat liver and BT-20 cells, medium 199 for the rat glial or McCoy\'s for the HT-29 cells respectively) containing 10% fetal bovine serum and \[^3^H\] AA (0.2 μCi/ml) was added and the cells incubated for 24-h. The cells were washed 4 times with MEM and incubated for various periods of time with 1.0 ml of MEM, medium 199 or McCoy\'s containing 1.0 mg bovine serum albumin (BSA)/ml and different concentrations of each test compound. The culture fluids were then decanted, centrifuged at 2000 × g for 10 min, and 200 μl of the supernate counted for radioactivity. Radioactivity recovered in the washes before the 6-h incubation was compared to input radioactivity to calculate the % radioactivity incorporated into the cells \[[@B5]\]. As measured by thin layer chromatography (TLC), of the \[^3^H\] released from radiolabelled methylcholanthrene transformed fibroblasts (about 20% after a 3-h stimulation by serum), 92% was AA, 4% was PGE~2~, 0.6% and 1% were PGF~2α~and phospholipids respectively \[[@B6]\]. When labelled to equilibrium, the \[^3^H\] AA had been incorporated into phosphatidylcholine (50%), phosphatidylethanolamine (36%), phodphatidylserine (9%) and triglycerides (10%) \[[@B7]\]. Such distribution varied among 12 cell lines \[[@B8]\]. The \[^3^H\] AA label released from human colorectal cancer cell lines (HCT116 and SW180) was AA as measured by TLC \[[@B9]\]. For PGI~2~production by the rat liver cells, 1.0 ml of MEM supplemented with 10% fetal bovine serum, lacking \[^3^H\] AA, was added after the first 24-h incubation. The cells were incubated for another 24-h, washed three times with MEM, then incubated with lactacystin and TPA in the presence and absence of the test compounds in MEM/BSA for various periods of time. The culture fluids were decanted and analyzed for 6-keto-PGF~1α~, the stable hydrolytic product of PGI~2~, by radioimmunoassay \[[@B10]\]. At the sub-confluent cell densities used in these experiments (about 5 × 10^4^or less per dish) only the major COX product of rat liver cells (97% 6-keto-PGF~1α~) could be measured. The other two COX products, PGE~2~(2%) and PGF~2α~(1%) could not be measured. The release of \[^3^H\] AA is presented as a percentage of the radioactivity incorporated by the cells. Except for the time-course experiments, which used duplicate dishes (Fig. [5](#F5){ref-type="fig"}), three to six culture dishes were used for each experimental point. The data are expressed as mean values ± SEM. The data were evaluated statistically by the unpaired *Student\'s t-test*. A *P*value \< 0.05 was considered significant. Results ======= AA Release ---------- The release of AA after incubation of the rat liver cells with tamoxifen, 4-OH-tamoxifen, LY117018 and raloxifene for 6 h is shown in Fig. [1-A](#F1){ref-type="fig"}. Tamoxifen is the most effective: it releases AA at 10 μM. 4-OH-tamoxifen, an active metabolite of tamoxifen \[[@B2]\] is also effective. Raloxifene and LY117018 release AA from the rat liver cells, but are about 5 times less potent (at the doses tested) than tamoxifen. At μM concentrations, all induce apoptosis \[[@B11],[@B12]\]. Tamoxifen, but not 4-OH-tamoxifen, has also been reported to induce apoptosis in p53(-) normal human mammary epithelial cells \[[@B13]\]. Yet, the metabolite binds to the ER with higher affinity than tamoxifen \[[@B14],[@B15]\]. The release of AA by tamoxifen and LY117018 is not unique to the rat liver cells. Both release AA from rat glial cells (Fig. [1-B](#F1){ref-type="fig"}). Tamoxifen is more effective at releasing AA than is LY117018. The highest level of \[^3^H\] AA released into the medium which contains BSA (1 mg/ml) would be 0.25 nM (the specific activity of the \[^3^H\] AA is 92 Ci/mmole). Since TLC or HPLC analyses were not done, the concentration of the total released AA could not be quantitatively estimated. However, even treatment of a colorectal cancer cell line with 200 μM AA for 48-h leads to apoptosis \[[@B9]\]. The release of AA is neither species nor tissue specific. Both drugs also release AA from human breast carcinoma (BT-20) and human colon carcinoma (HT-29) cells (Fig. [2-A](#F2){ref-type="fig"} and [2-B](#F2){ref-type="fig"}). Again, LY117018 is less effective than tamoxifen. PGI~2~production ---------------- Of the cultured cells studied, the COX mediated metabolic profile of AA has been described only in the rat liver cells. Thus, the effects of tamoxifen and LY117018 on PGI~2~production by these cells were determined. Tamoxifen stimulates basal PGI~2~production as well as that induced by lactacystin in the presence of TPA. LY117018 and raloxifene (Fig. [3-A](#F3){ref-type="fig"}) are ineffective in stimulating basal PGI~2~production. LY117018, at a concentration that does not affect PGI~2~production (50 μM), inhibits the PGI~2~production stimulated by lactacystin in the presence of TPA (Fig. [3-B](#F3){ref-type="fig"}). The increased PGI~2~production probably reflects some stimulation of both induced and basal COX activity. In mouse neuronal and embryonic rat mesencephalic cells, COX-2 is induced by lactacystin as measured by western and northern blot analyses \[[@B16]\]. Identification of the induced and basal isoforms stimulated by tamoxifen in these rat liver cells was deduced from the relative effectiveness of inhibition of PGI~2~production by celecoxib and piroxicam. Celecoxib is 7.5 times more effective than piroxicam as an inhibitor of COX-2, but 30,000 times less effective as an inhibitor of COX-1. Piroxicam is a selective inhibitor of COX-1 \[[@B17]\]. Celecoxib, the nonsteroidal anti-inflammatory drug (NSAID) that is a selective inhibitor of COX-2, inhibits both induced and basal COX activity 20--80 times more effectively than piroxicam (Fig. [4-A](#F4){ref-type="fig"} and [4B](#F4){ref-type="fig"}). Thus, based on inhibition by these two NSAIDs, COX-2 is the most likely isoform expressed both constitutively and after induction by lactacystin in the presence of TPA in rat liver cells. This conclusion was confirmed by western blot analyses (Levine L, Tashjian A, unpublished data). Is COX-2 production non-genomic and ER-independent? --------------------------------------------------- The time-course of stimulation by tamoxifen (15 μM) of basal activity is shown in Fig. [5](#F5){ref-type="fig"}. Even after a 5-minute incubation (not shown in Fig. [5](#F5){ref-type="fig"}), stimulation by tamoxifen is significant statistically \[0.08 ± 0.011 (n = 5) *vs*0.13 ± 0.011 (n = 5) ng 6-keto-PGF~1α~per ml in the absence and presence of tamoxifen, respectively\]. Similar to AA release \[[@B4]\], stimulation of basal PGI~2~production by tamoxifen is not affected by preincubation of the cells for 2-h with 1 μM actinomycin D. The induction obtained by incubation with lactacystin in the presence of TPA is inhibited (Fig. [6-A](#F6){ref-type="fig"}). Stimulation of basal PGI~2~production by tamoxifen is unaffected by ICI-182,780 (Fig. [6-B](#F6){ref-type="fig"}). This estrogen antagonist, which binds to the ER with high affinity \[[@B18]\] did not affect the release of AA by tamoxifen \[[@B4]\] but does partially inhibit release of AA stimulated by 17β-estradiol, 22 (R)-cholesterol, indomethacin, *trans*-retinoic acid and the tyrosine analog of thiazolidinedione, GW7845 \[[@B19]\]. The rapidity of stimulation (5 min) and the lack of effect of actinomycin D or ICI-182,780 suggest that the stimulation of PGI~2~production by tamoxifen is non-genomic and does not require ER occupancy. LY117018 blocks PGI~2~production stimulated by tamoxifen (Fig. [7](#F7){ref-type="fig"}). This blockade is unaffected by preincubation of the cells for 2-h with either actinomycin D or ICI-182,780 (data not shown). LY117018, however, does not inhibit the release of AA stimulated by tamoxifen; their combined effect is at least additive (Fig. [8](#F8){ref-type="fig"}). Discussion ========== These studies demonstrate that, at μM concentrations, tamoxifen and the raloxifene analog, LY117,018 stimulate the release of AA from cells in culture. At nM levels, the release is not observed. The effectiveness of these two compounds at prevention of estrogen-dependent breast cancer reflects competition for the ER \[[@B1],[@B2]\]. In addition to occupancy of the ER, I am postulating that these drugs, at μM concentrations, may also prevent breast cancer of estrogen independent tumors. Consistent with this hypothesis are the findings that even at 100 μM concentrations, ethanol extracts of Femora ^®^(letrozole) or Aramidex ^®^(anastrozole) do not release AA from cells in culture (unpublished data). These two drugs inhibit estrogen synthesis by blocking aromatase enzymes and also prevent estrogen-dependent breast cancer \[[@B20]\]. In view of the stimulations of AA release by tamoxifen and LY117018 from human colon carcinoma cells (Fig. [2-A](#F2){ref-type="fig"}), the occurrence of colon cancer in women undergoing long term treatment with high levels of tamoxifen would be of interest. A mechanism that most simply explains the release of AA by tamoxifen and LY117018 is the ability of such compounds to intercalate into cell membranes and affect phospholipase activities. The release of AA from endothelial cells by the Ca^2+^ionophore A-23187 reflects phospholipase activities. It is regulated by phosphorylation of the enzyme \[[@B21]\]. The enzymic properties of the altered membrane may impact signaling mechanisms e.g., pathways leading to apoptosis and COX induction. The intracellularly released AA also can serve as substrate for oxygenases. AA has intrinsic biologic activities that may also affect signaling pathways. Such changes would depend on the lipophilic properties of the compound and the composition of a particular membrane \[[@B22]\] and would vary from cell type to cell type, organelle to organelle and with the growth phase of the cell. The AA release by tamoxifen and other reagents studied in my laboratory occurs with μM concentrations \[[@B4],[@B5],[@B19],[@B23],[@B24]\]. These experiments were carried out in the presence of BSA (1.0 mg/ml), and therefore do not differentiate between the protein-bound and free reagent. Thus, they are likely to be overestimated values. Nevertheless, the possibility that general necrotic cell death may cause AA release, must be considered. Tamoxifen, was found not to be toxic at concentrations of 10 to 20 μM for A549 human lung adenocarcinoma (ER-negative) cells \[[@B25]\]. Nor was 10 μM tamoxifen toxic when tested on rat glial cells and breast cancer MCF-7 cells \[[@B26]\]. Even when cell viability of three different breast cell lines (ER-positive MCF-7; ER-negative MDA-MB-239 and ER-negative BT-20 cells) was measured after incubation with 25 μM tamoxifen for 24-h, the loss in viability was due to apoptosis \[[@B12]\] and was not the result of necrotic cell death. Concentrations of tamoxifen used in this report are comparable to those found to induce apoptosis, not necrotic cell death. The median concentration of tamoxifen and its metabolites for clinical effectiveness in the treatment of breast cancer varies from 0.8 μM to 2.4 μM, depending on the age of the woman \[[@B27]\]. When measured after a 6-h incubation, 10 μM tamoxifen stimulates deesterification of membrane phospholipids as measured by extracellular AA release, (Fig. [1](#F1){ref-type="fig"} and [2](#F2){ref-type="fig"}). The 6-h incubation may not be optimum for apoptosis to be observed, e.g. after a 6-h incubation of 5 μM tamoxifen with breast cancer cells, apoptosis was induced in about 10% of the cells compared to 8% in control cells \[[@B12]\]. After 24- and 48-h incubation, apoptosis in the tamoxifen treated cells increased to 40 and 70%, respectively. Apoptosis, induced by this membrane perturbation after 6-h incubation, increased 4 to 7 fold after longer incubation. It is apoptosis that may mediate the cancer preventative action of tamoxifen \[[@B11],[@B12]\]. Some tissue specific effects of tamoxifen and raloxifene, e.g. on endometrial cells of the uterus, may be related to COX activity. At the low cell densities used in the present studies, the PGI~2~produced by the rat liver cells can be quantitatively determined. The effects of tamoxifen and LY117018 on the COX activity of the rat liver cells may be similar to their effects on other cells that express COX. Tamoxifen and LY117018 affect COX activity differently; only tamoxifen stimulates PGI~2~production. AA, which regulates the production of lipoxygenase, COX and cytochrome P-450 epoxygenase products could impact many of the pharmacological actions of these two selective estrogen receptor modulators. AA can be oxygenated by COX isoforms, lipoxygenases, and cytochrome P-450 epoxygenases and their products converted to the prostaglandins, leukotrienes, epoxyeicostetranoic acids \[[@B28]\] and AA to the oxidative stress-related isoprostanes \[[@B29]\]. In the COX pathway, enzymes convert PGH~2~to the major physiologically active products, PGD~2~, PGE~2~, PGF~2α~, PGI~2~and thromboxane A~2~\[[@B28]\]. They, in turn, can be converted enzymatically and nonenzymatically to other biologically active compounds that possess different pharmacological properties. Often, only one major product is synthesized by the cell \[[@B8]\]. Thus, tissues can be affected differently by changes in COX activity. AA or its metabolites, would have different biological effects depending upon the genetic capabilities of the individual cell being affected. In addition to the spectrum of activities of the prostanoids, oxygenation of AA by 5-, 12- and 15-lipoxygenase and epoxygenases yield other products located in different cells and tissues \[[@B30],[@B31]\]. Competing interests =================== None declared. Author\'s contributions ======================= L.L. is the sole author and all of the experiments, with the exception of the western blots referred to as Levine L and Tashjian A (unpublished data), were conducted by L.L. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2407/4/49/prepub> Acknowledgements ================ My thanks to Dr. Armen H. Tashjian, Jr., Department of Genetic and Complex Diseases, Harvard School of Public Health, for his continuing interest in these studies and Hilda B. Gjika for preparation of the manuscript. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### AA release from (A) rat liver cells (C-9) stimulated by tamoxifen, 4-OH-hydroxy-tamoxifen, LY117018 and raloxifene and (B) from rat glial cells incubated with tamoxifen and LY117018. The cells were incubated for 6-h. The analyses were performed with triplicate dishes. Each bar gives the mean value and the brackets give the SEM. \* = statistically significant *vs*control. The data are representative of several independent experiments with similar results. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### AA release from (A) human breast carcinoma cells (BT-20) and (B) human colon carcinoma cells (HT-29) by tamoxifen and LY117,018. The cells were incubated for 6-h. The analyses were performed with triplicate dishes. Each bar gives the mean values and the brackets give the SEM. \* = statistically significant *vs*appropriate MEM/BSA control. The data are representative of several independent experiments. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Effect of tamoxifen, LY117018 or raloxifene on (A) basal 6-keto-PGF~1α~production and (B) the effect of tamoxifen or LY117018 on 6-keto-PGF~1α~production induced by lactacystin in the presence of TPA. Rat liver cells were incubated for 6-h. The analyses were performed with triplicate or quadruplicate dishes. Each bar gives the mean value and the brackets the SEM. \* = Statistically significant *vs*control stimulation. \*\* = Statistically significant *vs*lactacystin plus TPA control. The data are representative of several independent experiments with the same results. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Inhibition of (A) tamoxifen (16 μM)-stimulated basal and (B) tamoxifen-stimulated induced 6-keto-PGF~1α~production (lactacystin in the presence of TPA) by celecoxib (\*) and piroxicam (◊). Rat liver cells were incubated with either (A) tamoxifen (16 μM) or (B) lactacystin plus TPA in the presence of tamoxifen, for 6-h. The analyses were performed with triplicate dishes. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Time-course of 6-keto-PGF~1α~production by rat liver cells during incubation with 15 μM tamoxifen (γ) and MEM/BSA control (O). Analyses were performed with duplicate and triplicate dishes; mean values are shown. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### Effect of (A) actinomycin D on basal or induced (lactacystin plus TPA) 6-keto-PGF~1α~production stimulated by tamoxifen and (B) effect of ICI-182,780 on basal 6-keto-PGF~1α~production stimulated by tamoxifen. In (A), rat liver cells were preincubated with 1 μM actinomycin for 2-h and then incubated with tamoxifen (16 μM) or lactacystin plus TPA for 6-h. In (B), rat liver cells were incubated with tamoxifen (16 μM) and tamoxifen in the presence and absence of ICI-182,780 (16 μM) for 6-h. The analyses were performed with triplicate dishes. Each bar gives the mean value and the brackets give the SEM. Parts of the experiments were repeated several times with the same results. \* Statistically significant *vs*control. \*\* Statistically significant *vs*lactacystin plus TPA. :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### Effect of tamoxifen, LY117018 and tamoxifen plus LY118018 (all at 16 μM), on 6-keto-PGF~1α~production. The rat liver cells were incubated 6-h. The analyses were performed with quintuplicate dishes. Each bar gives the mean value and the brackets the SEM. \* Statistically significant *vs*tamoxifen. :::  ::: ::: {#F8 .fig} Figure 8 ::: {.caption} ###### AA release from rat liver cells by tamoxifen (12 μM), LY117018 (48 μM) and tamoxifen in presence of LY117018. Cells were incubated for 6-h. The analyses were performed with triplicate dishes. Each bar gives the mean value and the brackets the SEM. \* = Statistically significant *vs*MEM/BSA control. \*\* = Statistically significant *vs*tamoxifen. :::  :::

PubMed Central

2024-06-05T03:55:47.214084

2004-8-13

PMC514707

Background ========== Retinopathy of Prematurity (ROP) is a common retinal neovascular disorder of premature infants. It is of variable severity, usually heals with mild or no sequelae, but may progress in some infants to partial vision loss or blindness from retinal detachments or severe retinal scar formation \[[@B1]\]. ROP remains as one of the most frequent cause of blindness in children, in particular in countries with infant mortality rates between 10 and 60/1000 \[[@B2],[@B3]\]. Among 177 students attending schools for children with visual impairment in the city were this study was conducted, 107 (60.5%) had ROP \[[@B4]\]. The incidence of both any acute ROP, and of the more severe stages, varies inversely with gestational age at birth. ROP is unusual (except in the mildest forms) in infants of greater than 31 weeks gestation, and severe complications such as retinal detachment occur in less than one half of one percent of infants greater than 31 weeks gestation. However, more than 80% of infants less than 28 weeks gestation develop some ROP, and around 10% develop \"threshold ROP. In threshold ROP more that 40% of the cases progresses to retina folds or detachment, with its consequent blindness. In this stage, ablative surgery (cryotherapy or laser photocoagulation) to the peripheral avascular retina is recommended to reduce the risk of disease progression to retinal detachment \[[@B5]-[@B7]\]. Pre-threshold stage has been linked to bad results in the visual function: reduction of visual acuity, short-sightedness, amblyopic, etc. \[[@B8],[@B9]\]. A number of strategies have been developed to try to diminish the progress of ROP, but with limited success. These strategies include antioxidants such as vitamin E \[[@B10]\], D penicillamine \[[@B11]\] and allopurinol \[[@B12]\], reduction of exposure to light \[[@B13]\] and supplementation with oxygen \[[@B14]\]. The active disease appears in the premature about 4 to 8 weeks after birth. In this period, the levels of vascular endothelial growth factor (VEGF) increase in the retina, as well as other chemical mediators of inflammation such as platelets activator factor (PAF), prostaglandins (PGs), and eicosanoids which would put again under way the process of vascularization that had stopped in the period of oxidative injury. This vascularization is now degenerated and invasive \[[@B15]-[@B17]\]. In models of animal experimentation it was possible to diminish the degree of retinal neovascularization with use of indometacin \[[@B18]\], dexamethasone \[[@B19]\], rofecoxib \[[@B20]\], and bucillamine \[[@B21]\], and increased activity of cyclooxygenase 2 (COX2) was also demonstrated in vessels of neo-proliferation in retina and its inhibition decreases the neovascularization in 37% \[[@B20]\]. Ketorolac is a non-steroid anti-inflammatory drug (NSAID) derived from indometacin. Its mechanism of action is developed through the interruption of the synthesis of prostanoids, inhibiting the way of COX 1 and 2 in arachidonic acid metabolism; in this way the tissue levels of prostaglandin F2alfa and thromboxane B2 decrease. Its adverse effects are linked predominantly to their inhibitory action of platelet aggregation. It does not alter either the platelet count, or factors of clotting. High digestive hemorrhage is the principal adverse reaction. Nervous and cardiovascular systems are not generally affected by the use of ketorolac to habitual doses \[[@B22]-[@B24]\]. Ketorolac administered as conjunctival topical diminishes prostaglandin E2 concentration in aqueous humor, without modifying the intraocular pressure \[[@B24]\]. In addition it is effective in uveitis induced by tumors necrosis factor (TNF) \[[@B25]\]. Ketorolac is unquantifiable in plasma when administered in ophthalmic drops \[[@B24]\]. Ketorolac ophthalmic solution is usually used in older adults with retinal disorders. It is used to diminish the Cystoid Macular Edema that complicates the surgery of cataracts. In this pathology ketorolac has proved to be effective in diminishing the macular edema and improving the visual acuity, providing evidence that their conjunctival instillation produce effects on the most internal layers of the eye. The ophthalmic use of ketorolac only reports occasional episodes of discomfort and ocular burning \[[@B26]-[@B28]\]. The use of ketorolac ophthalmic solution in pediatrics is frequent as an analgesic in corneal abrasions, and in allergic and post surgical conjunctivitis. The FDA recognizes its indication for allergic conjunctivitis, ocular pain, post surgical ocular inflammation, ocular pruritus and photophobia \[[@B24],[@B29]\]. On the basis of experimental evidence and physio-pathogenic rationality, we treated preterm newborns admitted to neonatal intensive care unit (NICU) of our hospital with ketorolac in ophthalmic drops with the aim of decreasing the progression and severity of ROP. This study compares this group of children treated with ketorolac with historic controls to assess the impact of this treatment on the incidence of severe ROP and adverse effects. Methods ======= This is a preliminary report, that compares a cohort of children treated with ketorolac with historic controls that did not received such treatment. From January 2001 to December 2002 all preterm newborns with birthweight less than 1250 g or gestational age less than 30 weeks of gestational age, admitted in the NICU of the University Hospital of Maternity and Neonatology of the city of Córdoba, Argentina, were eligible for treatment with topical ketorolac. The comparison group included all preterm newborns, with the same inclusion criteria, admitted between January 1999 and December 2000. None of these newborns received treatment with ketorolac. There were no differences either in treatment guidelines, equipment or in the number of physicians and nurses that took care of the patients in the two periods. All the children were examined by the same group of ophthalmologists, which have many years of experience in treating patients with ROP. The international classification for ROP \[[@B30]\] was used to define the stages of the disease and verified by more than one observer. During the period in which ketorolac tromethamine was used, when risk signs for ROP were identified (zone I incomplete vascularization, vessels only in zone of transition I-II or anomalous ramification and equatorial incurvation of vessels in the avascular -- vascular junction) \[[@B31]\] treatment was started with a drop of ketorolac tromethamine (0.25 mgrs.) every 8 hours in each eye. The treatment continued until they presented signs of threshold ROP and cryotherapy was indicated, or till the resolution of the condition. Parents of these children gave their consent so that their children could receive treatment. The variables considered to assess the comparability among both groups were: birthweight, Apgar score at birth less than 6 at 5 minutes, duration of oxygen therapy, peri-intraventricular hemorrhage (PIVH) equal or greater than 3 degrees, necrotizing enterocolitis (NEC) equal of greater than II degrees, and late sepsis. The analyses included only those children that were alive at discharge, but the number of deaths was reported for both study periods. The presence of undesirable effects of ketorolac such as hemorrhages, oliguresis, local manifestations of intolerance, and conjunctival infection were analyzed. Fischer exact test was used to obtain two tailed p-values. Relative risks (RR) with 95% confidence intervals where also calculated. Adjusted analyses were performed by using Poisson regression with robust estimates \[[@B32]\]. Statistical analysis was carried out using the statistical software SPSS version 8.0, Stats Direct, and STATA version 8.0. Results ======= During the analyzed period 112 eligible preterm newborns were admitted, 53 between 1999 and 2000 (historic controls) and 59 between 2001 and 2002 (ketorolac group). The number of newborns that died before discharge was similar between groups (see table [1](#T1){ref-type="table"}). Table [1](#T1){ref-type="table"} presents the characteristic of those alive at discharge. Groups were comparable in terms of weight distribution, Apgar score at 5 minutes, incidence of sepsis, intraventricular hemorrhage and necrotizing Enterocolitis (see table [1](#T1){ref-type="table"}). The duration of oxygen therapy was significantly longer in the control group (see table [1](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Characteristics of newborns treated with ketorolac and historic controls. ::: **Historic controls** **Ketorolac** **p\*\*** ---------------------------------------------------- ----------------------- --------------- ----------- **Dead before discharge\*** 18 (34.0) 16 (27.1) 0.54 **Birthweight (g)** 0.28  \<= 750 4 (11.4) 8 (18.6)  751--1000 13 (37.1) 9 (20.9)  1001--1250 18 (51.4) 26 (60.59 **Apgar Score \< 6 at 5 minutes** 10 (28.6) 11 (25.5) 0.80 **Oxygen (days)** 0.01  0--10 9 (25.7) 21 (48.8)  11--27 10 (28.6) 15 (34.9)  \>= 28 16 (45.7) 7 (16.3) **Sepsis** 16 (45.7) 24 (55.8) 0.50 **Peri--intraventricular hemorrhage \> grade III** 8 (22.9) 6 (14.0) 0.38 **Necrotizing Enterocolitis \>= II degree** 5 (14.3) 9 (20.9) 0.56 † Excludes newborns that died before discharge. \* N = 53 in historic controls group and N = 59 in ketorolac group, that includes all eligible live newborn during the study period. \*\* P-value Fisher exact test. ::: In the ketorolac group 45 children were alive at the first ophthalmic control, two died before discharge due to late sepsis, so 43 received treatment with ketorolac. Nineteen children were discharged from NICU with treatment indication, and continued with ketorolac up to 44 weeks of gestational age, when the ophthalmologists considered that the risk for developing ROP was very small. The incidence of threshold ROP in newborns treated with ketorolac was significantly lower (Relative Risk Reduction 86%) than in the control group (see table [2](#T2){ref-type="table"}). Adjusting for duration of oxygen therapy did not significantly change these results (see table [2](#T2){ref-type="table"}). Further exploration of this relationship showed that duration of oxygen therapy equal or higher than 28 days was not associated with severe ROP in this data (Relative Risk 1.04, 95%CI 0.25 to 4.51, p = 0.28). This finding explains why duration of oxygen therapy does not act as a confounder. Hemorrhages were not observed in the vitreous after treatment with ketorolac. In four cases hemorrhages in the vitreous were already present at the beginning of therapy and they disappeared after 14 days of treatment. No signs of local intolerance, or conjunctival infection were observed. We did not find hemorrhages in other organs attributable to the drug, or signs of renal failure. Treatment was not suspended in none of the cases and all preterm babies received ketorolac until its interruption due to the resolution of ROP or the indication of cryotherapy. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Incidence of threshold retinopathy of prematurity in newborns treated with ketorolac and historic controls. Only those alive at discharged included in the analyses. ::: **Historic Controls** **Ketorolac** **Relative Risk (95%CI) P-value** ------------------------------------------ --------------- ----------------------------------- ------------------------------- -------------------------- **Threshold retinopathy of prematurity** 6/35 (17.1) 1/43 (2.3) 0.14 (0.00 to 0.80) 0.041\*\* 0.12 (0.02 to 0.92) 0.04 \* Adjusted by oxygen administration using robust Poisson Regression. \*\*Fisher exact test ::: Discussion ========== In this preliminary study we have shown that the incidence of severe ROP was significantly lower in very preterm newborns treated with ketorolac, compared with historic controls not receiving such treatment. These results suggest that administration of ketorolac as an ophthalmic solution might be an effective preventive strategy in patient at risk of developing severe ROP. The study was conducted on non-concurrent patient groups, and changes in the subjects risk profile or quality of care between the two study periods might have an impact in the risk of developing severe ROP. The magnitude of the effect of ketorolac on the incidence of severe ROP found in this study (from 17% in the control group to 2.3% in the ketorolac group) is large and in our opinion unlikely to be explained completely by confounding factors, although this possibility cannot be completely ruled out. To our knowledge, no significant change in the standard of care took place between the study periods, but mortality was lower during the period were ketorolac was used, although the difference was not statistically significant. The groups were comparable in terms of their birthweight distribution, Apgar score, and indicators of severe morbidity. A significant difference was found in duration of oxygen therapy, but in our data there was no association between duration of oxygen therapy and severe ROP, and adjusting for oxygen did not change the results. The principal factor involved in the genesis and severity of ROP is the injury due to O2 radicals; there were no substantial changes in the management of O2 supply to the children of our unit in both periods, the equipment used to measured O2 saturation and to administer the oxygen. The prevalence of severe ROP in our maternity was relatively high, considering that for 1998 the Vermont-Oxford Network reported an incidence of 9.5% for severe ROP and of 57.2% for ROP of any degree \[[@B33]\], but similar to the one reported by other units in our country and in others countries with similar development \[[@B34]-[@B36]\] Ketorolac was apparently safe. No patients in the group that received ketorolac presented oliguric, or with biochemical signs of renal failure during treatment. There were no hemorrhagic manifestations that could be attributed to ketorolac; hemorrhages observed in vitreous evolved favorably with collyrium. Neither local intolerance episodes to the drops nor purulent conjunctivitis among the treated cases were observed. The effect on ROP we have observed is compatible with an impact of Ketorolac in the active phase of neovascularization, a fact that is also observed in numerous reports in animal models of ROP with NSAIDs and steroid drugs administered systemically. The hypothesis of an inflammatory component in active ROP has been quite recently recognized by experts, extensive leukocyte adhesion was observed at the leading edge of pathological neovascularization \[[@B17]\]. Recent research work with inhibitors of COX2 shows that anti-inflammatory drugs apparently have an impact on the evolution of ROP \[[@B20]\]. Neufeld and collaborators found high plasmatic profiles of TNF and other cytokines during the phase of installation of the pre-threshold stage for ROP \[[@B37]\]. COX has a strong angiogenic action in the normal development of retina. The inhibition of COX2 diminishes the angiogenesis in cancer and rheumatoid arthritis \[[@B38]\]. The ganglion cells of retina secrete PGs that would interact with angiogenic substances as VEGF and Insulin like growth factor. The neo-vessels express a greater concentration of COX2 and inhibitors of COX2 stop the angiogenesis mediated by VEFG, which is the principal factor involved in the neovascularization and which would be regulated by the PGs secreted by ganglion cells of retina and endothelial cells \[[@B20]\]. The hyperoxia induces liberation of TNF with a powerful inflammatory action; dexamethasone interferes with TNF production attenuating the manifestations of retinopathy by hyperoxia \[[@B19]\]. Topical ketorolac is very effective in neutralizing uveitis generated by experimental infection with monoclonal TNF or bacterial endotoxins \[[@B25],[@B39],[@B40]\]. ROP is a problem that modern medicine has generated due to the survival of very small preterm children and which has been unable to solve so far. About 2% of the children with birthweight less than 1500 grams are blind because of this pathology. Current preventive interventions can prevent blindness in nearly 70% of the cases. Earlier treatment using ablation of the avascular retina in pre-threshold ROP has been proposed \[[@B41]\], but a preventive measure that could avoid the need of ablation would be much better. Conclusions =========== This preliminary report provides evidence suggesting that ketorolac in the form of an ophthalmic solution can be used to reduce the risk of developing severe ROP in very preterm newborns, without producing significant adverse side effects. These results, although promising, should be interpreted with caution because of the weakness of the study design. We cannot exclude the possibility of bias been responsible of the observed effect. On the other hand, if confirmed, these results can have important public health implications. This is an inexpensive and simple intervention that might ameliorate the progression of a disease with devastating consequences for children and their families. We believe that next logical step would be to assess the effectiveness of this intervention in a randomized controlled trial of adequate sample size to provide a definite answer to this research question. Competing interest ================== None declared. Authors\' contributions ======================= MAV and RM conceived the study. MAV, MLC, EB and RM design the study. MF carried out the revision of the clinical histories. MS carries out the ophthalmologic exams of all the cases. RM verified the ophthalmologic finds. EB carried out the statistical analysis. MAV and EB drafted the article. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2431/4/15/prepub> Acknowledgements ================ To PROGRESAR-ICMER that granted a fellowship to Medardo Avila-Vazquez to carry out a training in Epidemiology and Methodology of the Clinical Investigation in CLAP, PAHO-WHO where this report was elaborated and also generates a protocol of Randomized Clinical Trial to continue this investigation.

PubMed Central

2024-06-05T03:55:47.216069

2004-8-7

PMC514708

Background ========== Schizophrenia is a severe mental disorder characterized by hallucinations, delusions, disorganized thoughts, and various cognitive impairments. The life-time prevalence is about 1%, and genetic factors were known to play a critical role in its pathogenesis \[[@B1]\]. Based on the fact that phencyclidine (PCP) induces schizophreniform psychosis, a glutamatergic dysfunction hypothesis has been proposed for the pathogenesis of schizophrenia \[[@B2]-[@B4]\]. This hypothesis has been supported by recent multiple reports of association of schizophrenia with glutamate receptor genes and with the genes related to glutamatergic transmission, such as *G72*and *NRG1*\[[@B5]-[@B10]\]. In addition, other synaptic elements related to glutamate, such as excitatory amino acid transporters (EAATs), also potentially affect glutamatergic neurotransmission. EAATs maintain extracellular glutamate concentrations within physiological levels by reuptaking the synaptically released glutamate. A deficient uptake has been implicated in the pathogenesis of ischemic brain damage \[[@B11]\] and may be involved in neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS) \[[@B12]\]. Recently significant increases of mRNA expression of EAAT1 and EAAT2 have been reported in the thalamus of schizophrenics, suggesting the possibility that an excessive glutamate uptake is involved in schizophrenia \[[@B13]\]. On the other hand, a significant decrease of EAAT2 mRNA expression was observed in the parahippocampal gyrus of schizophrenics \[[@B14]\]. Therefore the EAAT genes are reasonable candidates for schizophrenia, as well as glutamate receptor genes. The EAATs family consists of five members (EAAT1-EAAT5). Their cellular localizations are different: EAAT1 and EAAT2 are astroglial, whereas EAAT3 EAAT4 and EAAT5 are neuronal \[[@B25]\]. Since EAAT2 accounts for approximately 90% of glutamate reuptake in the rodent forebrain \[[@B16],[@B17]\], we focused on the EAAT2 gene (*SLC1A2*) in association studies of schizophrenia. *SLC1A2*has been mapped to 11p13-p12 \[[@B18]\] and consists of 11 exons spanning over 165 kb. In this study we tested associations of schizophrenia with 11 SNPs distributed in *SLC1A2*with an average interval of 15.9 kb. To enhance the detection power of the study, we also examined the haplotype associations of the SNPs with the disease. Methods ======= Human subjects -------------- Blood samples were obtained from unrelated Japanese individuals who had provided written informed consent. We used two Japanese sample sets in this study. In the first one, Kyushu samples, 100 schizophrenia patients (mean age 49.5; 44.0% female) were recruited from hospital in the Fukuoka and Oita areas and 100 healthy unrelated controls (mean age 51.2; 44.0% female) were recruited from the Fukuoka area. In the initial SNP selection process, we used another 16 Japanese samples which are recruited in the Fukuoka area and informed in the same way. In the second one, Aichi samples, 100 schizophrenia patients (mean age 34.4; 44% female) and 100 healthy unrelated controls (mean age 39.9; 45% female) were collected in the Aichi area about 600 km east of Fukuoka. All patients fulfilled the DSM-IV criteria for schizophrenia \[[@B19]\]. All of the case and control samples are ethnically Japanese. DNA samples were purified from whole peripheral blood by the method previously described \[[@B20]\]. This study was approved by the Ethics Committee of Kyushu University, Faculty of Medicine and the Fujita Health University Ethics Committee. SNP selection in the SLC1A2 region ---------------------------------- We retrieved the primary SNP information from the dbSNP database <http://www.ncbi.nlm.nih.gov/SNP/>. Assuming the same size of the half length of linkage disequilibrium (LD) (60 kb) as reported in Caucasians \[[@B21]\], we initially intended to select common SNPs every 50 kb in *SLC1A2*. We tested 22 candidate SNPs including all of the exonic SNPs, in the 16 healthy Japanese samples by the direct sequencing method. Out of the 22 SNPs we selected the following 7 common SNPs with minor allele frequencies over 10% for further analyses: SNP1, rs1923295; SNP3, rs4534557; SNP6, rs1885343; SNP8, rs752949; SNP9, rs1042113; SNP10, rs3838796; SNP11, rs1570216. We also identified a novel SNP, SNP7, in intron 1 (conting location: 34105026). After the LD analyses described below, we noticed LD gaps (*D*\' \< 0.3) of the initial SNP set and examined additional 20 candidate SNPs. Out of the 20 SNPs, we selected the following 3 SNPs to fill the LD gaps: SNP2, rs4755404; SNP4, rs4756224; SNP5, rs1923298. The locations of the total 11 SNPs are shown in Figure [1](#F1){ref-type="fig"}. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Genomic organization of *SLC1A2*and locations of the SNPs.**Exons are shown as vertical bars with exon numbers. Eleven SNPs are indicated by circles. Distances between the SNPs are indicated above with kb. :::  ::: Genotyping ---------- Eleven SNPs were amplified as 11 individual fragments by PCR using the primers shown in Table 1 - [additional file 1](#S1){ref-type="supplementary-material"}. The reaction mixture for PCR was prepared in a total volume 10 μl with 5 ng of genomic DNA, 10 pmol of each primer (4 pmol of SNP3), 2.5 mM of MgCl~2~, 0.2 mM of each dNTP and 0.25 U of *Taq*DNA polymerase. An initial denaturing step of 1 min at 94°C was followed by 30, 35 or 40 cycles of 94°C for 30 sec, appropriate annealing temperature for 30 sec and 72°C for 30 sec. A final extension step was carried out at 72°C for 7 min. The nucleotide sequences of each primer, PCR conditions and genotyping methods for each SNP are shown in Table 1 - [additional file 1](#S1){ref-type="supplementary-material"}. We genotyped SNP3 by pyrosequencing analysis on a PSQ™96MA Pyrosequencer according to the manufacturer\'s specifications with a biotinylated reverse primer (5\'-CGCCTACTCCTGGTGACTTC-3\'), and the sequencing primer (5\'-CGCCCCCATGTGT-3\'). The other 10 SNPs were genotyped by direct sequencing, as previously described \[[@B7]\]. The raw data of direct sequencing were compiled on PolyPhred \[[@B22]\]. Statistical analyses -------------------- To control genotyping errors, Hardy-Weinberg equilibrium (HWE) was checked in the control samples by the χ^2^-test (d.f. = 1). We evaluated the statistical differences in genotype and allele frequencies between cases and controls by the χ^2^-test (d.f. = 2) and the Fisher\'s exact probability test (d.f. = 1), respectively. The magnitude of LD was evaluated in *D*\' and *r*^2^using the haplotype frequencies estimated by the EH program, version 1.14 \[[@B23]\]. Statistical analysis of the haplotype association was carried out as previously described \[[@B24]\]. The significance level for all statistical tests was 0.05. Results ======= Genotyping and SNP association analysis --------------------------------------- We selected 11 SNPs at average interval of 15.9 kb to cover the entire *SLC1A2*region with LD as described in Materials and Methods. Table 2 - [additional file 2](#S2){ref-type="supplementary-material"}. shows the results of genotype and allele frequencies of SNPs in case and control samples. No significant deviation from HWE in control samples was observed (data not shown). SNP2 showed significant differences in genotype (*P*= 0.013) and allele (*P*= 0.008) frequencies between cases and controls. After Bonferroni corrections, these two *P*values became non-significance levels (*P*~corr~= 0.143, *P*~corr~= 0.088, respectively). Pairwise linkage disequilibrium and haplotype association analyses ------------------------------------------------------------------ We compared the magnitude of LD for all possible pairs of the 11 SNPs in controls by calculating *D*\' and *r*^2^(Table 3 - [additional file 3](#S3){ref-type="supplementary-material"}. , upper diagonal), because LD around common alleles can be measured with a modest sample size of 40--50 individuals to a precision equal to 10--20% of the asymptotic limit \[[@B19]\]. We observed relatively strong LD (*D*\' \> 0.8) in the seven combinations: SNP4-SNP5 (*D*\' = 0.800), SNP7-SNP8 (*D*\' = 0.877), SNP8-SNP9 (*D*\' = 0.925), SNP4-SNP11 (*D*\' = 0.838), SNP5-SNP11(*D*\' = 0.999), SNP7-SNP11 (*D*\' = 0.816), SNP9-SNP11 (*D*\' = 0.819). Modest LD (*D*\' \> 0.4) was observed in the combinations of adjacent SNPs except for SNP5-SNP6 (*D*\' = 0.286) in the control samples. However, modest LD was detected in cases in the SNP5-SNP6 combination (*D*\' = 0.497). We constructed pairwise haplotypes for all of the 55 possible SNP pairs (Table 3 - [additional file 3](#S3){ref-type="supplementary-material"}. , lower diagonal). We observed significant associations with schizophrenia in eight combinations: SNP2-SNP3 (*P*= 0.0021), SNP2-SNP4 (*P*= 0.0274), SNP2-SNP5 (*P*= 0.0054), SNP2-SNP6 (*P*= 0.0178), SNP2-SNP7 (*P*= 9.4 × 10^-5^), SNP2-SNP9 (*P*= 0.0354), SNP2-SNP10 (*P*= 0.0089) and SNP2-SNP11 (*P*= 0.0216). The combination of SNP2-SNP7 was the only one remained significant after Bonferroni correction (*P*~corr~= 0.0052). Cumulative analysis using the second sample set ----------------------------------------------- In this study, we detected significant associations of one haplotype in the *SLC1A2*region with schizophrenia in the Kyushu samples. To confirm the positive finding, we investigated the second Japanese sample set recruited from the Aichi area. Although significant association of the disease was observed with neither genotype, allele frequencies of SNP2 (*P*= 0.195, *P*= 0.178, respectively), nor haplotypes of SNP2-SNP7 (*P*= 0.084) in the second sample set, the significant haplotype association of SNP2-SNP7 was replicated in the cumulative analysis including the two sample sets (*P*= 5.0 × 10^-4^) (Table 4 - [additional file 4](#S4){ref-type="supplementary-material"}. ). Discussion ========== *SLC1A2*is located on the chromosomal region of 11p13-p12, to which no evidence has been reported for linkage of schizophrenia, \[[@B25],[@B26]\]. However, there is still a possibility that *SLC1A2*is a candidate for schizophrenia susceptibility genes, because linkage studies could only detect genes with the large genotype relative risk \[[@B27]\]. We carried out the genotyping of 100 cases and 100 controls for 11 SNPs, which were selected to cover the entire *SLC1A2*region with LD. Since minor allele frequencies of each SNP we tested ranges from 0.220 to 0.485, the expected detection power of our case-control study is from 0.89 to 0.94 for the susceptibility gene assuming 2 for genotype relative risk \[[@B28]\]. Modest LD (*D*\' = 0.925 \~ 0.409) was observed in the combinations of neighboring SNPs except for SNP5-SNP6 (*D*\' = 0.286) in the control samples, suggesting that there may be a recombination hot spot present in the small region (7.8 kb) between the two SNPs (Table 3 - [additional file 3](#S3){ref-type="supplementary-material"}. ). We plotted the magnitude of LD with the physical distance for each pair of the SNPs, and estimated the average half-length of LD to be 31.8 kb by assuming a linear regression (Fig. [2](#F2){ref-type="fig"}). This is approximately half of the previously estimated size 60 kb in a United States population of north-European descent \[[@B21]\]. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **A plot of pairwise linkage disequilibrium (LD) vs. physical distance between the SNPs in the *SLC1A2*region.***D*\' were plotted with filled diamonds, and *r*^2^with open diamonds. From the regression line, the half-length of LD was estimated to be 31.8 kb in the *SLC1A2*region. :::  ::: Significant associations of schizophrenia with genotype (*P*= 0.013) and allele (*P*= 0.008) frequencies of SNP2 (rs4755404) were detected (Table 2 - [additional file 2](#S2){ref-type="supplementary-material"}. ). However, none of these \"single-marker\" associations survived after Bonferroni corrections. An A-G transition in codon 206, causing a substitution of serine for asparagine, was identified in the exon 5 of *SLC1A2*in a heterozygous sporadic ALS patient \[[@B29]\]. Since located in a putative glycosylation site, the nonsynonymous SNP is potentially involved in the pathophysiology of schizophrenia through affecting the glycosylation status and the transport activity of *SLC1A2*\[[@B30]\]. No occurrence of the G allele of the SNP in 124 Italian schizophrenic and 50 control subjects has been reported \[[@B30]\]. We found also only A allele of the SNP in the 100 controls and 100 cases of the Kyushu samples (data not shown). In pairwise haplotype association analyses, SNP2 consistently showed significant haplotype associations. The *P*value of the combination SNP2-SNP7 was still significant even after Bonferroni correction (*P*= 9.4 × 10^-5^, *P*~corr~= 0.0052). In our second sample set, the Aichi sample, no significant association of SNP2 was observed in any of the analyses of genotypes, alleles and haplotypes. Cumulative analyses of the two sample sets, however, provide the replication of the significant haplotype association of SNP2-SNP7 with schizophrenia (*P*= 5.0 × 10^-4^). The frequency of the G-C haplotype in schizophrenics (26.6%) was notably higher than in controls (5.6%), suggesting that the G-C haplotype may be a risk haplotype for schizophrenia. We observed that the G-C haplotype frequency of schizophrenics (20.0%) was only slightly higher than controls (14.2%) in the Aichi sample, suggesting a less contribution of this locus on schizophrenia pathogenesis in the Aichi sample, although no apparent difference in clinical subtypes between both sample sets studied in this paper. The positive association reported here needs to be validated in larger sample sets, and it would also be worthwhile to search for functional SNPs in the region spanning SNP2-SNP7. Conclusion ========== We concluded that at least one susceptibility locus for schizophrenia is probably located within or nearby *SLC1A2*in the Japanese population. Competing interests =================== None declared. List of abbreviations used ========================== SNP; single nucleotide polymorphism DSM-IV; dianostic and statistical manual of mental disorders, 4^th^edn PCR; polymerase chain reaction HWE; Hardy-Weinberg equilibrium LD; linkage disequilibrium EAAT; excitatory amino acid transporter Authors\' contributions ======================= XD carried out genotyping, statistical analyses and drafted the manuscript: HS participated in design of this study and statistical analyses: HN, NT, NI and NO participated in collecting specimens and clinical data: YF conceived of the study and participated in its design and coordination. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-244X/4/21/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 PCR primers for genotyping of SNPs in SLC1A2. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 Genotype and allele frequencies of SNPs in SLC1A2 in Kyushu samples. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 Pairwise linkage disequilibrium and haplotype association in SLC1A2. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 4 Association analysis of the SNP2-SNP7 haplotype using two sample sets. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We are grateful to all the medical staff involved in collecting specimens. This work was supported in part by a Grant-in Aid for Scientific Research on Priority Areas \"Medical Genome Science\" and other grants from the Ministry of Education, Culture, Sports, Science and Technology and the Ministry of Health, Labor and Welfare of Japan.

PubMed Central

2024-06-05T03:55:47.217903

2004-8-6

PMC514709

Background ========== Influenza has the potential to cause serious complications (chiefly viral and bacterial pneumonia, with attendant mortality and morbidity), especially in the older adult (\>= 65) population and in those with comorbid illness \[[@B1]-[@B3]\]. Influenza vaccination is demonstratively safe, effective \[[@B4]-[@B6]\], and cost-effective \[[@B7],[@B8]\], and current guidelines in both Canada and the United States recommend that the vaccine be administered yearly to all individuals over the age of 65 and to all residents of Long Term Care Facilities \[[@B9],[@B10]\]. Despite current evidence and recommendations, influenza vaccination rates remain suboptimal. Estimates of vaccine coverage in high-risk groups (including older adults) range from 10--40% in the UK \[[@B11]\], and 45--68% in the United States \[[@B8],[@B12],[@B13]\]. A 1993 Canadian study found that 57.5% of community-dwelling Albertans over the age of 65 had been vaccinated within the past 12 months \[[@B14]\]. A question about influenza vaccination included in the 1991 Statistics Canada General Social Survey found that 44.8% of randomly sampled community dwelling individuals aged 65 and older had been vaccinated during the 1990/91 flu season \[[@B15]\]. A Canadian study of vaccination rates among residents of long term care (LTC) facilities reported 79% coverage for the 1990/91 season, increasing to 83% in 1998/99 \[[@B16]\]. To our knowledge, no previous studies have examined predictors of vaccination in community-dwelling older Canadians. The objectives of our study were to determine the vaccination rate in this community-dwelling population of older Canadians and to identify factors predictive of influenza vaccination, in order to determine which segments of the older adult population should be targeted to achieve better coverage in influenza vaccination programs. Methods ======= Sample/study population ----------------------- The Canadian Study of Health and Aging (CSHA) is a population-based representative sample drawn from Canadians over the age of 65, designed to study the prevalence, incidence, and risk factors for development of dementia \[[@B17]\]. A technical report providing details on sampling, design and measurement is available elsewhere \[[@B18]\]. Briefly, in 1991 and 1992, data were collected from 10,263 older adults: 9008 community-dwelling and 1255 in Long Term Care Facilities (LTCF). Those in the community were randomly selected from medicare lists (or the Enumeration Composite Record in Ontario), and institutionalized individuals were randomly selected from stratified random samples of institutions in each region. Individuals excluded were residents of the Yukon and Northwest Territories, those living on Aboriginal reserves or military bases, or those with life-threatening illnesses (examples cited include terminal cancer or conditions requiring life support). A Self-Administered Risk Factor Questionnaire (SARFQ) was completed by the non-demented community-dwelling participants in the study. The SARFQ included demographic questions and addressed issues of lifestyle, medical and family history, and medication use. From these questions a fitness/frailty scale was derived, that recognizes seven levels, from most fit (= 1) to most frail (= 7) \[[@B19]\]. A question on immunization history was asked, in which respondents were asked whether they had received the influenza vaccination, and if yes, approximately how many times they had had received it and the year of their most recent immunization. The sample population used in this study includes all participants who completed the SARFQ for whom influenza vaccination status could be determined based upon their answers. The derivation of the study population is shown in Figure [1](#F1){ref-type="fig"}. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Derivation of the study population. :::  ::: Statistical analysis -------------------- Respondents were designated vaccinated if they reported having had an influenza vaccine within the 2 years preceding completion of the SARFQ. As the data were collected over a two-year period, whereas the question about the last vaccination asked for a date, we designated as unvaccinated those who did not report receiving a vaccination within the previous 2 years. Vaccinated and unvaccinated respondents were compared; as a first step crude analyses were done taking each variable individually (not adjusted for any others). Chi-squared test or Fisher\'s exact test was used for categorical variables and one-way ANOVA was used for continuous variables. Characteristics analyzed included age, gender, education, region of residence, marital status, smoking status, alcohol intake, self-assessed health status, exercise, diagnosis of dementia, Modified Mini-Mental State Examination (3MS)\[[@B20]\] score, number of comorbidities, and urban *vs*. rural residence. As a second step, a multivariate analysis done by stepwise selection of parameters found to be significant in the univariate analysis was then conducted. Analyses comparing responders *vs*. non-responders to the relevant SARFQ questions as well as comparing respondents who reported first-time *vs*. regular influenza vaccination were also undertaken. The possibility of interaction between variables was considered. We included in the multivariable model interaction terms of pairs of variables where interaction was plausible (between smoking and drinking, age and frailty according to the frail scale, and region of residence and urban/rural dwelling). Results ======= Influenza vaccination status could be determined for 5007 (76.8%) of the 6521 non-demented, community-dwelling participants in the CSHA who completed the SARFQ. Of these, 2763 (55.2%) reported having received at least one influenza vaccination within the previous two years. The univariate analysis, in which each variable was examined individually without adjusting for the effects of others, showed several differences between the demographics, region of residence, lifestyle, and health status of the two groups (Table [1](#T1){ref-type="table"}). Immunized older adults were on average more highly educated, and being married (living with a partner or spouse) was another significant predictor of vaccination. Urban dwellers in general, and residents of Ontario were significantly more likely to have been immunized than those living elsewhere in Canada. The lowest self-report vaccination rates were found in Quebec. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Univariate analysis of vaccinated vs. unvaccinated community dwelling older Canadians ::: **Risk Factor** **Vaccinated (n = 2763)** **Unvaccinated (n = 2244)** **OR (95% C.I.)** **P value** -------------------------------- --------------------------- ----------------------------- ------------------- ------------- Age mean (SD) 75.1 (6.5) 74.6 (6.7) 1.01 (1.00--1.02) 0.0056 Sex  Male (%) 1162 (56.8) 883(43.2) 1.00 0.0527  Female (%) 1601 (54.1) 1361(45.9) 0.89 (0.80--1.00) Education mean years (SD) 11.0 (3.7) 10.3 (3.7) 1.05 (1.04--1.07) \<0.0001 Region  Atlantic (%) 471 (54.1) 399 (45.9) 1.00  Quebec (%) 453 (48.2) 486 (51.8) 0.79 (0.66--0.95) \<0.0001  Ontario (%) 568 (64.6) 311 (35.4) 1.55 (1.28--1.88)  Prairies (%) 540 (55.6) 431 (44.4) 1.06 (0.88--1.28)  BC (%) 731 (54.2) 617 (45.8) 1.00 (0.85--1.19) Current Marital Status  Not married (%) 1197 (52.6) 1078 (47.4) 1.00 0.0007  Married (%) 1566 (57.4) 1163 (42.6) 1.21 (1.08--1.36) Smoker  No (%) 1232 (52.9) 1095 (47.1) 1.00 0.0032  Yes (%) 1495 (57.1) 1122 (42.9) 1.18 (1.06--1.32) Alcohol  No (%) 1590 (53.2) 1399 (46.8) 1.00 0.0012  Yes (%) 1133 (57.9) 824 (42.1) 1.21 (1.08--1.36) Self-Rated Health  Not Good/Very Poor (%) 502 (60.6) 326 (39.4) 1.00 0.0006  Very/Pretty Good (%) 2254 (54.1) 1912 (45.9) 0.77 (0.66--0.89) Regular Exercise  No (%) 893 (52.0) 824 (48.0) 1.00 0.0014  Yes(%) 1815 (56.8) 1382 (43.2) 1.22 (1.08--1.36) 3MSE mean score (sd) 89.9 (6.4) 89.1 (6.4) 1.02 (1.01--1.03) \<0.0001 No. of comorbidities mean (sd) 2.7 (1.7) 2.3 (1.6) 1.15 (1.11--1.20) \<0.0001 Geography  Urban (%) 2442 (55.8) 1934 (44.2) 1.00 0.0274  Rural (%) 320 (51.1) 306 (48.9) 0.83 (0.70--0.98) ::: Of note, smokers were significantly more likely than non-smokers to have received the influenza vaccination. The same was true of individuals who consumed alcohol regularly when compared with those who drank rarely or not at all. Regular exercise was found to be a predictive factor for vaccination. Those who had received the vaccination had significantly more health problems, as did those who saw themselves as being in poorer health. In the multivariate analysis (Table [2](#T2){ref-type="table"}) predictive factors for immunization that retained significance after adjusting for the effects of other variables include older age, higher level of education, being married, smoking, engaging in regular exercise, and having more co-morbid illnesses. Interaction terms describing interaction between age and frailty, smoking and drinking, and region of residence and urban/rural dwelling did not retain significance in the multivariable model, suggesting that there was no statistically significant interaction between these pairs of variables. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Multivariate analysis comparing vaccinated and unvaccinated community-dwelling older Canadians ::: **Risk Factor** **Odds Ratio** **95% C.I.** **p value** ------------------- ---------------- -------------- ------------- Older Age 1.02 1.01--1.06 \<0.0001 More Education 1.05 1.03--1.07 \<0.0001 Currently Married 1.29 1.14--1.47 \<0.0001 Smoking 1.14 1.01--1.30 0.0401 Drinking 1.51 1.01--1.31 0.0358 Regular Exercise 1.25 1.10--1.42 0.0004 Region  Atlantic 1.00 \-\--  Quebec 0.94 0.77--1.14 \<0.0001  Ontario 1.48 1.21--1.82  Prairies 0.99 0.82--1.20  BC 0.80 0.67--0.97 More Comorbidity 1.18 1.140--1.226 \<0.0001 ::: To investigate the impact of response bias, we compared those who had answered the influenza vaccination question and the 23.2% of the original sample who had not (Table [3](#T3){ref-type="table"}). Non-responders were more likely to be rural residents and to drink alcohol regularly. Age, 3MS score, self-rated health, and number of comorbidities also differed between the groups: non-responders were more likely to be younger, healthier, and to have better self-rated health and higher 3MS scores. Several of the same characteristics (age, number of comorbidities, self-rated health, region of residence and urban vs. rural living) also differed between regular and first time vaccination users (Table [4](#T4){ref-type="table"}). Repeat users of the influenza vaccine (i.e. those who reported at least two past vaccinations) were slightly older (75.3 *vs*. 74.1 years, p \< 0.0001), had more comorbid illness (2.7 *vs*. 2.5, p = 0.0011), and poorer self-rated health (81.7% poor health *vs*. 77.9% good health, p = 0.045). Residents of Atlantic Canada (75.1% *vs*. 79.3% residing elsewhere, p = 0.0003) and those living in rural areas (79.3% of urban *vs*. 73.8% of rural-dwellers, p = 0.12) were less likely to be regular users. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Analysis of responders vs. non-responders to questions relating to influenza vaccination in the SARFQ ::: **Characteristic** **Value Responders (n = 5007)** **Value Non-responders (n = 1514)** **p value** -------------------------------- --------------------------------- ------------------------------------- ------------- Age mean (SD) 74.9 73.4 \<0.0001 Sex  Male (%) 76.9 23.1 0.8416  Female (%) 76.7 23.3 Education mean (SD) 10.7 10.7 0.8164 Region  Atlantic (%) 71.6 28.4  Quebec (%) 78.9 21.1 \<0.0001  Ontario (%) 69.4 30.6  Prairies (%) 70.1 29.9  BC (%) 92.1 7.9 Current Marital Status  Not married (%) 77.1 22.9 0.5742  Married (%) 76.5 23.5 Smoker  No (%) 78.0 22.0 0.1888  Yes (%) 76.6 23.4 Alcohol  No (%) 82.4 17.6 \<0.0001  Yes (%) 75.8 24.2 Self-Rated Health  Not Good/Very Poor (%) 82.4 17.6 \<0.0001  Very/Pretty Good (%) 75.8 24.2 Regular Exercise  No (%) 77.8 22.2 0.6263  Yes(%) 77.3 22.7 MMMSE mean (sd) 89.6 90.1 0.0071 No. of comorbidities mean (sd) 2.5 2.1 \<0.0001 Geography  Urban (%) 77.7 22.3 \<0.0001  Rural (%) 70.7 29.3 SARFQ = Self-Administered Risk Factor Questionnaire in the Canadian Study of Health and Aging ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Characteristics of regular vs. first time influenza vaccine users ::: **Risk Factor** **Regularly Vaccinated** **First Time Vaccinated** **P value** -------------------------------- -------------------------- --------------------------- ------------- Age mean years (SD) 75.3 (6.7) 74.1 (6.7) \<0.0001 Sex  Male (%) 79.5 20.5 0.29  Female (%) 78.9 22.0 Education mean years (SD) 11.0 (3.8) 10.8 (3.8) 0.27 Region  Atlantic (%) 75.1 24.9 0.0003  Quebec (%) 83.7 16.3  Ontario (%) 81.5 18.5  Prairies (%) 78.2 21.8  BC (%) 75.5 24.5 Current Marital Status  Not married (%) 79.2 20.8 0.45  Married (%) 78.1 21.9 Smoker  No (%) 78.3 21.7 0.73  Yes (%) 78.8 21.2 Alcohol  No (%) 78.4 21.6 0.83  Yes (%) 78.7 21.3 Self-Rated Health  Not Good/Very Poor (%) 81.7 18.3 0.045  Very/Pretty Good (%) 77.9 22.1 Regular Exercise  No (%) 80.2 19.8 0.086  Yes(%) 77.5 22.5 3MSE mean score (sd) 89.9 (6.4) 89.7 (6.4) 0.48 No. of comorbidities mean (sd) 2.7 (1.7) 2.5 (1.7) 0.0011 Geography  Urban (%) 79.3 20.7 0.012  Rural (%) 73.8 26.2 ::: Discussion ========== Our data were collected in the early 1990s and may therefore not reflect current practice patterns. Although the Canadian Study of Health and Aging was not designed to study influenza vaccination, it does provide a large and unique data set of primarily community-dwelling older Canadians, and is therefore potentially useful in the examination of health-related risk factors and demographics that may influence decisions to vaccinate. Given the importance of influenza vaccination in the prevention of significant morbidity and mortality in populations at risk, the vaccination rate of 55.2% in our community-dwelling sample of older adults is concerning. People who were not vaccinated tended to be younger, non-smokers and to have fewer co-morbid illnesses. They were also found to have a lower level of education, not to be married and not to engage in regular exercise. These were the factors that retained statistical significance in the multivariable analysis, suggesting that these associations are unlikely to have arisen due to confounding by any of the variables investigated. Vaccination has previously been studied in the CSHA, but only in relation to its status as a potentially protective factor with respect to cognitive impairment \[[@B21]\]. The weight of evidence derived from re-examination of large databases is less than that derived from specifically designed trials, however this method still maintains an important role in epidemiological research. For example, evaluation of systematic problems can be used to help develop targeted efforts in improving vaccination rates. Our data do not include information on institutionalized older adults, where it might be expected that at-risk profiles would vary, and where vaccination rates are generally higher \[[@B16]\]. Another important source of potential error is our reliance on self-reported immunization status. However, self-report of influenza vaccination status in elderly outpatients has been found to be highly sensitive and moderately specific when checked against medical record documentation \[[@B22]\]. Other studies have identified predictive factors for vaccination in other countries \[[@B8],[@B12]\]. An American study found that patients with more health conditions, higher rates of use of health care resources, and history of pneumonia were more likely to be vaccinated, while non-vaccinated individuals were older and more likely to have dementia or stroke \[[@B8]\]. An Iowa study identified a number of factors associated with the receipt of both influenza and pneumococcal vaccines: age over 70, self-owned residence, working, increased number of medical conditions, current prescription medication, and a physician visit within the past year. Geography (rural vs. urban living) was unrelated to vaccine receipt \[[@B12]\]. Our data suggest that there appears to be an important degree of targeting of vaccination resources within the older adult population to people who are less healthy. Even within these higher risk groups, however, immunization is incomplete. Perhaps such people are perceived by their health care providers as being at higher risk from influenza infection, and are thus more likely to be immunized. Similar explanations may account for the higher vaccination rates among smokers and those who consumed more alcohol. However, according to the current Canadian guidelines \[[@B9]\], influenza vaccination is indicated for all persons 65 years of age and over. If it is the case that those at the younger end of this age group are less likely to be vaccinated, as our study suggests, more must be done to ensure that vaccination is reaching its entire target population. Regular exercise was shown to be a factor predictive of influenza vaccination. Interestingly, this association may have been confounded by generally health-protective behaviour (which might be expected to be associated with both regular exercise, the explanatory variable, and vaccination, the outcome), given that individuals exhibiting healthy lifestyle choices (such as exercise) may have been more likely to have sought preventative health care and to have visited their health care providers more regularly, thus providing more opportunity to be vaccinated. Regular exercise did, however, retain significance in our multivariate analysis. The regional differences identified in our study may point to geographic differences in access to influenza vaccination, although the general milieu and level of awareness in both the medical community and society at large may also be significant. These regional differences suggest that certain areas may benefit from targeting of vaccination efforts. However, over and above questions of health policy and public health education, the question of access to vaccination is of vital importance if we are to achieve acceptable rates of coverage in this target population. The finding that rural residence is a negative predictor of vaccination is particularly concerning, and points to the larger equity issue of how uptake of influenza vaccination can be improved outside of major urban centers. However, our finding that rural *vs.*urban residence did not retain statistical significance in the multivariable regression model suggests that this crude association may be due to confounding by other factors. Our analysis comparing non-responders with those for whom influenza vaccination status was known on the basis of the SARFQ showed that response bias was indeed likely. Non-responders were more likely to drink more alcohol, a factor which was associated with being immunized. However, several factors which were associated with not being immunized, including younger age, good self-rated health, having fewer comorbidities, and residence in a rural area, were more prevalent among non-responders. This over-representation of a number of factors predictive of non-immunization among non-responders suggests that our results may well have been influenced by response bias. For example, we may have over-estimated the prevalence of vaccination in this population. The analysis comparing regular users of the influenza vaccine with those who reported first-time immunization in the SARFQ demonstrated a number of factors that differed between the two groups. The finding that younger age is associated with first time vaccination may be at least partially explained by those at the younger end of the \> = 65 age group having spent less time in this \"target group\" for vaccination. As such, they would be more likely to be receiving the vaccination for the first time. The effect of age may also play a role in the finding that those with better self-assessed health and fewer comorbidities were more likely to be first-time users (as they may also have been younger). However, being in better health was also a predictive factor for non-vaccination within the previous two years, as was rural residence, suggesting that these factors may influence decisions to initiate as well as to sustain influenza vaccination over subsequent years. Conclusions =========== Despite current recommendations and proven benefit, influenza vaccination rates in our Canadian sample were suboptimal; only 55.2% of older adults in our representative sample reported being vaccinated within the past two years. The predictive factors identified in this analysis may facilitate the study of targeting of older adults in future vaccination programs. Competing interests =================== SM and HM have been involved in trials of Influenza vaccines funded by ID Biomedical. Neither has received any personal financial compensation for their involvement. MKA and KR have no competing interests. Author contributions ==================== MKA wrote the first draft, and reviewed each of the analyses. SM & KR reviewed the analyses and revised the manuscript. KR is a co-Principal Investigator of the CSHA. HM did the statistical analyses. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/4/36/prepub> Acknowledgements ================ This analysis was supported by a grant from the Canadian Institutes for Health Research (CIHR) MOP 62823. The data reported in this article were collected as part of the Canadian Study of Health and Aging. The core study was funded by the Seniors\' Independence Research Program, through the National Health Research and Development Program (project no. 6606-3954-MC(S)). Additional funding was provided by Pfizer Canada Incorporated through the Medical Research Council/Pharmaceutical Manufacturers Association of Canada Health Activity Program, the National Health Research and Development Program (project no. 6603-1417-302(R)). The study was coordinated through the University of Ottawa and the Division of Aging and Seniors, Health Canada. KR receives support from the CIHR through an Investigator awards, and from the Dalhousie Medical Research Foundation as Kathryn Allen Weldon Professor of Alzheimer Research. MKA receives support from the Commonwealth Scholarship Commission in the United Kingdom.

PubMed Central

2024-06-05T03:55:47.219677

2004-8-11

PMC514710

Background ========== Gene therapy protocols would strongly benefit from the development of a one step technique that would allow cells to be genetically modified through the introduction of several therapeutic genes. In order to induce the translation and expression of exogenous cDNAs, carried by a single vector, researchers have cloned internal ribosomal entry sites (IRES) upstream from these exogenous cDNAs. This approach should lead to the translation of three cistrons from an unique mRNA and therefore to the consequent expression of the three encoded proteins \[[@B1]-[@B6]\]. In most cases, the IRES from EMCV is cloned into polycistronic vectors as it induces high levels of DNA translation \[[@B3],[@B7],[@B8]\]. The capacity of other IRESes to induce high levels of exogenous cDNA expression in different cell types has been compared to the capacity of the EMCV IRES \[[@B2]-[@B4],[@B9]-[@B11]\]. However, in most cases, these comparisons were carried out after different IRESes had been inserted into a single, characterized, dicistronic (one IRES) or tricistronic (two IRESes) mRNA and after the *in vitro*vector translation efficiency had been established \[[@B9]-[@B14]\]. These studies are useful in choosing the IRES that will drive the *in vivo*expression of heterologous proteins, they do, however, give little information as to the potential *in vivo*interactions that occur between different IRESs cloned into the same MuLV-based retroviral vector. We cloned tricistronic vectors encoding three different cDNAs. This involved using at least two IRESes to translate the second and third cistrons. Using the same IRES twice in a single vector could, however, induce recombination events and the loss of the second IRES and cistron. In the same way, using the same cistron twice could lead to a competition between the two IRESs for the binding to cell type specific translation factors. For these reasons, we chose to clone and compare the efficiency of different IRESes cloned into the same vector. We chose the IRES of EMCV (IRES^EMCV^), the IRES of the *c-myc*proto-oncogene (IRES^C-MYC^), the IRES of FGF-2 (IRES^FGF-2^) and the IRES of the HTLV-1 lentivirus (IRES^HTLV-1^) \[[@B1],[@B8],[@B15]-[@B22]\]. The vectors were constructed so that the third cistron was translated from the IRES^EMCV^and the second cistron was translated from the IRES^EMCV^, IRES^C-MYC,^, IRES^FGF-2^or IRES^HTLV-1^. Recently, it has been shown that retroviral vectors derived from MuLV contain an additional IRES on the 5\' gag sequence \[[@B20],[@B23]\]. The vectors described here contained three IRESes: the IRES from MuLV located between the LTR and the Psi sequence controlling the translation of the first cistron, the IRES from a different origin and the IRES from EMCV respectively controlling the translation of the second and third cistrons (Figure [1A](#F1){ref-type="fig"} and [1B](#F1){ref-type="fig"}). The exogenous genes cloned into the tricistronic vectors were chosen for their potential use in clinical trials. They code for co-stimulatory molecules known to enhance tumor cell immunogenicity: CD80, a member of the B7 family and CD70, a member of the TNF family \[[@B7],[@B24]-[@B26]\]. These molecules act in synergy to enhance the induction of Ag-mediated anti-tumor immunity when co-expressed with tumor antigens \[[@B7],[@B24],[@B25],[@B27]\]. We generated retroviral vectors that encoded the two co-stimulatory molecules CD70 and CD80, and a selection gene. We compared the efficacy of these vectors in their capacity to genetically modify various human and murine cells, and also observed how they affected the selection and culture of these cells following transduction. We then compared the expression of the three exogenous genes within the genetically modified cells. Murine melanoma cells were then tested in two different murine tumor models for their ability to induce an *in vivo*anti-tumor immune response, regardless of the percentage of co-stimulatory molecules expressed by the transduced cells. Results ======= Construction of tricistronic retroviral vectors expressing CD70 and CD80 ------------------------------------------------------------------------ We constructed tricistronic vectors that would induce the expression of three cDNAs (CD70, CD80 and a selection gene) from one promoter (LTR viral promoter) (Figure [1](#F1){ref-type="fig"}). The constructions are described in the *Methods*. Expression of the first open reading frame (cDNA encoding a co-stimulatory molecule) occurs from one IRES (MuLV) located at the 5\' extremity of the mRNA. The second and third open reading frames are translated from two identical or two different IRESes. TFG^EMCV^NEO or TFG^EMCV^ZEO were constructed so that the translation of the selection gene and the second co-stimulatory molecule could be induced from two identical IRES^EMCV^(Figure [1A](#F1){ref-type="fig"}). TFG^HTLV-1^, TFG^C-MYC^and TFG^FGF-2^were constructed to allow the expression of the selection gene from IRES^HTLV-1^, IRES^C-MYC^or IRES^FGF-2^while the translation of the third open reading frame was under the control of IRES^EMCV^(Figure [1B](#F1){ref-type="fig"}). Efficiency of the different IRESes in inducing the expression of the selection gene (NEO or ZEO) in different cell types ------------------------------------------------------------------------------------------------------------------------ We generated retroviral vectors using the different plasmids described in Figure [1](#F1){ref-type="fig"} and transfected ψCRIP cells or triple-transfected the 293T packaging cell line. We obtained viable G418 or zeocin-resistant ψCRIP cells after transfection with TFG^EMCV^NEO, TFG^EMCV^ZEO, TFG^HTLV-1^ZEO, TFG^HTLV-1^NEO, TFG^FGF-2^ZEO or TFG^FGF-2^NEO but none after transfection with TFG^cMYC^ZEO or TFG^cMYC^NEO. We tested the supernatants from ψCRIP or 293T transfected cell suspensions for the presence of replication competent retroviruses (RCR) through the dosage of reverse transcriptase activity. The supernatants from all the transfected packaging cells were free of reverse transcriptase activity. The titers of the different retroviruses produced different results, depending on the experiments, varying from 10^4^to 10^6^particles/ml. We used 10^5^particles of each type of retrovirus that was produced (TFG^FGF-2^ZEO, TFG^HTLV-1^ZEO or TFG^EMCV^ZEO) or the 48 hour supernatants from transiently transfected TFG^cMYC^ZEO ψCRIP or 293T cells to transduce different types of mammalian cells. The murine cells used were either NIH-3T3 fibroblasts or B16.F10 melanoma cells. To establish a human model, we cultured melanoma cells from biopsies as described in the *Methods*. During the course of this study we obtained 23 melanoma biopsies. Fifteen cell cultures were obtained from these biopsies which represents a yield of 65 %. The quality and characteristics of the melanoma cells were determined by immunohistochemistry, on cytospin cells, using anti-cytokeratin (KL-1) (negative control), anti-S100 protein, anti-melanA/MART1 and anti-HMB-45/gp100 antibody staining as described in the *Methods*. The 15 human cell cultures obtained all stemmed from melanoma cells (data not shown). The transduced cells were selected for their resistance to G418 or zeocin. Fifteen days after the transduction with each retroviral vector, we selected G418 or zeocin-resistant cells. Apart from the TFG^CMYC^ZEO transduced cells, we obtained roughly the same number of zeocin or G418 resistant clones with the different retroviral constructs. Both murine and human cells could be successfully genetically modified using the engineered trigenic retroviral vectors. No vector was statistically more efficient in obtaining a higher yield of resistant clones. However, the TFG^EMCV^ZEO, TFG^FGF-2^ZEO and TFG^HTLV-1^ZEO cells were long lasting, viable and could be expanded, whereas neither murine nor human TFG^cMYC^ZEO transfected cells displayed long-term viability. The selected genetically modified cells expressed co-stimulatory molecule mRNAs ------------------------------------------------------------------------------- Cells were stably transduced with MFG derived vectors encoding CD70, CD80 and an antibiotic resistance gene. After selection with the appropriate antibiotics (G418 or Zeocin), the stably transduced cells were analyzed for the expression of the different mRNAs, by RT-PCR using specific primers. The aim of the RT-PCR analysis was to show, through the expression of the full length RNA, that successful transcription of the construct had been achieved. We were uncertain whether the transcription of the ectopic DNA was complete. RT-PCR analysis was performed using cells that were resistant to Zeocin. We are convinced that the mRNA transcribed from the ectopic DNA contains the IRES zeocin cassettes. As we can noticed in figure [2A](#F2){ref-type="fig"}, the primers used for PCR analysis n°4 and 5 overlapped. They hybridized with the sequence corresponding to the ZEO gene. Our hypothesis is that if we can amplify the two segments of the construction (CD70-IRES-ZEO and ZEO-IRES-CD80 or CD80-IRES-ZEO and ZEO-IRES-CD70) at the same time from the ectopic RNA, then the full length RNA had been transcribed. We have already performed other RT-PCR analyses using this the long ranger taq polymerase from Applied (France). We have been able to amplify full length RNA from all constructs (data not shown). Taken together our data strongly suggests that the full length RNA was transcribed from all the viral constructs. These RT-PCR results suggest that there was no downregulation of CD70 or CD80 expression at the transcriptional and post-transcriptional levels. The selected genetically modified cells expressed the co-stimulatory molecules at different levels -------------------------------------------------------------------------------------------------- The cell populations were tested for co-stimulatory molecule expression by flow cytometry using Abs that were directed against CD70 and CD80. Within a given population, a large percentage of cells expressed the two co-stimulatory molecules in a stable manner and at high levels. Figure [3F](#F3){ref-type="fig"} shows the results of the flow cytometric analysis of human melanoma cells transduced with TFG^FGF-2^ZEO. Figure [3](#F3){ref-type="fig"} is representative of the cytometric analyses carried out on the murine B16.F10 melanoma cells transduced with 48 hr supernatants of 293T cells transfected with the 4 different types of retroviral vectors. A high percentage of B16.F10 cells transduced with TFG^FGF2^ZEO or TFG^HTLV-1^ZEO retroviral vectors (Fig. [3D](#F3){ref-type="fig"} and [3E](#F3){ref-type="fig"} respectively) expressed both co-stimulatory molecules. The percentage of B16.F10 cells, transduced with TFG^EMCV^ZEO or TFG^c-MYC^ZEO, expressing only CD80 was higher than the percentage of cells expressing only CD70 or both molecules (Fig. [3B](#F3){ref-type="fig"} and [3C](#F3){ref-type="fig"} respectively). The level of expression of the two molecules differed depending on the vectors used and the cell types that were transduced, as shown in Table 1 - see [additional file 1](#S1){ref-type="supplementary-material"}. In theory, selected clones expressed both molecules. In fact, in any selected clone, we found cells that expressed only one of the molecules (CD70 or CD80) and cells that expressed the two molecules. ### Murine cells In murine cells (either fibroblasts or melanoma cells), a high percentage of cells only expressed CD70. This was not due to the expression of the CD70 molecule, in itself. Indeed, when we used constructs with CD80 as the first cistron, we obtained a high percentage of cells that only expressed CD80. This percentage could be attributed to a negative regulation of the translation of the third cistron. Even after the selection of transfected cells, we obtained cells which did not express detectable levels of co-stimulatory molecules. Indeed, in TFG^HTLV-1^ZEO, TFG^FGF-2^ZEO or TFG^EMCV^ZEO transduced murine NIH-3T3 cells, we respectively found 35%, 31% or 44% of cells that expressed undetectable levels of CD70 or CD80. Expression levels were assessed regardless of the resistance gene that was used (NEO or ZEO). We tested the expression of the two co-stimulatory molecules immediately after transduction or after selection. No correlation could be established between the time of analysis and the percentage of cells expressing both molecules. One clone could be composed of a majority of cells (60%) expressing both molecules at day 5 post selection, only have a small percentage of these cells at day 15 and up to 95 % on day 25. Indeed, even within a selected population of cells, the percentage expressing both co-stimulatory molecules could vary from 5 up to 95 % of the total cell number. After transduction with a single vector, 20 clones were selected. Within the same clone, some cells expressed only one co-stimulatory molecule whereas others expressed both molecules. These observations are reflected by the high SD of the percentage of cells expressing co-stimulatory molecules (Table 1 - see [additional file 1](#S1){ref-type="supplementary-material"}). None of the vectors used were found to be adequate for the transfection of murine cells. ### Human cells For human cells, we chose to use a pool of transduced cells, rather than clones, to stay as close as possible to the reality of human clinical protocols. When we used TFG^HTLV-1^ZEO (CD80 first) or TFG^HTLV-1^ZEO2 (CD70 first) we obtained cells that expressed only the first cistron. We obtained a better yield of cells expressing the two co-stimulatory molecules when using the TFG^FGF-2^ZEO construct (70% of cells expressing the two molecules in a stable manner). Again, even within a selected population of cells, the percentage of cells expressing both co-stimulatory molecules could vary considerably (from 45 to 95 % of total cell number). In human melanoma cells, the use of tricistronic vectors in which IRES^FGF-2^induced the translation of the second cistron, and IRES^EMCV^the translation of the third cistron, led to a high percentage of cells expressing the three cistrons. Expression of the tricistronic transgene slows the tumor growth rate after s.c. injection of B16.F10 in an established model ---------------------------------------------------------------------------------------------------------------------------- Among selected cells, the percentage expressing both co-stimulatory molecules varied. We have previously shown that melanoma cells expressing high levels of CD70 alone or in combination with CD80 induced *in vitro*splenocyte proliferation \[[@B7]\]. Using a pool of selected cells that were genetically modified by TFG^HTLV-1^ZEO (52 % CD70, 21 % both CD70 and CD80), TFG^FGF-2^ZEO (5 % CD70, 33 % both CD70 and CD80), TFG^EMCV^ZEO (27 % CD70, 16 % both CD70 and CD80) or double-transfected with DFG CD70 and DFG CD80 (\>65 % both CD70 and CD80), we showed that different percentages of cells expressing both co-stimulatory molecules (even as little as 20 %) could induce a proliferative response in splenocytes. A substantial increase in spleen cell proliferation was observed when the two molecules were expressed following genetic modification with all the viral vectors. This increase was similar to that induced by the use of double transfected cells (data not shown). To determine whether the local expression of CD70 and CD80 could affect tumor establishment, we sub-cutaneously injected 10^5^cells, from each type of tumor, into the flanks of C57BL/6 mice (Figure [4A](#F4){ref-type="fig"}). These cells had previously been used in splenocyte proliferation experiments. Subcutaneous injection of B16.F10 cells into B6 syngeneic immunocompetent mice led to the development of tumors. However, there was a delay in the appearance of tumors derived from transduced or double-transfected cells compared with tumors induced by the inoculation of parental or mock-transfected cells (control B16.F10 cells transfected with a retroviral vector encoding the zeocin-resistance gene). Although, all mice had palpable tumors before day 10, the growth rate was significantly slower for TFG^FGF-2^ZEO or TFG^HTLV-1^ZEO transduced cell tumors compared to parental cell tumors (p \< 0,01) and double transfected cell tumors (p \< 0.001) on day 20 (Figure [4A](#F4){ref-type="fig"}). This decrease in tumor growth was less pronounced in tumors derived from cells modified by tricistronic vectors than in tumors derived from cells modified by two independent vectors. This could be due to the level of expression of the co-stimulatory molecules on the cell surface. We have previously shown that co-expression of the two co-stimulatory molecules is necessary to induce an *in vivo*immune response. Here we show that, for the B16.F10 melanoma cell model, we had to inject as many as 6 × 10^4^cells (60 % of the injected cells) expressing both co-stimulatory molecules to observe an immune response. We have previously shown that the anti-tumor effect is more difficult to obtain in a MHC class I loss variant than in a MHC class I positive model \[[@B7]\]. To confirm how important the percentage of cells expressing both co-stimulatory molecules is in inducing an *in vivo*immune response, we transduced TS/A adenocarcinoma cells using the same retroviral vectors, as previously described. We pooled the zeocin resistant cells and chose two pools. In the first pool of cells transduced with TFG^EMCV^ZEO, 34 % only expressed CD70 and 23 % expressed both CD70 and CD80. In the second pool of cells transduced with TFG^FGF-2^ZEO, 73 % only expressed CD70 and 20 % expressed both CD70 and CD80. These two pools were injected sub-cutaneously into syngeneic immunocompetent BALB/c mice. These injections led, in all cases, to the development of tumors. However, when the cells were double-transfected with two independent vectors and more than 70 % of cells expressed both co-stimulatory molecules, most of the palpable tumors spontaneously decreased within 10 days. In contrast, only a delay in the appearance of the tumors derived from transduced cells could be observed. This delay was more or less pronounced depending on the percentage of cells expressing both molecules. There was no delay in the appearance of tumors derived from the pool containing 34 % of cells expressing only CD70 and 23 % of cells expressing CD70 and CD80. A delay was observed in the appearance of tumors derived from the pool containing 73 % of cells expressing only CD70 and 20 % of cells expressing CD70 and CD80 (figure [4B](#F4){ref-type="fig"}). Discussion ========== In this study, we have constructed and tested different tricistronic retroviral vectors containing IRES elements from different origins. As IRESes are remarkably efficient when used in bicistronic vectors, it was particularly interesting from a biotechnological point of view (for gene therapy protocols) to design polycistronic vectors that could allow the expression of several proteins from the same mRNA. Several authors have reported the construction of different vectors using the IRES^EMCV^or the IRES^FMDV2A^to trigger the high level expression of an exogenous gene\[[@B2]-[@B4],[@B10],[@B11],[@B28]\]. Prats and coll. have described several other IRESes including the IRESs from the *c-myc*proto-oncogene, the IRES from FGF-2 and the IRES from HTLV-1 \[[@B17],[@B18],[@B20],[@B21]\]. The first aim of our work was to determine whether one of these IRESes could induce a higher level of exogenous protein expression than the IRES from EMCV in different cell lines. The second aim was to design polycistronic vectors carrying different IRESes to avoid the risk of recombination between IRESes. The third aim was to obtain genetically modified melanoma cells through transduction with tricistronic retroviral vectors. These tumor cells would therefore be genetically modified to express two co-stimulatory molecules, CD70 and CD80, which are known to induce an anti-tumor response in syngeneic mice. As we have generated retroviral vectors derived from the well known MFG vector (MuLV vector), we already had one additional IRES upstream from the first ATG codon (initiation codon of the first exogenous cDNA). This IRES has been described by Vagner and coll \[[@B20]\]. The first IRES (MulV) and the third IRES (EMCV) were conserved in all the vectors. The second IRES responsible for inducing the translation of the selectable marker, was the IRES from either EMCV, FGF-2, *c-myc*or HTLV-1. The choice of IRES that was used to express the cDNA encoding the G418 or zeocin-resistance genes was unimportant as we obtained resistant cells in all the cell lines tested: murine tumor cells (B16.F10, TS/A), murine fibroblasts (NIH3T3), packaging cell line (ψCRIP), human melanoma and lung adenocarcinoma cells (A549). However, when the IRES from the *c-myc*proto-oncogene was used, we never obtained long-lasting zeocin or G418 resistant murine or human cells, whether these were tumor cells or fibroblasts. So far, most of the cellular mRNAs that contain IRESes and code for proteins involved in the control of cell proliferation and differentiation, require stringent regulation like for example the *c-myc*mRNAs \[[@B13],[@B15],[@B29],[@B30]\]. These genes need to be expressed at very specific stages of the cell cycle and/or in response to different stimuli \[[@B17],[@B19]\]. This has also been shown for FGF-2. Indeed the CUG-initiated isoforms of FGF-2 are translationally activated in response to stress \[[@B21]\]. Such observations suggest that when the IRES^c-myc^is used, translation is strongly downregulated \[[@B30]\]. We obtained long-lasting viable resistant cells when we used the IRESes from EMCV, FGF-2 or HTLV-1. The number of clones obtained after transduction or transfection was roughly the same depending on the experiments and the cell lines tested. This indicated that only a few (if any) recombination events occurred when we used the same IRES (EMCV) twice in the same retroviral vector. The first and third IRESes are responsible for inducing the translation of the two co-stimulatory molecules. These IRESes competed to induce the expression of the two exogenous cDNAs. Indeed within the same population of selected cells, whatever the retroviral vector used, we were able to obtain cells that only expressed the first exon, or only the third exon or both exons. The percentage of cells that expressed both co-stimulatory molecules varied with the cell passage and from one selected clone to another. This is true regardless of the cell line tested: murine tumor cells (B16.F10, TS/A), murine fibroblasts (NIH3T3), murine packaging cell line (ψCRIP), human melanoma cells and human lung adenocarcinoma cells (A549). There is a difference in the IRES-dependent mechanism that occurs in cellular and viral internal initiation. Currently, two cellular *trans*-acting factors, the La antigen and PTB have been found to bind to picornavirus IRES elements and to be essential for their internal initiation of translation \[[@B29],[@B31],[@B32]\]. However, these proteins do not specifically bind to eukaryotic cellular mRNAs with the same efficiency. IRES function must require either different amounts of translation initiation factors or, more likely, additional proteins similar to those required for the cap-dependent initiation of protein synthesis \[[@B29],[@B33]\]. Borman and coll. have recently shown that the recognition of different IRES elements varies within different tissue culture cell lines \[[@B9],[@B12]\]. The activity of a particular IRES within a cell may be dependent on the relative level of stimulatory and inhibitory molecules \[[@B34]\]. It is possible that different *trans*-acting factors that are dependent on a specific IRES may be required. It could be that the MuLV IRES (first exon) binds those *trans*-acting factors with a higher affinity than the IRES from EMCV (third exon). Anthony and Merrick suggested that translation factors, that have a higher affinity for the cap structure than for the IRES element, could be sequestered at the m7GpppN cap structure and would therefore be unavailable for the internal initiation of translation at saturating concentrations of capped bicistronic RNA \[[@B1],[@B35]\]. This would lead to an increased translation of the first cistron and a decreased translation of the second and third cistrons that are thought to be dependent on internal initiation of translation. In our case, we have an IRES (IRES MulV) on the 5\' end of the mRNA. This IRES induces the translation of the first co-stimulatory molecule. However, while the IRES from EMCV or HTLV-1 could interact with other IRESes present within the retroviral construct (the IRES from gag or the IRES from EMCV), we found that the IRES from FGF-2 induced the expression of exogenous cDNA in human melanoma cells without any positive or negative regulation from the other IRESes. We have previously shown that two co-stimulatory molecules (CD70 and CD80), expressed on the surface of tumor cells could induce an anti-tumor immune response when the cells were injected into a syngeneic animal \[[@B7],[@B27]\]. In tumor cells that were genetically modified by the tricistronic retroviral vector, we attempted to induce an *in vivo*anti-tumor response in murine models. We first showed that B16.F10 cells that were genetically modified by the tricistronic vectors, could induce *in vitro*proliferation of spleen cells. These B16.F10 cells were then injected into syngeneic animals and tumor growth was monitored. We observed that in this murine model (C57BL/6) or in the BALB/c murine model (breast adenocarcinoma TS/A cells), co-expression of the two co-stimulatory molecules by at least 50 % of cells was necessary to induce an anti-tumor response. CD70 expression, alone or in association with a low level of CD80 expression, was not sufficient to induce anti-tumor immunity. These findings show that at least 50 % of the genetically modified cells must express the CD70 and CD80 co-stimulatory molecules before starting an immunotherapy protocol. Establishing cultures of human melanoma cells derived from biopsies was very difficult, we obtained a low yield of 63 %. We could infect human melanoma cells with the tricistronic retroviral vector with an efficiency of 50 %. However, within the 50 % of cells that were genetically modified we could only obtain a high percentage of cells expressing both co-stimulatory molecules when we used the tricistronic recombinant amphotropic viruses obtained with the IRES FGF-2. Conclusion ========== The ability of retroviral vectors carrying IRESes to deliver genes, *in vitro*and *in vivo,*to a variety of dividing cell types has been applied to research and gene therapy for the past 10 years. Our work shows that it is difficult to chose which IRES must be inserted into a polycistronic gene therapy vector, when the aim is to ensure a high level of translation of the exogenous gene. This level of expression will depend on the type of cell that is transduced but also on the presence of other genetic elements within the vector. Methods ======= Cell lines ---------- The murine melanoma B16.F10 cell line and the mouse mammary adenocarcinoma TS/A cell line, previously described, were cultured in RPMI 1640 supplemented with 2 mM glutamine and 10% fetal calf serum (FCS) (GIBCO-BRL, Cergy-pontoise, France). NIH-3T3 cells and the cysteine-rich intestinal protein (ψCRIP) and 293T packaging cell lines, were purchased from the American Type Culture Collection (Rockville, Md, USA). These three cell lines were cultured in Dulbecco\'s modified Eagle\'s medium supplemented with 10 % FCS \[[@B27]\]. All cell lines were periodically tested for mycoplasma infection using a DNA hybridization probe (Stratagene, La jolla, CA, USA). Primary culture of human melanoma cells --------------------------------------- Melanoma tumor biopsies were dissected into small explants and then enzymatically digested with collagenase (5 mg/ml) and hyaluronidase (3 mg/ml) (SIGMA, Saint Quentin Fallavier, France) for 1 hour at 37°C under agitation. The cells were then centrifuged (5 minutes at 1200 rpm) and transferred to a culture flask and left to proliferate for 10 days. These cells were cultured in RPMI 1640 supplemented with 2 mM glutamine, 10% FCS, non-essential amino acids and vitamins (SIGMA, France). During the first 4 days after the first passage, cells were cultured with 0.1 mg/ml of G418 (GIBCO-BRL, Cergy-pontoise, France) to remove fibroblasts as described by Mouriaux and coll\[[@B36]\]. The quality and characteristics of the melanoma cells were studied by immunohistochemistry using anti-cytokeratin (KL-1), anti-S100, anti-HMB45/gp100 and anti-melanA/MART1 (clone A103) antibodies (all from DAKO SA, Trappes, France). Retroviral constructs --------------------- The pMDgag/pol and pBA-GALV plasmids were obtained from the Genethon (Evry, France). The tricistronic vectors are MFG-based retroviral vectors. These vectors were derived from the DFG-human-CD80 (hCD80) and DFG-human-CD70 (hCD70) vectors that have previously been described \[[@B7],[@B27]\]. The first cistron (cDNA encoding hCD70 or hCD80) was cloned between the Nco1 and BamH1sites of the MFG vector. The AUG of this molecule corresponds to the AUG of the *env*gene of the MFG vector. ### Construction of TGF^EMCV^ZEO or NEO In these vectors, the translation of the genes encoding the two co-stimulatory molecules and the cDNA encoding the G418 or zeocin resistance genes are under the control of the IRES^EMCV^. We digested the Blue-script plasmid (PKs) (Stratagene, La jolla, CA, USA) with Xho1 and EcoRV to insert the cDNA encoding IRES^EMCV^(636 bp), obtained from PCR amplification of pIRES-EGFP (Clontech, Palo Alto, CA, USA), and obtained a PKs IRES^EMCV^vector. This vector was digested by Nco1 and EcoR1 and ligated with the cDNA encoding either CD70 or CD80, obtained from the PCR amplification of DFG-CD70 or DFG-CD80 respectively. The PKs IRES^EMCV^CD70 or PKs IRES^EMCV^CD80 vectors were digested by BamH1. These inserts have been cloned into either DFG CD80 NEO or ZEO by partial digestion using BamH1, generating the (T) tricistronic vectors TFG^EMCV^NEO or TFG^EMCV^ZEO with CD80 as first cistron and CD70 as third cistron or in DFG CD70 NEO or ZEO generating the same type of tricistronic vectors but where CD70 is the first cistron and CD80 is the third cistron. ### Construction of TGF^FGF-2^ZEO, TGF^HTLV-1^ZEO, TGF^C-MYC^ZEO or NEO Through PCR amplification, we cloned the DNA encoding the IRES of the human basic Fibroblast Growth factor (FGF-2) (547 bp), the IRES of the *c-myc*oncogene (cMYC)(602 bp) and the IRES of the HTLV-1 lentivirus (249 bp) upstream from the cDNA encoding the G418 or zeocin resistance genes in the PKs plasmid. We generated different combinations of cDNAs with BamH1 sites on both the 5\' and 3\' ends: IRES^HTLV-1^-NEO, IRES^HTLV-1^-ZEO, IRES^FGF-2^-NEO, IRES^FGF-2^-ZEO, IRES^cMYC^-NEO and IRES^cMYC^-ZEO. These constructs were inserted downstream from the gene encoding the first co-stimulatory molecule on the BamH1 site. We generated the TFG^FGF-2^NEO or ZEO, or TFG^c-MYC^NEO or ZEO and TFG^HTLV-1^NEO or ZEO tricistronic vectors (Figure [1A](#F1){ref-type="fig"}). All the vectors were sequenced. Transfection and transduction ----------------------------- CRIP packaging cells were transfected with the different plasmids using the lipofectamine technique followed or not by selection with G418 (1 mg/ml) (from Life Technologies, GIBCO-BRL, France) or Zeocin (0.2 mg/ml) (CAYLA, France). The resistant CRIP clones were expanded and screened for viral titers which were of approximately 10^4^viral particles/ml/48 hours/10^6^CRIP cells. 293T cells were triple-transfected with the different plasmids: TFG, pMDgag/pol and pBA-GALV. Estimation of the viral titer in transiently transfected 293T cells showed the presence of 10^6^viral particles/ml/48 hours/10^6^293T cells. Fibroblasts or tumor cells (B16.F10, TS/A or human melanoma) were transduced with the retroviral vectors using polybrene (8 μg/ml) (SIGMA, France). The stably transduced cells were selected using G418 or Zeocin depending on the viral particles used, and the resistant cells were either cloned or pooled and used in subsequent experiments. RT-PCR ------ Cells were cultured in T-75 cm^2^culture-flasks for 72 h until they reached sub-confluence. Total RNA was isolated from a suspension, as described by Choczynski and Sacchi (1987) \[[@B37]\], of 5 × 10^6^cells using TRIZOL™ reagent following the manufacturer\'s recommendations (Life Technologies, Cergy Pontoise, F). RNA solutions were treated with 5 units of DNAse 1 (Roche Diagnostics, Mannheim, D) to remove any contaminating genomic DNA. mRNA was transcribed into cDNA using the Ready-to-go™ kit (Amersham Pharmacia Biotech. Inc. Piscataway, NJ, USA) and random primers were purchased from Life Technologies. Amplification of cDNA was carried out using 1U/100 μl of Taq DNA polymerase (Roche Diagnostics, Mannheim, D) in a PTC-100™ Programmable Thermal Controller (MJ Research, Watertown, Mass, USA) after 33 temperature cycles consisting of denaturation at 94°C (60 s), annealing at 65°C (60 s) and elongation at 72°C (120 s). The following primers were used: MFG sense primer (Trigen 1): 5\'-TGTAAAACGACGGCCAGTCACGTGAAGGCTGCCGACC-3\', ZEO anti-sense primer (trigen 5): 5\'-CAGGAAACAGCTATGACCCACCGGAACGGCACTGGTC-3\', ZEO sense primer (trigen 4): 5\'-TGTAAAACGACGGCCAGTGACCAGTGCCGTTCCGGTG-3\' and MFG anti-sense primer (trigen 10): 5\'-CAGGAAACAGCTATGACCGCCTGGACCACTGATATCCTGTC-3\'. PCR products and the lambda/hindIII molecular weight marker (Promega, Lyon, F) were separated by electrophoresis on 0,8% agarose (Roche, Mannheim, D), Tris Borate EDTA (Interchim, Monluçon, F) gels and visualized by staining with ethidium bromide. Immunostaining and flow cytometric analysis ------------------------------------------- Transduced cells were stained for membrane expression of the two co-stimulatory molecules using PE-conjugated mouse anti-human CD70 mAb and FITC-conjugated mouse anti-human CD80 mAb (Pharmingen, Hamburg, D) as previously described \[[@B7]\]. Stained cells were analyzed using a FACScalibur (Becton Dickinson, Mountain View, USA). Establishment of murine tumor models ------------------------------------ Female C57BL/6 (H-2b), or BALB/c (H-2d) mice were obtained from CERJ Janvier (St Quentin-Fallavier, France). All ear-tagged mice were kept in the special pathogen-free animal facility in our institution and were used for experiments between the age of 6 to 8 weeks. To establish subcutaneous (s.c.) tumors, 105 cells (twice the minimal tumorigenic dose) were suspended in 0.1 ml of PBS and injected s.c. into the flank. Animals were examined daily until the tumor became palpable, the diameter of the tumor, in two dimensions, was then measured twice a week. The animals were sacrificed when the tumor size reached 2.5 × 2.5 cm in the control group. The statistical significance of the data was established using the two sample student\'s t-test. A p-value of less than 0.01 was considered to be statistically significant. Authors\' contributions ======================= VDE carried out the cellular and *in vivo*studies, SB the synthesis of the retroviral vectors (molecular biology and virology). PR performed the characterization of the melanoma biopsies. LC provided all the IRES sequences. BC and ACP are responsible for conceiving this work. BC participated in the design, coordination and drafting of this manuscript. GF is the director of the laboratory. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ The authors wish to acknowledge Lourdes Gasquet and Christiane Pages from the animal facility for their excellent assistance, Jean Charles Faye for helpful discussion and Nathalie Bajonc and Daniele Berg for their technical assistance. We thank Catherine Greenland for the correction of the manuscript. Victorine Douin is supported by the French Ligue contre le Cancer. This work has been supported by a grant from the Association de Recherche sur le Cancer (ARC). Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Structure of retroviral constructs.**TFG vectors are MFG-based retroviral plasmids into which the hCD70 cDNA and hCD80 cDNAs have been cloned, with either a neoR or a zeoR selectable marker gene. The three genes are co-translated from a tricistronic mRNA which was transcribed from the 5\' long terminal repeat. The translation of the first cDNA (CD70 or CD80) depends on the 5\' long terminal repeat-gag IRES, while the translation of the selectable marker and the second cDNA (CD70 or CD80) are dependent on two IRESes from different origins. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **RT-PCR analysis of mRNA transcribed from the tricistronic vectors in stably transduced melanoma cells.**2A : Localization of the different primers. Figure 2B: cDNA obtained after reverse transcription was amplified using primers 1 and 5 (lanes 1 to 4) or primers 4 and 10 (lanes 5 to 8) as described in *Methods*. Lanes 1 and 5: RT-PCR performed on mRNA isolated from TFG^FGF-2^ZEO transduced cells (1819 and 1640 bp amplified respectively). Lanes 2 and 6: RT-PCR performed on mRNA isolated from TFG^EMCV^ZEO transduced cells (1819 and 1640 bp amplified respectively). Lanes 3 and 7: RT-PCR performed on mRNA isolated from TFG^HTLV-1^ZEO transduced cells (1510 and 1640 bp amplified respectively). Lanes 4 and 8: RT-PCR performed on mRNA isolated from mock retroviral vector transduced cells (negative controls). :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Example of co-stimulatory molecule expression by melanoma cells after transduction with TFG tricistronic vector: Immunostaining with hCD80 --(x-axis) and hCD70 -- (y-axis)3A**: non transduced melanoma cells were cultured. Ccells (3 × 10^5^) was stained for surface expression of CD70 and CD80 using specific antibodies as described in the *Methods*. **3B-3E**: Murine B16.F10 melanoma cells were transduced with the four TFG retroviral vectors. Forty eight hours after transduction, pools of cells (3 × 10^5^) were stained for surface expression of CD70 and CD80 using specific antibodies as described in the *Methods*. 2B: TFG^EMCV^ZEO transduced cells, 2C: TFG^c-MYC^ZEO transduced cells, 2D: TFG^FGF-2^ZEO transduced cells, 2E: TFG^HTLV-1^ZEO transduced cells.**3F**: Human melanoma cells were transduced with the TFG^FGF-2^ZEO. A selected pool of cells (3 × 10^5^) was stained for surface expression of CD70 and CD80 using specific antibodies as described in the *Methods*. The samples were subjected to two-color analysis by flow cytometry. Data are shown as immunofluorescence profiles with arbitrary fluorescence units (log) of FITC on the x-axis and immunofluorescence profiles with arbitrary fluorescence units (log) of PE on the y-axis. To evaluate the percentage of co-stimulatory molecules expressed by transfected cells, the marker was set to allow \< 5% positive cells in the non-transduced cells. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Effects of CD70 and CD80 co-expression by single transduced tumor cells on tumor growth in an established model.4A**: 10^5^B16.F10 wt (control cells), or CD70 and CD80 double transduced B16.F10, or CD70 and CD80 single transduced B16.F10 were injected s.c. on day 0 into female C57BL/6 mice. **4B**: 10^5^TS/A wt (control cells), or CD70 and CD80 double-transfected TS/A, or CD70 and CD80 single-transfected TS/A were injected s.c. on day 0 in female BALB/c mice. Tumor growth was monitored twice a week. The results are expressed as mean size (mm^2^) of tumors from groups that each contain five mice ± SD. The data represent mean +/- SD of a single experiment representative of three.pouet :::  :::

PubMed Central

2024-06-05T03:55:47.222281

2004-7-27

PMC514711

Background ========== Biosensors based on GFP-fusion proteins are powerful tools for observing real-time events within living cells. Insertion of GFP within another protein has produced biosensors capable of signaling intracellular events through intrinsic fluorescence changes \[[@B1],[@B2]\], fluorescence resonance energy transfer (FRET) \[[@B3],[@B4]\], and changes in sub-cellular localization \[[@B5]\]. The difficult task of finding the right insertion site to produce a biosensor can be accelerated by screening libraries of random GFP insertions \[[@B6]-[@B8]\]. The insertional strategies described to date, however, are limited in two ways. First, each insertion produces only one kind, or color, of fluorescent fusion protein. Creating the multicolored libraries necessary for co-expression or FRET analyses requires either separate rounds of insertion and screening for each fluorescent protein or additional subcloning to exchange fluorophores. Second, the efficiency of any random approach is limited to a maximum of 1:6 because a fusion protein can only be produced if the *GFP*coding sequence lands in the correct orientation and reading frame with respect to the target coding sequence. We reasoned that it might be possible to overcome these limitations by placing two different fluorescent protein coding sequences in a single transposon, either in tandem or antiparallel. Here we present the results of our initial tests with these designs. Results and Discussion ====================== The mosaic ends (MEs) that define the hyperactive Tn*5*transposon \[[@B9]\] have two possible open reading frames (ORFs) through them. We used one of these reading frames to construct the Either-Or transposon (\<EYOR\>, Figure [1A](#F1){ref-type="fig"}). \<EYOR\> carries the sequence encoding the yellow fluorescent protein (YFP) at its 5\' end, flanked by two 8 bp restriction sites (*Asc*I -- 5\' and *Srf*I -- 3\'). An identical cassette encoding cyan fluorescent protein (CFP) flanked with *Asc*I and *Srf*I sites, is positioned in the same orientation at the 3\' end of the transposon. Tn*5*transposition, *in vitro*, is only \~1% efficient \[[@B10]\], so the kanamycin resistance gene (*Kan*^*R*^) was incorporated between the two fluorescent protein cassettes. Since the *YFP*sequence has no start codon, it should only be translated if it inserts within another protein coding sequence in the correct orientation and reading frame. Plasmids with transposon insertions that are in-frame with respect to the target coding sequence can be rapidly identified by screening for YFP fluorescence in transiently transfected mammalian cells. Each of these clones produces a truncated fusion protein with YFP at the C-terminus. A stop codon is positioned downstream of the *Srf*I site to prevent translation beyond the YFP coding sequence. These truncated fusion proteins may provide additional information about which parts of the primary sequence contain trafficking signals. Full-length YFP and CFP fusion proteins are then generated in parallel from each clone by digestion with *Srf*I or *Asc*I and re-ligation. By producing identical full-length YFP and CFP fusion proteins from each in-frame insertion \<EYOR\> should double the efficiency with which multicolored fusion protein libraries can be generated. To test the \<EYOR\> transposon we targeted Prestin, an integral membrane protein expressed in outer hair cells of the cochlea and believed to be the motor responsible for their rapid changes in length in response to fluctuations in membrane voltage \[[@B11]\]. The 2.2 kb cDNA encoding Prestin was expressed in a 4.8 kb Ampicillin resistant (*Amp*^*R*^) CMV expression plasmid, pBNJ12.5. Transposon insertions that disrupt the plasmid origin or *Amp*^*R*^(together \~1.5 kb) are not recovered \[[@B8]\], so the predicted number of in-frame insertions in Prestin, is \~7%. After transposition, plasmids conferring Amp^R^and Kan^R^were isolated by standard mini-prep procedures and transiently expressed in HEK-293 cells in a 96-well format. In a random sample of 192 transposed clones, 32 produced detectable fluorescence. Eighteen of these (\~9%) were clearly localized to intracellular membranes. (Figure [2A](#F2){ref-type="fig"}). Of the remaining fluorescent proteins 11 appeared to be YFP alone, with evenly distributed fluorescence throughout the entire cell, and 3 were too dim to determine any sub-cellular localization. Sequencing revealed that the 18 proteins targeted to intracellular membranes were truncated Prestin-YFP fusion proteins resulting from in-frame insertions at 12 unique sites (Figure [2D](#F2){ref-type="fig"}). One of the YFP-like clones was an in-frame insertion in the intracellular N-terminus of Prestin (at amino acids 28--30) upstream of the first predicted transmembrane domain. The remaining 13 fluorescent proteins resulted from \<EYOR\> insertions outside the Prestin coding sequence, most of them being clustered just downstream of the CMV promoter. We could not identify an in-frame start codon (AUG) in any of these clones. There is however, an in-frame CUG codon at the 5\' end of the *Tn*5 ME sequence that may present an alternate translation initiation site in the presence of the strong CMV promoter \[[@B12]\]. To verify that \<EYOR\> could be used to generate full-length fusions with either YFP or CFP unique in-frame clones were digested in parallel with *Srf*I or *Asc*I and re-ligated. The resulting fusion constructs were transiently expressed in HEK-293 cells and screened for YFP and CFP fluorescence (Figure [2B,2C](#F2){ref-type="fig"}). All 13 unique insertion sites produced fluorescent full-length fusions with both CFP and YFP and all were localized to intracellular membranes. The \<EYOR\> design could be expanded for a wide range of protein tagging applications by replacing the secondary CFP cassette with another open reading frame. With such a transposon, YFP fluorescence would be used as a reporter to rapidly identify random in-frame insertions. Subsequent digestion with *Asc*I and re-ligation could then generate fusion proteins that might otherwise be difficult to screen for such as epitope tags, protease cleavage sites, or even a new N-terminus complete with a secretory signal peptide. Several groups have reported similar strategies based on multi-domain transposons for the random insertion of small peptide tags \[reviewed in:\[[@B13]\]\]. Like \<EYOR\>, these transposons utilize a primary reporter domain to identify in-frame insertions. Subsequent excision of the reporter domain (and selectable marker) then restores the full-length target coding sequence with an inserted peptide tag. The \<EYOR\> design is unique, however, in that its overlapping pairs of *Asc*I and *Srf*I restriction sites, allow the user to create identical full-length fusion proteins from both the reporter domain and the secondary coding sequence. The second transposon design, the Double-Barrel transposon (\<DBT\>, Figure [1B](#F1){ref-type="fig"}), encodes green and red fluorescent proteins (GFP and DsRed) in opposite orientations. This means that any \<DBT\> insertion within another protein coding sequence has a 1:3 chance of being in-frame regardless of its orientation. Therefore, \<DBT\> should double the efficiency of random fusion protein generation, by producing equal numbers of GFP and DsRed fusions. In addition to their antiparallel orientation, the *GFP*and *DsRed*coding sequences in \<DBT\> each use a different relative reading frame through the Tn*5*MEs. As in \<EYOR\>, GFP fusion proteins are created by insertions after the third nucleotide of a target codon. The *DsRed*coding sequence, however, has been shifted by 1 nucleotide relative to the Tn*5*MEs. Therefore, DsRed fusion proteins are generated by transposon insertions between the second and third nucleotides. Using different reading frames for *GFP*and *DsRed*doubles the total number of insertion sites in the target coding sequence from which fusion proteins could potentially be made. While this does not alter the frequency of in-frame insertions, it does reduce the screening cost of saturating a target clone by increasing the probability of recovering unique in-frame insertions. To test the efficiency of fusion protein generation with the \<DBT\> transposon, we targeted pCMVI-9, a CMV expression plasmid carrying cDNA encoding the type 1 IP~3~receptor (IP~3~R) \[[@B14]\]. The IP~3~R is a ligand gated Ca^2+^channel, composed of 4 homomeric subunits, expressed within the membranes of the endoplasmic reticulum. Each IP~3~R subunit is over 2700 amino acids, and creating a full-length fluorescent IP~3~R fusion protein with such a large cDNA presents a formidable challenge for traditional molecular biological techniques. The high ratio of coding sequence to vector makes it an excellent target for transposition, however, with a predicted frequency of in-frame insertions of \~11% per fluorophore for \<DBT\>. At 24 hrs post-transfection, visual screening for fluorescence of 288 Amp^R^+ Kan^R^clones in HEK-293 cells identified 44 clones that produced green fluorescent proteins. Of these, 35 displayed a uniform cytoplasmic distribution and exclusion from the nucleus (Figure [3A](#F3){ref-type="fig"}), 3 showed fluorescence throughout the entire cell and 6 were too dim to determine any sub-cellular localization. Screening at several time points between 2 and 4 days post-transfection identified 7 clones producing red fluorescent proteins, 2 of which were clearly excluded from the nucleus (Figure [3B](#F3){ref-type="fig"}). The remaining 5 red proteins were too dim to determine any sub-cellular localization. Sequencing out of the transposon confirmed that 41 of the clones encoding green proteins (\~14%) and 2 clones encoding red proteins represented in-frame insertions. Consistent with the results of the Prestin transposition, all of the proteins with sub-cellular localization different from that seen with GFP alone were the product of in-frame insertions. After digestion and religation, only 20 of the full-length fusion proteins retained detectable levels of fluorescence (18 green, 2 red). These full-length proteins displayed a dramatic shift in their distribution, with clear ER localization (Figure [3C](#F3){ref-type="fig"}). We chose *GFP*and *DsRed*to build the \<DBT\> transposon because their coding sequences are so dissimilar. Our concern was that if we chose two similar sequences, *CFP*and *YFP*for example, the antiparallel orientation of these coding sequences could produce extensive mRNA hybridization and secondary structure that would inhibit protein translation. It appears however, that DsRed is not well suited for insertion within other proteins. Indeed, DsRed has not been reported as a fusion protein in the middle of another protein, and even N- and C-terminal fusions with DsRed can be problematic \[[@B15]\], perhaps due to its being an obligate multimer \[[@B16]\]. Despite the low yield of DsRed fusions, these results demonstrate that \<DBT\> can be used to simultaneously generate full-length fusion proteins in two different reading frames. As novel fluorescent proteins are isolated from new species, or old ones are altered, this type of bi-directional transposon could potentially double the output of the screening process. Conclusions =========== The transposons described here should greatly accelerate the creation of multicolored libraries of fluorescent fusion proteins. By creating identical full-length YFP and CFP fusion proteins from each in-frame insertion, the \<EYOR\> transposon not only facilitates the generation potential FRET pairs, it enables the direct comparison of different fluorophores in otherwise identical fusion proteins. The \<DBT\> design, on the other hand, has the capacity to double both the throughput of fusion protein generation by virtue of its bi-directionality as well as the total output of novel fusion proteins through the simultaneous use of multiple reading frames. Ultimately, the ability to generate large numbers of novel fusions proteins in days rather than months, should shift the limiting rate at which novel fluorescent protein biosensors are identified to functional screening rather than protein design and construction. Methods ======= Plasmids -------- PCR and standard subcloning procedures were used to create the plasmids encoding \<EYOR\> (pBNJ55b.1), \<DBT\> (pBNJ38.5) and Prestin (pBNJ12.5) (full sequences in supplementary material). The fluorescent protein coding sequences used were Venus (YFP) \[[@B17]\], ECFP-N164H (CFP), EGFP (GFP) and DsRed2 (Clontech). The *Kan*^*R*^gene was obtained from pUniV5-His-TOPO (Invitrogen). The construction of the IP3 receptor expression plasmid, pCMVI-9, was previously described \[[@B14]\]. Tn5 transposition and plasmid isolation --------------------------------------- Transposons were amplified from their host plasmids via PCR with a single primer complementary to the 19 bp Tn*5*ME (5\'-CTGTCTCTTATACACATCT-3\') and purified as previously described \[[@B8]\]. Purified transposon and target concentrations were each quantified against an independent DNA standard using a DynaQuant 200 fluorimeter. The transposition reaction was performed according to manufacturer\'s recommendations (Epicentre Technologies, Madison, WI) with 200 ng of target DNA and a molar equivalent of purified transposon. Electrocompetent XL-10 Gold *E. coli*(Stratagene, La Jolla, CA) were transformed with 0.5 μL of the transposition reaction and plated on LB agar with Ampicillin (100 μg/mL) and Kanamycin (50 μg/mL). Parallel plating of the transformation on LB agar with Ampicillin alone was used to establish the transposition efficiency. Transposed plasmids were isolated in a 96-well format from 1.25 mL LB cultures with Eppendorf PerfectPREP-96 Vac Direct Bind miniprep kits on a PerkinElmer MultiPROBE II HT liquid handling robot and eluted in 70 μL of ddH~2~O. Visual screening ---------------- HEK-293 cells (American Type Culture Collection CRL-1573) were plated 24 hr prior to transfection, in 96-well glass bottom tissue culture plates (NalgeNUNC) at 6 × 10^4^cells in 100 μL of MEM-E with 10% fetal bovine serum. Transfections were performed with \~300 ng of plasmid DNA and 0.3 μL of Lipofectamine 2000™ (Gibco BRL) in a total volume of 50 μL of Opti-MEMI (Gibco BRL) per well. The cells were screened for fluorescence 24 hr after transfection with a 20× objective on a Zeiss inverted microscope with excitation and emission filter sets optimized for CFP/YFP or GFP/DsRed imaging (Omega, Brattleboro, VT). Sequencing and generation of full-length fusion proteins -------------------------------------------------------- Exact insertion sites were identified for all fluorescent transposed clones by sequencing 5\' out of the transposon with a primer complimentary to the \<EYOR\>*YFP*coding region (5\'-CTGCAGGCCGTAGCC-3\') or \<DBT\>*GFP*coding region (5\'-TGGCCGTTTACGTCGCCGTCCA-3\'). To generate full-length fusion proteins, plasmids with unique in-frame insertions were digested and re-ligated. After restriction digestion (100 ng of plasmid DNA and 0.5 U of *Asc*I or *Srf*I in 10 μL total volume), 1 μL of the digest reaction (\~20 ng DNA) was re-ligated with Fast-Link™ ligase (Epicentre Technologies) for 15 min at room temperature (0.5 mM ATP, 1X Fast-Link™ buffer, 1 U ligase, 7.5 μL total volume). After heat inactivation (70°C for 15 min.), XL-10 Gold *E. coli*were transformed with 0.5 μL of the ligation reaction and plated on LB agar with Ampicillin. The following day, colonies were co-inoculated in LB with Ampicillin and Ampicillin + Kanamycin to verify loss of the *Kan*^*R*^prior to plasmid isolation. Authors\' contributions ======================= D.S. conceived of the transposon designs and carried out their construction, performed the transposition and screening for fluorescent fusion proteins, and drafted the manuscript. T.H. participated in the study design, coordination, and analysis. All authors have read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 **Supplementary Material-hughes**This is a PDF file containing both plasmid maps and full sequence data for the plasmids used in this study. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ The authors would like to thank Martina Brauns and the members of the Friday Afternoon Lab Meeting for their invaluable input on this project. We would also like to thank Atsushi Miyawaki for the Venus (YFP) coding sequence, Catherine Berlot for the ECFP-N164H (CFP) coding sequence, Joseph Santos-Sacchi for the Prestin cDNA and Barbara Ehrlich for pCMVI-9. We are grateful to the staff of the HHMI Biopolymer/Keck Foundation Biotechnology Research Laboratory for their rapid and reliable DNA sequencing. This work was supported by NINDS R21 NS044883-01. D.S. is a Howard Hughes Medical Institute Predoctoral Fellow. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Two new transposons for generating multicolored GFP fusion protein libraries.**In-frame insertions of the Either-Or transposon, \<EYOR\> (A), within another coding sequence initially create truncated C-terminal YFP fusions. Alternate restriction digestion with *Asc*I or *Srf*I removes one of the fluorescent proteins and the *Kan*^*R*^. Subsequent re-ligation produces identical full-length YFP and CFP fusion proteins. The fluorescent protein is flanked by 9 amino acid linkers encoded by the Tn*5*MEs and restriction sites. The Double Barrel transposon, \<DBT\> (B), encodes green and red fluorescent proteins (GFP and DsRed) antiparallel to one another. Therefore, \<DBT\> insertions with another coding sequence have a 1:3 chance of being in-frame regardless of orientation. The *GFP*coding sequence uses the same reading frame through the Tn*5*MEs as *YFP*in \<EYOR\>. *DsRed*however, has been shifted by 1 bp to use a different reading frame through the Tn*5*MEs, doubling the number of usable insertion sites within a target coding sequence **(Note that the reading frame shown in \<DBT\> is that used for *DsRed*, translated from the lower DNA strand, and is read from right to left)**. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **In-frame \<EYOR\> transposition events in the voltage sensitive integral membrane motor protein, Prestin.**Visual screening of transiently transfected HEK-293 cells revealed truncated Prestin-YFP fusion proteins generated by in-frame \<EYOR\> transpositions (A) (scale bar = 20 μm). Alternate digestion and re-ligation of an in-frame clone with either *Srf*I or *Asc*I produces identical full-length YFP- (B) and CFP-fusion proteins (C). In-frame \<EYOR\> insertions in Prestin are identified by the first of 3 target amino acids duplicated during transposition (D). Redundant insertions are indicated by the number of recovered clones (e.g. 3x). :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Transposition of the type 1 IP~3~R with the Double Barrel Transposon (\<DBT\>).**Transient expression in HEK-293 cells allows identification of in-frame insertions producing truncated GFP- (A) or DsRed- (B) IP~3~R fusion proteins. Digestion and subsequent re-ligation produces a full-length fluorescent IP~3~R fusion protein (C) (Scale bar = 20 μm). :::  :::

PubMed Central

2024-06-05T03:55:47.225161

2004-8-18

PMC514712

Background ========== The regulation of cellular ion concentrations is an essential process in all organisms, necessary to sustain a multitude of physiological processes including pH balance and ion homeostasis. This process is accomplished mainly through membrane ion transporters. In *Escherichia coli*, among the membrane proteins that catalyze the exchange of ions across the cell membrane \[[@B1]\] are the Na^+^/H^+^antiporters NhaA, NhaB and ChaA, which are involved in sodium ion extrusion. Within *E. coli*and other enteric bacteria, antiporters encompass the primary systems responsible for adaptation to growth in conditions of high Na^+^concentrations and varying pH \[[@B2]-[@B8]\]. It is common for bacteria to have multiple systems for a similar function. The use of one system is preferred depending on the stress as a means to adapt to varying environmental conditions \[[@B9],[@B10]\]. Of the Na^+^/H^+^antiporters, ChaA is unique in that it also shows pH-independent Ca^+^/H^+^antiporter activity. ChaA is also regulated by Mg^2+^, which inhibits both its Na^+^/H^+^and Ca^+^/H^+^antiporter activity \[[@B11]\]. The Cha operon consists of 3 genes, *chaA*, *chaB*and *chaC*found at \~27 minutes on the *E. coli*chromosome \[[@B12]\]. Both ChaB and ChaC are proposed to be regulators of ChaA however, the biological function for either remains to be established. ChaB is a 76-residue protein that contains a conserved 60-residue region found in several other bacteria and baculoviruses. We report here the three-dimensional structure of ChaB determined by NMR spectroscopy and examine key differences between the ChaB families of proteins. Results and Discussion ====================== Assignment of resonances ------------------------ The ^1^H-^15^N HSQC spectrum of ChaB (Fig. [1](#F1){ref-type="fig"}) is well dispersed suggesting ChaB is a globular, folded protein. Complete ^1^H, ^15^N and ^13^C backbone assignments were made for ChaB, except residues S39 and H40, which yielded no apparent amide cross peaks. Virtually complete assignments (\> 98%) were made for the ^1^H, ^13^C, and ^15^N side chain resonances. Resonance assignments have been deposited at BMRB (code 6117). Six signals in the ^1^H-^15^N HSQC spectrum originated from the 21 residue N-terminal His-tag (Fig. [1](#F1){ref-type="fig"}). The low heteronuclear NOE values (Fig. [2A](#F2){ref-type="fig"}) and the relatively low number of long range NOE\'s (Fig. [2B](#F2){ref-type="fig"}) for these residues indicate the His-tag to be flexible in solution. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **^1^H-^15^N HSQC spectrum of ChaB at 600 MHz, pH 6.3 in 50 mM CaCl~2~**. Folded resonances are indicated by asterisks. Numbering shown includes the N-terminal His-tag (residues 1--21). The native ChaB sequence starts at P22. The unassigned peak is denoted as a question mark. Folded arginine/lysine side chain resonances are indicated by SC. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Plots of ^1^H-^15^N heteronuclear NOE, NOE constraints and RMSD statistics for ChaB.**(A) ^1^H-^15^N heteronuclear NOE acquired at 600 MHz. (B) Summary of all unassigned unambiguous NOE constraints: intra-residue, sequential, medium and long range NOEs are shown as blue, green, red and black bars respectively. (C) Backbone RMSD\'s calculated for the 17 lowest energy ChaB structures based on superposition of residues P22-S96. :::  ::: Solution structure of ChaB -------------------------- The 3D structure of ChaB (Fig. [3](#F3){ref-type="fig"}) is well defined by the structural constraints (Table [1](#T1){ref-type="table"}) and is dominated by two, relatively long central helices comprising residues H40-Q55 (helix α2) and D65-E83 (helix α3) and a small N-terminal helix (helix α1, E31-K34), which is terminated by a proline (P30). At the C-terminus, a short two-strand β-sheet is observed involving residues Y84-K86 and W92-K94. A tightly packed hydrophobic core stabilizes the overall fold of ChaB. The following hydrophobic residues have \< 10% of their surface area exposed to the solvent: Y23, L29, V33, L37, A41, I44, Y45, A48, F49, A52, A72, A76, V80, Y84, A85 and W92. Many of the hydrophobic contacts are between the two long helices (α2 and α3). The C-terminal β-sheet acts as a \"cap\" for hydrophobic residues from loop 1 (V36, L37), which connects helices α1 and α2, and residues at the N-terminus of helix α2 (A41) and the C-terminus of α3 (V80). Both central helices are largely amphipathic, with residues D43, K46, E47, D54, of helix α2 and E70, K74 and K81 of helix α3 exposed to the bulk solvent and contributing to a highly charged ChaB surface. Most notable, an area of negative charge is observed at the highly mobile loop 2 and the helices immediately surrounding it. In addition, K74, K81, K86, and K95 contribute to a positively charged area while Y56, V75 and A79 present a small hydrophobic patch. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Constraints and structural statistics for ChaB ::: [Constraints used for structure calculation (all residues)]{.underline} ------------------------------------------------------------------------------------------------- ------------- ----------------- Total NOE constraints 2140 Intraresidue NOEs (n = 0) 515 Sequential NOEs (n = 1) 432 Medium Range NOEs (n = 2,3,4) 379 Long Range NOEs (n \> 4) 486 Total Unambigous NOEs 1794 Ambiguous NOE restraints 346 Dihedral angle constraints 49 ^15^N-^1^H residual dipolar couplings 58 [Average RMSD to mean structure (Å) (residues 22--96)]{.underline} Backbone atoms 0.397 All heavy (non-hydrogen atoms) 0.807 [Average energy values (kcal mole^-1^) quoted for residues 1--96]{.underline} E~total~ -332.42 ± 9.46 E~bond~ 11.93 ± 0.89 E~angle~ 84.06 ± 1.56 E~improper~ 16.35 ± 0.58 E~VdW~ -515.11 ± 10.31 E~NOE~ 41.39 ± 3.75 E~dihedral~ 0.86 ± 0.21 E~sani~ 28.09 ± 3.11 [Deviation from idealised covalent geometry]{.underline} Bonds (Å) 0.0028 ± 0.0001 Angles (°) 0.4470 ± 0.0042 Improper (°) 0.353 ± 0.006 [RMSD from experimental data]{.underline} Distance restraints (Å) 0.016 ± 0.0007 Dihedral angle restraints (°) 0.302 ± 0.094 [Average Ramachandran statistics for 17 lowest energy structures (residues 22--96)]{.underline} Residues in most favored regions 78.2% Residues in additional allowed regions 8.6% Residues in generously allowed regions 3.3% Residues in disallowed regions 0.0% [Analysis of residual dipolar coupling]{.underline} RMSD (Hz) 1.495 ± 0.097 Q-factor 0.138 ± 0.0059 Correlation coefficient 0.98 ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Solution structure of ChaB**. Ensemble of the 17 lowest energy structures showing (A) backbone and (B) heavy-atom traces. Superposition was made over residues comprising the native ChaB sequence (P22-S96). (C) Ribbon representation of the lowest energy ChaB conformer. :::  ::: In general, the secondary structure elements of ChaB are well defined exhibiting RMSD\'s of 0.14 Å and 0.39 Å for backbone and all non-hydrogen atoms, respectively. This is confirmed by the heteronuclear NOE data, which show a 10% trimmed weighted mean of 0.77 ± 0.03 in the structured regions and indicate lack of motions on the nanosecond timescale (Fig. [2A](#F2){ref-type="fig"}). Regions connecting the secondary structural elements exhibited lower heteronuclear NOE values. In particular, the loops connecting helices α2 and α3 (loop 2, Y56-D64) and the two strands of the β-sheet (loop3, G87-K91) exhibited NOE values below 0.65, indicating large amplitude nanosecond motions in these regions. These motions manifest as regions with large RMSD values in the structural ensemble (Fig. [2C](#F2){ref-type="fig"}). The small sheet region at the C-terminus was also seen to exhibit some motional freedom particularly for the second strand. However, it is notable that its NOE values are substantially higher than the surrounding loop. The ChaB family and structurally similar proteins ------------------------------------------------- Figure [4](#F4){ref-type="fig"} shows the alignment of protein sequences related to *E. coli*ChaB. The most conserved residues make up the hydrophobic core of ChaB, particularly the two long helices and the small sheet. With the exception of P38 and A79 all these residues exhibit \< 5% solvent accessibility. These residues are critical for defining the overall fold of ChaB and suggest that all proteins within this family adopt a similar fold. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Sequence alignment of the family of ChaB protein**s. (A) Alignment of ChaB from *E. coli*aligned with a series of related proteins identified by Pfam \[34\]. In bold above alignments, are residues most conserved among ChaB proteins. Cartoon diagram above represents the secondary structure of ChaB. ChaB from *E. coli*(in bold) *Salmonella typhi*(Q8XGJ2)and *Methanosarcina mazei*(Q8PYS9) are classified as group I ChaB proteins. Group II ChaB proteins are all found in Baculoviridae. The figure was created using BOXSHADE (EMBnet). Identical amino acids are highlighted in black and homologous residues in gray. (B) Sequence alignments of ChaB and related proteins σ^70^and σ^RN^based structural composition (see text for details). :::  ::: Structural homologues of ChaB in the PDB were identified using the DALI server \[[@B13]\] (Fig. [5](#F5){ref-type="fig"}). This yielded several matches with fragments of other structures, the best match being sigma factor σ^70^(PDB code 1SIG, \[[@B14]\]) with a DALI Z-score of 4.7. A DALI-Z score greater than 2.0 is considered structurally similar. Another sigma factor, σ^RN^\[[@B15]\], with little sequence homology to σ^70^was identified with a DALI-Z score of 3.5. ChaB, however, exhibits no significant sequence similarity with these proteins. Sigma factors are proteins, which bind to DNA dependent RNA polymerases to form the holeoenzyme \[[@B16],[@B17]\]. Although the observed structural similarities do not define a functional role for ChaB it is worth noting that the σ^RN^domain is classified in the cyclin-box fold of proteins \[[@B15]\], a class of proteins that bind a diverse set of proteins and nucleic acids \[[@B18]\]. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Comparison of ChaB with structurally similar proteins.**Structural similarities between (A) ChaB, (B) σ^70^(PDB code 1SIG \[14\]) and (C) σ^RN^(PDB code 1H3L, \[15\]) proteins identified from the DALI server. The three helices in ChaB are colour coded, with the equivalent helices in σ^70^and σ^RN^similarly coloured. :::  ::: An interesting observation can be made when aligning the sequences of ChaB, σ^70^and σ^RN^based on their structural similarity (Fig. [4B](#F4){ref-type="fig"}). Residues in the three structures with \< 10% of the surface exposed to the solvent are highlighted in red. Clearly, the hydrophobic core, critical for the ChaB fold (marked above the sequence alignments in Fig. [4](#F4){ref-type="fig"}) is also important for the fold observed in the sigma factors. Key hydrophobic residues appear in similar locations in their \"structural space\" between the three proteins, forging contacts important for stabilizing the fold. These residues are among the most conserved in the two sigma factor families and within ChaB proteins. Loop 2 has weak affinity for divalent ions ------------------------------------------ Given the proposed function of ChaB as a regulator and the effect of magnesium as an inhibitor of the Ca^+^/H^+^antiporter ChaA, we examined the influence of calcium and magnesium ions on ChaB ^15^N-^1^H chemical shifts. The pattern of perturbed shifts (summarised in Fig. [6A](#F6){ref-type="fig"} for Ca^2+^) indicates that the highly charged (Fig. [6B](#F6){ref-type="fig"}) flexible loop 2 and surrounding regions are most important for binding. Chemical shift perturbations of similar magnitude and direction were witnessed upon addition of MgCl~2~indicating that Mg^2+^has a similar binding site and affinity for ChaB as Ca^2+^. The observed association constant for CaCl~2~is weak (the K~D~was estimated to be \> 10 mM by NMR) and not likely to be physiologically significant. Given ChaB\'s proposed role as a regulatory protein, it is possible that the affinity for Ca^2+^or Mg^2+^is increased in the presence of ChaA or ChaC. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **CaCl~2~titration and surface potential map of ChaB.**(A) Summary of the perturbations of CaCl~2~on the backbone ^1^H-^15^N chemical shifts of ChaB. The change in chemical shifts (determined from a weighted vector sum of ^1^H and ^15^N ppm deviations) are mapped onto the structure of ChaB using a colour gradient from blue, to red to yellow, where yellow is the largest perturbation and blue the smallest. Residues that could not be analyzed such as overlapping residues or residues that do not exhibit NH resonances are coloured grey. (B) Potential map of the surface of ChaB calculated using MOLMOL \[35\] shown in the same orientation as (A). Residues most perturbed by Ca^2+^cluster around a highly negatively charged patch on the ChaB surface comprising a flexible loop. :::  ::: A functional role for loop 2? ----------------------------- ChaB proteins are classified into two major groups based on their sequence alignments (Fig. [4](#F4){ref-type="fig"}). Group I consists of ChaB proteins found in bacteria (*E. coli*and *Salmonella typhi*) and archeabacteria (*Methanosarcina mazei*), while Group II contain ChaB related proteins that are found in Baculoviridae. Thus far, no ChaB domains have been identified in vertebrate and plant species. One major difference between the two classes of ChaB proteins is the presence of the charged loop (loop 2, Fig. [4](#F4){ref-type="fig"}) that we have shown to bind weakly to Ca^2+^and Mg^2+^ions. The EMBL European bioinformatic database annotates ChaB proteins found within group II (Baculoviridae, Fig. [4](#F4){ref-type="fig"}) as putative DNA binding proteins. Interestingly, the σ-factor domains that are structurally similar to ChaB are known to bind DNA at the position equivalent to helix α3 in ChaB. However, the composition of the corresponding loop in sigma factors is more hydrophobic and/or shorter than in ChaB. Members of the ChaB sequences belonging to group I (Fig. [4](#F4){ref-type="fig"}) are annotated as cation transport regulators based on being part of the ChaA operon. The alignment results suggest that the loop 2 region, which is only observed in the group I family of ChaB, is correlated to its function as a cation transport regulator protein. Clearly, further experiments are required to test this hypothesis. Conclusion ========== As part of the *E. coli*structural genomics project, we report the first 3D structure of a member from the ChaB protein family. *E. coli*ChaB is a putative cation transport regulator protein whose structure resembles the cyclin-box fold. ChaB was shown to have weak affinity for calcium and magnesium ions at a highly charged and mobile loop that is only present in ChaB family members associated with a cation transporter. We hypothesise that this loop may play a role in the function of ChaB as a regulator of cation transport. Methods ======= Cloning, expression and purification of ChaB -------------------------------------------- The gene encoding full length ChaB (residues 1--76) was amplified from genomic *E. coli*DNA strain O157:H7 using oligos OPI403 AAAAAAG[GATCCC]{.underline}CGTATAAAACGAAAAGCGACCTG and OPI499 AAAAAA[GAATTC]{.underline}TTACGATTTTTTATGCCATTTATCATCA. Underlined are restriction sites BamHI and EcoRI for both oligos, respectively. The product was cloned into the BamHI/EcoRI site of pFO-1. The plasmid pFO-1 is a tailored pET-15b vector (Novagen Inc., Madison, WI), which contains an extended poly-linker region and an 8× N-terminal histidine tag with a modified thrombin cleavage site. The ChaB construct (plasmid ID: pPI489) was expressed in *E. coli*BL21-Gold (DE3) cells (Stratagene) as an 8× His-tagged fusion protein. At an OD~600~of 0.8, the cells were induced with 1 mM IPTG and grown for another 3 hours at 30°C. The protein was purified to homogeneity by absorption onto a Ni^2+^charged chelating sepharose column (Amersham Biosciences) under native conditions. The recombinant protein used for NMR studies consists of the ChaB sequence with an extra 21 residues (MGSSHHHHHHHHSSGFNPRGS) at the N-terminus containing the 8× His-tag and a thrombin cleavage site. This tag replaced the first residue, a methionine, of the genomic sequence of ChaB. In the analysis *vide ante*ChaB refers to residues P22-S96 of our construct (i.e. the wild type ChaB sequence excluding the first Met residue). Thus, ChaB begins at P22 in our numbering scheme, corresponding to P2 in the native ChaB sequence. The mass of ChaB was confirmed by SDS-PAGE and electrospray mass spectroscopy. NMR Spectroscopy ---------------- Uniform enrichment of ChaB with ^15^N and/or ^13^C was achieved by growing the bacteria in M9 minimal medium supplemented with BME vitamins (SIGMA) and (^15^NH~4~)~2~SO~4~and/or ^13^C~6~-glucose as the sole nitrogen and carbon sources at 37°C. ChaB was purified as described above. NMR samples were obtained by exchanging ChaB into an NMR buffer comprising 50 mM CaCl~2~at pH 6.3 using a PD-10 column and subsequent concentration to \~200--300 μL using an Amicon Ultra-4 (5 KDa cutoff, Millipore). Typical protein concentrations ranged from 1.5--2.0 mM. NMR spectra for resonance assignments were recorded at 303 K on a Bruker Avance DRX 600 MHz spectrometer equipped with a triple-resonance CryoProbe and processed with NMRPipe \[[@B21]\]. Backbone ^1^H, ^13^C and ^15^N assignments were completed from CBCA(CO)NH, CBCANH and HBHA(CBCACO)NH spectra using a combination of NMRView \[[@B22]\] and SMARTNOTEBOOK (a module designed for semi-automated assignment in NMRView) \[[@B23]\] packages. ^1^H, ^13^C and ^15^N sidechain assignments were obtained by manual analysis of the H(CC)(CO)NH, C(C)(CO)NH and HCCH-TOCSY experiments using NMRView and in-house written scripts. ^1^H, ^13^C and ^15^N chemical shifts were referenced to DSS according to the IUPAC recommendation \[[@B24]\]. Distance constraints were obtained from a simultaneous 3D ^13^C/^15^N-edited NOESY experiment (τ~m~= 120 ms) in 90% H~2~O/10% D~2~O, and ^13^C-edited NOESY (τ~m~= 100 ms) and ^13^C-edited NOESY (aromatic region) (τ~m~= 100 ms) experiments acquired in 99.9% D~2~O. The experiments in D~2~O were acquired at 800 MHz on a Varian INOVA spectrometer at NANUC). A 4D ^13^C-^13^C edited NOESY experiment (τ~m~= 100 ms) was acquired at 600 MHz to resolve ambiguities involving methyl groups. For all experiments at 600 MHz, the minimal number of scans dictated by the phase cycle was used in combination extensive folding in ^15^N and ^13^C to reduce experimentation time. Additional restraints used in structure calculations were: dihedral restraints, derived from ^3J^HN-Cα coupling constants obtained from the HNHA experiment \[[@B25]\] and ^1^H-^15^N residual dipolar couplings extracted from comparison of IPAP-HSQC experiments recorded on ChaB with and without 11 mg/mL Pf1 phage \[[@B26]\]. For the measurement of dipolar couplings, the NMR buffer was altered to 50 mM phosphate and 100 mM NaCl, pH 6.3 since the presence of CaCl~2~precipitated the Pf1 phage. Steady state {^1^H}-^15^N NOE spectra were acquired in an interleaved manner in which each individual FID was collected with and without presaturation and a recycle delay of 4 s \[[@B27]\]. Saturation was achieved using a train of 120° pulses separated by 5 ms for a total irradiation time of 3 s. Structure calculations ---------------------- A set of unambiguous NOE constraints were extracted from the 3D-NOESY spectra and used in conjunction with dihedral angle restraints to generate a preliminary fold of ChaB using CNS1.1 \[[@B28]\]. The resulting structures were used as model templates for automated assignment of NOE peaks using the ARIA 1.1 package \[[@B29]\]. In many cases, the 4D ^13^C-^13^C NOESY experiment was important for manually assigning a number of ambiguous assignments. A total of 1794 unambiguous and 346 ambiguous NOE restraints were obtained from this method and used in combination with dihedral restraints to calculate an ensemble of ChaB structures using CNS \[[@B28]\]. These structures were further refined using residual dipolar coupling restraints. The axial and rhombic components of the alignment tensor were obtained from the histogram method \[[@B30]\] and optimized by a grid search \[[@B31]\] and determined to be D~a~= 13.7 and R = 0.325. Only residues exhibiting a heteronuclear NOE \> 0.65 were included as residual dipolar couplings. Seventeen lowest energy structures with the fewest violations were selected to represent the ChaB structure. No NOE violations over 0.2Å were observed. Structural statistics for this ensemble as calculated by CNS \[[@B28]\], PROCHECK \[[@B32]\] and SSIA \[[@B33]\] are summarised in Table [1](#T1){ref-type="table"}. The coordinates have been deposited in the RCSB under PDB code 1SG7. Titration with calcium and magnesium ------------------------------------ The effect of calcium on ChaB was determined from addition of aliquots of 5 M CaCl~2~or 2 M MgCl~2~to ^15^N labeled ChaB. Prior to titration, metal impurities were removed by addition of EDTA to the ChaB sample followed by exchange into 20 mM Bis-Tris buffer, pH 6.3 using a PD-10 column. Aliquots of CaCl~2~or MgCl~2~were added up to a final concentration of 50 mM. Minimal changes in pH and volume were ensured throughout. Chemical shift perturbations were measured as a weighted vector sum of the ^1^H and ^15^N chemical deviations: {\[(Δ^1^H ppm)^2^+ (Δ^15^N ppm × 0.2)^2^\]^0.5^}. List of abbreviations ===================== NMR: nuclear magnetic resonance NOE: nuclear Overhauser enhancement HSQC: heteronuclear single quantum coherence PPM: parts per million RMSD: root mean squared deviation PDB: Protein Data Bank Authors\' contributions ======================= MJO expressed and purified isoptically enriched ChaB, collected all NMR spectra at 600 MHz, processed and analyzed NMR data, performed structural calculations and structural refinement. NS identified ChaB among a series of *E. coli*proteins cloned as part of the structural genomics initiative. PI completed the initial cloning of chaB. NS expressed, purified and characterized ChaB by mass spectrometry. MJO drafted and NS contributed to the written manuscript. KG coordinated and provided financial support for this study. Acknowledgements ================ The Montreal-Kingston Bacterial Structural Genomics Initiative supported this work under a Canadian Institute of Health Research (CIHR) grant to KG. We would like to acknowledge the Canadian National High Field NMR Center (NANUC) for their assistance and use of their facilities. CIHR, the Natural Science and Engineering Council of Canada (NSERC) and the University of Alberta fund the operation of NANUC.

PubMed Central

2024-06-05T03:55:47.227279

2004-8-11

PMC514713

Background ========== It is generally believed that a major portion of post-traumatic neuronal necrosis after brain injury does not result from diffuse primary injury, but rather from a secondary process. The injury appears to trigger a cascade of molecular events that lead to gradual vascular and neuronal tissue degeneration, thus destroying the anatomical substrate necessary for the neurological recovery. A large body of evidence obtained from a wide variety of experimental studies of acute CNS injury strongly suggest that various reactive oxygen species (ROS) and nitrogen species (RNS) have been implicated in the progressive secondary degeneration that follows the injury \[[@B1]-[@B3]\]. ROS and RNS have been associated with secondary injury that amplifies the magnitude of final neuronal damage. Both biochemical analyses and studies with transgenic mice has shown that ROS/RNS production persists for many hours after the initial insult. This offers a potential therapeutic window for pharmacologic intervention of clinical relevance. Several classes of pharmacologic mimetics of superoxide dismutase/catalase have been synthesized. Evaluation of these catalytic antioxidants in laboratory models of acute brain injury has shown both robust neuroprotection and a prolonged therapeutic window at doses apparently devoid of neurotoxicity \[[@B4],[@B5]\]. Another important aspect determining extent of oxidative stress mediated post-reperfusion injury subsequent to ischemia is the antioxidant status of the affected tissue as it is of great importance for the primary endogenous defence against free radical attack. Various types of antioxidant enzymes like Cu-Zn superoxide dismutase (SOD), peroxidases such as glutathione peroxidase (GPx), catalase have been reported to be neuroprotective \[[@B5],[@B6]\]. Rubia cordifolia (RC), Fagonia cretica linn (FC) and Tinospora cordifolia (TC) are tropical herbs and have been extensively used in the treatment of various types of haematological, hepatic, neurological and inflammatory conditions \[[@B7]\]. The antioxidant and anti-inflammatory and immuno-modulatory properties of RC and TC has also been well documented \[[@B8]-[@B11]\]. Although RC has been reportedly used as an Ayurvedic medication in a wide variety of conditions, reports regarding the use of FC and TC are not available. In light of the aforementioned properties of RC and TC and relatively scant studies on FC we hypothesized that these herbs may overcome the oxidative stress mediated injury during ischemic neuronal injury via modulating the antioxidant pool of the cells. In order to test this hypothesis we devised a two pronged strategy in the present study to evaluate the effect of the drugs on the status of antioxidant enzymes such as GPx and Cu-Zn SOD, the levels of reduced glutathione (GSH), the principal redox regulator of the cell, and the status of nitric oxide (NO) generation the in the hippocampal slices subjected to oxygen-glucose deprivation (OGD). Methods ======= Reagents -------- All reagents unless stated otherwise were obtained from Sigma Chemical Co. USA, Merck, India and Loba Chemie, India. All animals used were as per the institutional animal ethics committee approval. Preparation of hippocampal slices --------------------------------- Method for preparation of slices was similar to those of Taylor & Weber\[[@B12]\]. Normal New Zealand male Wistar rats weighing between 180--220 gm. were anaesthetized with ether and decapitated, whole brain was removed and placed in ice-cold oxygenated artificial cerebrospinal fluid (aCSF) containing (in mM) NaCl -- 125, KCl -- 3.5, CaCl~2~-- 2.0, MgSO~4~-- 1.0, NaHCO~3~-- 26, Na~2~HPO~4~-- 1.25 and D-glucose -- 10 mM. The chilled brain was removed and the hippocampal regions were dissected, slices of 0.6 mm were obtained on a modified Stadie -- Riggs microtome and were immersed in oxygenated aCSF between 22--28°C for 120 min to allow the tissue to get stabilized. Induction of in vitro oxygen-glucose deprivation (OGD) hypoxic ischemia ----------------------------------------------------------------------- Hippocampal Slices were subjected to OGD (oxygen glucose deprivation) \[[@B12]\]. (Taylor + Weber) by suspending the slices in D-Glucose deficient aCSF equilibrated with 95% Nitrogen gas and 5% CO~2~, until the partial oxygen pressure was less then 20% as that of normoxic aCSF (pO~2~was measured using a blood gas analyzer and was found to be ≤ 35 mm Hg). Slices were incubated in oxygen glucose deficient aCSF for 30 min at 37°C. Transferring the slices to reperfusion chamber containing aCSF started reperfusion. The extent of injury was estimated by assaying LDH release in the medium using an LDH kit. Study groups ------------ Hippocampal slices were divided into 3 groups of OGD for present study. a\) The overall control group consisted of the slices immediately stabilized for 2 hours at 22--26°C in normoxic aCSF. b\) The experimental group consisted of stabilized slices subjected to OGD without any drug treatment. c\) The experimental treated group consisted of OGD slices reperfused in the medium (aCSF) containing 1\) 2 mg/ml concentration of RC, TC and FC for 30 min. 2\) Ascorbic acid and Reduced glutathione equivalent to their content in RC, FC and TC for 30 min. Biochemical assays ------------------ In all the above study groups the following biochemical parameters were monitored. 1\) Free Cu-Zn superoxide dismutase (Cu-Zn SOD) was assayed using the method of Beyer et al\[[@B13]\]. 2\) Reduced Glutathione (GSH) was assayed by the method of Teitz \[[@B14]\]. 3\) Glutathione peroxidase (GPx) was assayed by the method of Paglia et al. \[[@B15]\]. 4\) Nitric oxide (NO) was measured as nitrite (NO~2~) in the supernatant by the method of Green et al. \[[@B16]\]. 5\) Protein was estimated by the method of Lowry et al. \[[@B17]\]. Electron Paramagnetic Resonance (EPR) measurement ------------------------------------------------- EPR (Magnettech X-band Miniscope MS-100, Berlin Germany) was used to establish the antioxidant effect of RC, FC and TC with respect to superoxide anion (O~2~^.^), hydroxyl radicals (^.^OH), nitric oxide (NO) radical and peroxynitrite anion (ONOO) generated from pyrogallol, menadione, DETA-No and Sin-1 respectively. All the free radical donors (100 μM -- 500 μM) were incubated in phosphate buffered saline (pH 7.4, 37°C) containing the spin-trap, Tempone-H (1 mM), oxidation of which generates 4-oxo-tempo with a characteristic three-line EPR signal centred at 3365 G. Development of this signal was monitored for 60 min from addition of the oxidising species and compared to parallel incubations containing RC, FC and TC (10 or 50 μM, n = 3). The amplitude of the first line of the spectrum was measured; data are expressed in arbitrary units. The EPR parameters for these experiments were as follows: Microwave frequency -- 9.4 GHz; microwave power -- 20 mW; modulation frequency -- 100 kHz; modulation amplitude -- 1500 mG; centre field 3365 G; sweep width 50 G; sweep time 20 sec; No. of passes -- 1; receiver gain 3E1. Reverse Transcriptase Polymerase chain reaction (RT-PCR) -------------------------------------------------------- GCLC (Glutamyl -- cysteinyl ligase catalytic subunit), iNOS (inducible nitric oxide synthase) and Cu-Zn SOD mRNA was isolated using TriZOL reagent from the hippocampal slices and reverse transcribed to study the effect of RC, FC and TC on the expression status of the above genes in OGD untreated slices after a period of 24 and 48 hours post OGD. Gene expression in treated groups was studied after 24 hours. GAPDH served as the house-keeping gene. After an initial reverse transcription, the cDNAs obtained were amplified using the following respective primers and PCR conditions: GCLC was amplified by 32 thermal cycles of 94°C for 30 s, 55°C for 30 s, 72°C for 2 mins followed by an extension at 72°C for 10 mins using primers: *for*\' 5\'gtggtactgctcaccagagtgatcct and *rev*\' 5\'tgatccagtaactctgggcattcaca. iNOS was amplified at 94°C for 1 min, 60°C for 1 min, and 72°C for 1 min for a total of 27 cycles followed by a 10-min extension at 72°C using primers: *for*\'5\'-gtgttccaccaggagatgttg-3\' and *rev*\' 5\'-tggggcagtctccattgcca-3\'. Cu-Zn SOD was amplified at 94°C for 45 s, 56°C for 30 s, and 72°C for 45 s for a total of 23 cycles followed by a 10-min extension at 72°C using primers: *for*\'5\'-tctaagaaacatggcggtcc-3\' and rev\'5\'-cagttagcaggccagcagat-3\'. GAPDH was amplified using 20 thermal cycles of 94°C for 45 s, 60°C for 45 s, and 72°C for 1 min 30 s, followed by final extension for 10 mins at 72°C, primer used were *for\'*5\'ccacccatggcaaattccatggca and *rev*5\'tctagacggcaggtcaggtcaacc. Results ======= Intracellular levels of GSH is differentially modulated by RC, FC and TC ------------------------------------------------------------------------ Figure [1](#F1){ref-type="fig"} depicts the effect of the three herbs RC, FC and TC on the intracellular GSH levels in treated rat hippocampal slices post glutamate toxicity. A significant elevation in the GSH level was recorded in untreated controls as compared to the stabilized group (p \< 0.001). Where as RC and FC did show an increase in the GSH levels as compared to TC. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Effect of RC, FC and TC on the cytosolic GSH levels during OGD in rat hippocampal slices. \* p \<0.001 compared to stabilized, and + p \< 0.001 compared to OGD. Results expressed are a mean of 6 independent experiments ± SEM. :::  ::: Effect of RC, FC and TC on GCLC gene expression ----------------------------------------------- Figure [2](#F2){ref-type="fig"} shows the effect of RC, FC and TC on GCLC (Glutamyl -- cysteinyl light chain) RT-PCR gene expression, all the three herbs were found to be inducers of the gene expression as compared to the untreated controls. Expression of housekeeping gene GAPDH was unaltered in all the lanes. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### a: Representative GCLC mRNA expression in relation to GAPDH in control, OGD and OGD + drug treated rat hippocampal slices. Lane 1 -- Control, Lane 2 -- OGD 24 hr, Lane 3 -- OGD 48 hr, Lane 4 -- OGD + FC, Lane 5 -- OGD+TC, Lane 6 -- OGD+RC. b: Percentage densitometric expression of GCL mRNA expression in relation to GAPDH in control, OGD and OGD + drug treated rat hippocampal slices. \* p \< 0.01 and \*\* p \< 0.001 versus control, + p \< 0.01 versus \* & \*\* and ++ p \< 0.001 versus \* & \*\*. Results expressed are a mean of 6 independent experiments ± SEM. :::  ::: RC, FC and TC directly scavenge free radicals --------------------------------------------- RC, FC and TC contain polyphenols, a class of compound that can directly interact with electrophillic species and thus can act as a direct scavenger of free radicals. To test this hypothesis we employed EPR spectroscopy to study the interaction of RC, FC and TC with certain free radicals such as O~2~^.-^, ^.^OH, NO and ONOO using specific donors. EPR spectroscopy revealed that addition of RC, FC or TC (10 μg/ml) to the O~2~^.-^generator, pyrogallol (100 μM; n = 3, Fig [3a](#F3){ref-type="fig"}) or the ^.^OH generator, menadione (500 μM; n = 3, Fig [3b](#F3){ref-type="fig"}), NO generator, DETA-NO (100 μM; n = 3, Fig [3c](#F3){ref-type="fig"}) and ONOO generator Sin-1 (100 μM; n = 3, Fig [3d](#F3){ref-type="fig"}) in the absence of a spin trap failed to generate a spin signal over a 60 min period. This indicated, either the herbs interacted with the free radicals covalently and therefore yielded no paramagnetic signals or did not interact at all. However, 3-line spin signals characteristic of formation of the stable radical 4-oxo-tempo developed in a time-dependent manner when the same oxidant generating substances were incubated with the recognised spin trap, tempone-H (1 mM). Co-incubation of RC, FC and TC with pyrogallol (100 μM) in the presence of tempone-H caused a significant inhibition of spin signal development over a 60 min time period (p \< 0.001, 2-way ANOVA -- results of Bonferroni post-hoc analyses and non-linear curve fits are shown on the figure; n = 3). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Non-linear curve fit analysis of the effect of RC, FC and TC on scavenging free radicals such as A) O~2~^-^, B) OH^.^, C) NO and D) ONOO^-^generated by pyrogallol, menadione, DETA-NO and SIN-1 respectively. The reference curve fit is depicted in the respective groups. Results expressed are a mean of 3 independent experiments ± SEM. :::  ::: Intracellular levels of GPx are increased by RC, FC and TC ---------------------------------------------------------- Figure [4](#F4){ref-type="fig"} depicts the effect of the three herbs RC, FC and TC on the intracellular GPx levels in treated rat hippocampal slices post glutamate toxicity. A significant rise in the GPx level was recorded in treated slices as compared to the stabilized and control group (p \< 0.001). Here too RC and FC show an increase in the GPx levels as compared to TC. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Effect of RC, FC and TC on the glutathione peroxidase levels in rat hippocampal slices. + p \< 0.01 versus stabilized and ++ p \< 0.001 versus OGD. Results expressed are a mean of 4 independent experiments ± SEM. :::  ::: Cu-Zn SOD gene expression and enzyme level is upregulated by RC, FC and TC -------------------------------------------------------------------------- Figure [5](#F5){ref-type="fig"} depicts the effect of the three herbs RC, FC and TC on the expression and activity levels of intracellular Cu-Zn SOD in stabilised, untreated and treated rat hippocampal slices post OGD. A significant drop in the SOD level was recorded in untreated OGD slices as compared to the stabilized group (p \< 0.001). A significant rise in the levels of both SOD gene expression and the enzyme activity was observed for RC, FC and TC (p \< 0.001). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Effect of RC, FC and TC on Cu-Zn SOD gene expression and cytosolic free levels during OGD in rat hippocampal slices. RT-PCR pictograph of Cu-Zn SOD is shown aligned to the corresponding groups. \* p \< 0.001 versus stabilized, + p \< 0.001 versus OGD. Results expressed are a mean of 6 independent experiments ± SEM. :::  ::: RC, FC and TC and exogenous antioxidants decrease nitric oxide generation and iNOS gene expression -------------------------------------------------------------------------------------------------- The effect of exogenously added antioxidants namely ascorbic acid and reduced glutathione equivalent to their concentrations found in three herbs RC, FC and TC on nitric oxide generation in comparison to the three herb extracts is shown in figure [6](#F6){ref-type="fig"}. All the three herbs along with the exogenously added GSH and Vit C significantly inhibited NO~2~generation in the treated OGD slices. The decrease in the level of NO~2~was found to parallel a decrease observed for the expression of the iNOS gene in the same group of the hippocampal slices (Figure [7](#F7){ref-type="fig"}). ::: {#F6 .fig} Figure 6 ::: {.caption} ###### Effect of RC, FC and TC and amounts of ascorbic acid and GSH equivalent to that present in the herbs (RC\*, FC\* and TC\* respectively) on NO~2~generation in hippocampal slices. + p \<0.001 versus control (100%) and \*\* p \< 0.001 versus OGD. Results expressed are a mean of 3 independent experiments ± SEM. :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### Effect of RC, FC and TC on the iNOS gene expression in rat hippocampal slices subjected to OGD and OGD+RC/FC/TC. Lane 1 -- Control, Lane 2 -- OGD 24 hr, Lane 3 -- OGD 48 hr, Lane 4 -- OGD + FC, Lane 5 -- OGD+TC, Lane 6 -- OGD+RC. \*p \< 0.05 vs control, \*\*p \< 0.001 vs control, +p \< 0.001 vs \*\*. Results expressed are a mean of 6 independent experiments ± SEM. :::  ::: Discussion ========== Ischemic cells are known to be under oxidative stress and hence experience oxidative injuries like membrane alterations, lipid peroxidation, increased ionic influx, especially Ca and Na etc. \[[@B18]-[@B20]\] elevated levels of O~2~^-^, NO, NO~2~, decreased activity of Ca^2+^Mg^2+^ATPase and Na^+^/K^+^ATPase \[[@B21]\]. In the present study we have reported the protective effects of RC, FC and TC during OGD insult to rat hippocampal slices. GSH was protected from depletion during the insult and this protection could be correlated to an elevation in the expression of the g-GCS gene. The peroxide scavenging enzyme, glutathione peroxidase (GPx) activity was also significantly restored by the three plant extracts. Treatment of the OGD-hippocampal slices with pure vitamin C and GSH in proportions equivalent to that found in the extracts exhibited a parallel effect on NO generation as to that found with the drugs. RC, FC and TC effectively reduced free radical levels by mechanisms involving increased expression of Cu-Zn SOD, decreased expression of iNOS and simultaneous scavenging of the free radicals such as O~2~^-^, OH^.^, NO and ONOO. Overall, RC, FC and TC exhibit potential cytoprotective ability in rat hippocampal slices subjected to OGD. In addition to its role as an antioxidant, the GSH status of a cell is critical for various other biological events that include transcriptional activation of specific genes and modulation of redox-sensitive signal transduction and hence pro-inflammatory processes during cerebral ischemia \[[@B22]\]. GSH also plays a crucial role in the regulation of expression of several redox-sensitive antioxidant and anti-inflammatory genes \[[@B23]\], processes which are aggravated especially, post-ischemic insult as a result of reperfusion of white blood cells to the injured area \[[@B24]\]. As a result there is a rapid loss of reducing equivalents of the cell and hence an onset of oxidative stress. The oxidative stress further leads to the upregulation of expression of a wide variety of pro-inflammatory cytokines, including adhesion molecules, all of which contribute to tissue injury, apoptosis/necrosis \[[@B25],[@B26]\]. Therefore maintenance of GSH pool and other antioxidant levels is critical to cell survival and adaptation to the ischemic injury \[[@B27]\]. In response to the battery of free radicals generated during ischemia, the cells initially neutralize the oxidative challenge via GSH mediated antioxidant mechanisms. However, a rapid decline in the levels of GSH soon follows which ultimately leads to tissue injury. Therefore it is imperative that any therapeutic intervention should be able to cater to this deficiency observed during ischemia/OGD. Our results show that RC, FC and TC were able to reverse the GSH levels, which was significantly depleted during OGD. This restorative/protective activity of RC, FC and TC could be partly attributed to their native antioxidant contents (GSH = 8.33 ± 0.5, 10.26 ± 0.55 and 6.94 ± 0.49, Vit C = 27.52 ± 0.93, 32.99 ± 1.03 and 41.86 ± 0.68 and Polyphenols = 18.33 ± 2.02, 11.88 ± 1.33 and 21.00 ± 1.26 mg/g extract respectively). However, it was not clear at this juncture as to how much of these antioxidants are actually bio-available, an area which is out of scope of the present study. On the other hand RC, FC and TC may exert such a restorative effect by increasing the synthesis of GSH in the cells. To test this hypothesis we studied the expression status of γ-GCS gene during OGD in both untreated and treated hippocampal slices. RC, FC and TC showed a positive inductive effect on the γ-GCS gene expression after 48 hours. Although significant, it is to be noted that the three drugs exhibited partial differences in their response towards the expression of the GCLC gene expression and in restoration of the GSH levels. This indicates that the drugs may act via other mechanisms that may have a sparing effect on the GSH levels. Scavenging of the free radicals is one such mechanism whereby depletion of GSH is prevented \[[@B28],[@B29]\]. In order to test the later hypothesis we examined whether or not RC, FC and TC could directly interact with the free radicals such as O~2~^-^, OH^.^, NO and ONOO using pyrogallol, menadione, DETA-NO and Sin-1 as respective donors. Electron paramagnetic resonance study has revealed a significant scavenging effect of the three drugs on the free radicals chosen for the study. The results were more pronounced for OH^.^and NO radicals as depicted by the non-linear curve fit analysis. The effects on O~2~^-^, and ONOO radicals were also found to be quite promising. Therefore, direct scavenging of the free radicals is an important mechanism by which the drugs may exert their cytoprotective effect, not only by sparing GSH utilization by free radicals but also preventing the free radical mediated tissue damage. GSH depletion can also be brought about by its abnormal redirection towards neutralisation of the oxidants. Within the normal metabolic course of the cell, reduction of the prostanoid hydroperoxides to their respective hydroxides, is catalysed by the enzyme glutathione peroxidase (GPx), which requires GSH as one of the co-substrates for the reaction \[[@B30],[@B31]\]. This is basically a protective mechanism of the cell against the harmful lipid peroxides generated during their synthesis and oxidative stress. In the present study we have recorded a significantly decreased GPx activity in OGD hippocampal slices as compared to the controls. The observed decrement may be attributed to the depleted GSH levels either due to increased utilisation and/or diminished activity of γ-GCS. Furthermore, diminished GPx activity, at least in part, indicates cellular accumulation of the lipid hydro peroxides, which can potentially turn on a chain reaction wherein more unsaturated lipids become targets for further peroxidative tissue injury. The ability of RC, FC and TC to enhance the GPx activity is therefore an important finding since one of the protective mechanisms of the herbs under study might be mediated via upregulation of the GPx activity. This effect may further be explained in view of the fact that the herbs themselves contain an appreciable amount of GSH and ascorbate. A large body of evidence suggests that another important intracellular enzyme Cu-Zn SOD is found to be neuroprotective in nature along with GPx and GSH \[[@B32],[@B33]\]. We observed a significant drop in the cytosolic free levels of Cu-Zn SOD in the untreated OGD slices as compared to the stabilized and the three herbs RC, FC and TC significantly restored the levels of the antioxidant enzyme. However, it was not clear at this juncture as to the mechanism involved in such a restoration. Oxidative stress is known to induce the synthesis of Cu-Zn SOD as a defensive mechanism aimed at containing the oxidant levels generated during inflammatory/ ischemic conditions \[[@B34]\]. Cu-Zn SOD is one of the first lines of defense against free radicals such as O~2~^-^and NO and acts as a direct scavenger of these extremely potent hazardous species. We therefore investigated whether the antioxidant properties exhibited by the three herbs also involve their modulatory effect on the expression of the Cu-Zn SOD expression. The increased expression of the Cu-Zn SOD gene in RC, FC and TC treated OGD slices clearly confirm the positive modulatory effect the three drugs have on the cellular antioxidant system. Increased expression of Cu-Zn SOD in response to these herbs is a crucial finding especially since this enzyme is directly implicated in the scavenging of not only O~2~^-^but also the more potent toxicant ONOO \[[@B35]\]. Thus the three herbs show potent antioxidant properties via modulating the expression of the antioxidant genes such as GCLC and Cu-Zn SOD. The mechanism of antioxidant and neuroprotective effects reported above for RC, FC and TC are new findings. However, as to what components of the drugs bring about such cell-protective effects could not be fully ascertained. Preliminary analysis has revealed that all the three herbs have significant amounts of GSH and Vit C and another important antioxidant, polyphenol. In addition inductively coupled plasma spectroscopic analysis have also revealed the presence of important trace elements in the three herbs (personal observations (Zn- 13.81, 18.19 and 14.94, Cu- 3.16, 9.46 and 13.79, Vd- 30.00, 18.55 and 26.00, Se- 1.79, 3.33 and 0.28 and Mo- 0.39, 1.64 and 1.15 ppm in RC, FC and TC respectively). We therefore thought that the three herbs exert their ameliorative properties due to their antioxidant and trace element contents. In order to test this hypothesis we investigated the effect of GSH and Vit C on NO generation in OGD treated and untreated hippocampal slices. The amount of GSH and Vit C used were equivalent to that found in the three herbs. Our results indicate that GSH and the Vit C components of the herbs are at least partly responsible for the attenuation of NO generation in the treated OGD slices. The effects of RC, FC and TC were almost similar to those recorded for GSH and Vit C. NO is an important neurotransmitter in the brain and also can be rendered harmful due to unprecedented generation during hypoxic/ischemic/inflammatory conditions \[[@B20],[@B36]\]. In addition, NO can combine with O~2~^-^to form a more toxic ONOO, which is extremely deleterious to the cells \[[@B37]\]. Hence our observation of the ability of the herbs to attenuate NO generation is not only new but also is indicative of the possible mechanism by which these herbs might exert their protective functions. It is interesting to note that RC, FC and TC also increase Cu-Zn SOD expression thus contributing to the decrease in ONOO formation due to an increased scavenging of O~2~^-^. Furthermore the three herbs were found to repress the expression of the iNOS gene, which is considered to be an important damaging factor during hypoxia/ischemia \[[@B37]\]. Thus the decrease in generation of NO by RC, FC and TC is not only due to direct scavenging by the herbs (fig [3](#F3){ref-type="fig"}) but also via the transcriptional modulation of the iNOS gene, which is induced during OGD. Conclusions =========== In conclusion, RC, FC and TC exert cell/neuroprotective properties via preventing the depletion and increasing GSH levels by inducing GCLC expression, by reducing oxidant levels via direct scavenging, decreasing iNOS expression and by increasing the antioxidant gene Cu-Zn SOD. Further protective ability may be attributed to the enhanced activity of GPx brought about by the herbs. The antioxidant contents of the herbs appear to be important components for the observed effects. However, further investigations are required to ascertain the role of individual constituents in the efficacy of the above described properties of the herbs in order to ascribe potential pharmacological applications to these herbs. Competing interests =================== Authors do not have any competing interest with anyone whatsoever. Authors\' contributions ======================= AKR -- This work is a part of AKR\'s PhD thesis and he has done all the experiments, data collection and processing and manuscript preparation. MGM -- He is co-supervisor for AKR and was involved in the ideology, manuscript preparation and providing facilities for the work. SKB -- He is the supervisor for AKR and was involved in the main ideology and study design of the work, manuscript preparation and performing EPR studies and training AKR for the respective techniques. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1472-6882/4/11/prepub>

PubMed Central

2024-06-05T03:55:47.229623

2004-8-13

PMC514714

Background ========== Pneumonia (with influenza) is the leading infectious disease cause of death in the United States and the sixth leading cause of death overall \[[@B1]\] According to national mortality data from the CDC blacks suffer disproportionately from this disease, with blacks having a higher incidence of pneumonia and a 1.4 times higher age-adjusted odds of death from pneumonia as compared to non-Hispanic whites \[[@B1]-[@B4]\] In contrast several studies of racial variations in short-term mortality have demonstrated that black patients hospitalized with pneumonia are less likely to die in the hospital than whites. \[[@B5],[@B6]\] As part of the Pneumonia Medical Quality Improvement Study (MQIS) a national expert panel established a set of process of care measures for patients hospitalized with CAP \[[@B7]\] Several studies have demonstrated that some of these processes of care, especially prompt antibiotic administration within 4 or 8 hours of presentation, is associated with decreased mortality for inpatients with CAP \[[@B7]-[@B9]\] Although several studies have demonstrated significant racial differences in pneumonia care no one has examined whether there are racial variations in all of these explicit processes of care for patients with CAP. \[[@B5],[@B6],[@B10]-[@B13]\] The aims of this paper are to 1) examine whether there are significant racial differences in the processes of care that have been associated with mortality for patients hospitalized with CAP, and 2) to examine the relative risks of death within 30-days for blacks versus whites. Methods ======= Study patients -------------- KePRO, the Medicare Peer Review Organization for Pennsylvania, obtained these data as part of the Pneumonia MQIS project, whose goals is to assess and improve the quality of care for Medicare patients hospitalized with CAP. The study population was Medicare fee-for-service inpatients hospitalized at participating hospitals in Pennsylvania between 10/1/1998 and 9/30/1999. Inclusion criteria included having a primary ICD-9 diagnosis of pneumonia (480.0--483.99; 485--487.0), or a primary diagnosis of respiratory failure (518.81) or sepsis (038. XX) with a secondary diagnosis of pneumonia. Only the first qualifying discharge was considered for each patient. Among the 204 hospitals functioning in PA during the study period, 101 agreed to participate in this study. For each hospital, a random sample of up to 60 discharges with qualifying ICD-9 codes was selected. For hospitals with fewer than 60 qualifying discharges, all charts were selected. In most cases, chart review data was collected by trained record abstractors either on site from the original record, or from photocopies sent to the offices of the Quality Improvement Organization. In two cases, data were collected by the hospital\'s own staff using an approved QIO data collection instrument (n = 2). Patients were excluded if they had no working diagnosis of pneumonia on admission or received care limited to comfort measures, left the hospital \"against medical advice\", or were transferred from another acute care hospital. Patients whose race was not white or black were also excluded. Data abstraction ---------------- Chart review data included demographics, comorbid conditions, physical exam findings, laboratory data, and chest radiograph information. In addition, data on important processes of care for patients hospitalized with CAP were obtained by chart abstraction. These processes of care included: first antibiotics within 8 hours of admission, collection of blood cultures prior to antibiotic administration, oxygen saturation measurement within 24 hours of presentation, and concordance of antibiotic therapy with national guidelines. \[[@B7]\] After initial training the abstractors performed data collection on charts were assessed using gold-standard cases that had been previously evaluated by multiple expert abstractors. If the abstractors did not achieve 95% accuracy, they underwent further training until they had an error rate of less than 5%. In addition, 10% of charts were reabstracted during the review process to monitor the accuracy of chart review. The error rate for these reabstracted charts remained less than 5%. Risk adjustment --------------- The pneumonia severity index (PSI) was used to assess severity of illness at presentation \[[@B14]\] The PSI is a validated prediction rule for 30-day mortality in patients with CAP. Patients are classified into one of five risk classes based on three demographic characteristics, five comorbid illnesses, five physical examination findings, and seven laboratory and radiographic findings at the time of presentation. The PSI was developed and validated using data from a large prospective cohort study, in which 30-day mortality ranged from 0.1% for Class I to 27% for Class V for patients. \[[@B14]\] Sampling -------- Due to the relatively small number of black patients in the cohort (n = 240) we performed a modified resampling procedure of the white cohort with matching to the black patients on age and gender \[[@B15]\] We included all black patients in the study sample, and performed multiple resampling of three white patients matched for age (\< 65, 65--74, 75--84, and ≥ 85) and gender to each black patient in the sample. Matching was used to filter out demographic imbalances between the populations. This resampling was performed 10 times and the results were pooled for analysis. Statistical analyses -------------------- Univariate statistics were used to compare sociodemographic and clinical characteristics between white and blacks patients. Categorical variables were analyzed using the Chi-square test and continuous variables were analyzed using Student\'s t-test. Separate discrete conditional logistic regression models were estimated for each of the individual process of care measures, and for 30-day mortality \[[@B16]\] The PSI score and race were entered as independent variables into the models. In addition we assessed the significance of any clinical variables not included in the PSI and significant at P \< 0.10 into regression models using a step-wise forward method. However none of these additional variables were significant so they were excluded from the models. Interactions terms were assessed for each of the models however none were statistically significant so they were not included in any of the models. Results ======= Of 4889 charts requested, the complete medical record was available for 4823. Of these, 4034 patients, 240 of whom were black, were eligible for inclusion in the study. Patients were excluded because they were neither white or black (N = 231) or because they had no working diagnosis of pneumonia on admission (n = 413), their care was restricted to comfort measures (n = 173), they were transferred from another acute care facility (n = 37), or they left \"against medical advice\" (n = 14). For each of the ten resamplings 720 white patients were sampled and matched to the 240 black patients based upon the age and gender as previously discussed. The clinical and demographic characteristics of the study population are presented in Table [1](#T1){ref-type="table"}. For our analysis of racial differences in care, the age and gender distribution of the whites was similar to that of the blacks because of our matching strategy. However, blacks continued to have higher PSI scores, indicating greater severity of illness, as well as more commonly having each of the comorbid conditions (malignancy, chronic renal disease, liver disease, congestive heart failure and history of stroke) that contribute to the PSI. There were no other statistically significant differences between the two groups. In univariate analysis mortality at 30-days was 7.8% for whites and 5.8% for blacks (p = 0.3), and 82.1% of whites received antibiotics within 8 hours as compared to 75.7% of blacks (p = 0.04). Regarding blood culture performance, 96.4% of white and 97.1% of blacks had blood cultures obtained within 24 hours, and 84.8% of whites and 77.8% of blacks had blood cultures obtained prior to antibiotics (p = 0.03). Oxygenation saturation was assessed within 4 hours of 88.9% of whites and 93.9% of blacks (p = 0.03). Figures [1](#F1){ref-type="fig"} through [5](#F5){ref-type="fig"} are forest plots that demonstrate the effect of race on the dependent variables. These plots show each of the 10 samplings and the results of the pooled analysis. These figures demonstrate the significant variability between the random samples for the different dependent measures. In the regression models, after adjusting for severity of illness with the PSI, black patients were significantly less likely to receive antibiotics within 8 hours with an odds ratio (OR) of 0.63 and 95% confidence interval (CI) of 0.41 to 0.97. Black patients also had a trend towards decreased all-cause mortality at 30-day with an OR of 0.4 and 95% CI of 0.16 to 1.0. There were no significant differences between whites and blacks in regards to obtaining blood cultures prior to antibiotics (OR 0.69, 95% CI 0.32--1.47), oxygenation assessment within 24 hours (OR 1.61, 95% CI 0.85--3.04), or use of guideline concordant antibiotics (OR 0.86, 95% CI 0.62-1.71). Discussion ========== This study found significant racial differences in an important process of care for patients with CAP, specifically time to antibiotic administration. Our results support the previous studies of racial variation in pneumonia care which demonstrated racial variations in care for patients hospitalized with community-acquired pneumonia. \[[@B10]-[@B13]\] Our study also suggests that these variations may have clinically important outcomes since the process measures used to assess quality of care in this study have been previously associated with increased 30-day mortality. \[[@B7]-[@B9]\] Our study is also consistent with previous studies which found that blacks hospitalized with CAP have lower short-term mortality rates as compared to do whites \[[@B5],[@B6]\] It is unclear why this would be the case. Possible explanations include confounders that we were not able to control for, or other important factors, which were not examined that may significantly vary by race such as sociodemographic characteristics or differences in immune response. Racial variations in CAP are important to assess since unlike coronary artery bypass surgery, hemodialysis, and many other conditions that have been studied, the inpatient treatment of CAP is largely outside of the control of the patient. Although the patient has input into being admitted to the hospital, after that point the patient has little input into the processes of care such as choice and timing of antibiotics, diagnostic testing or location of care. This has several advantages in studies where researchers seek to determine if racial differences in care reflect patient preferences, provider decisions or some negotiation between them. There are several possible explanations for our findings of racial variations in these processes of care. Besides the obvious conclusion that there may be biases that affect care there are several other possible factors that may be responsible that we are not able to examine. One possible factor is that there are geographic or other factors that results in blacks presenting for admission at hospitals with overall lower quality of care for patients with CAP. Although we were not able to control for this factor other studies have suggested that the reverse is usually true. \[[@B11],[@B13]\] That is, minorities are more likely than whites to receive their care at tertiary teaching hospitals, which on average provide superior care as compared to other hospitals. \[[@B12],[@B17]\] To attempt to adjust for imbalances between black and white patients we used a modified resampling technique to generate 10 samples of white patients, which were matched to the black population. We then pooled these results over the 10 samples. This approach allowed us to obtain a more robust estimate of the effect racial variation may have on mortality and processes of care then would be obtained from a single random matched sample \[[@B15]\] Interestingly it also demonstrates the potential biases that may be present if only a single sampling is performed for a matched analysis. The forest plots (figures [1](#F1){ref-type="fig"},[2](#F2){ref-type="fig"},[3](#F3){ref-type="fig"},[4](#F4){ref-type="fig"},[5](#F5){ref-type="fig"}) demonstrate that for some individual samples obtained significantly different results than the pooled analysis. We feel that this technique strengthens our demonstrated results. There are several limitations that should be acknowledged. First we did not have information on the physicians, hospitals, or the geographic locations of the providers so we were not able to adjust for clustering. Second our study was limited to Medicare patients hospitalized in Pennsylvania. It will also be important to examine whether patients with other types of insurance, such as Medicaid and managed care, and from other states have similar outcomes. In addition we were also unable to assess the robustness of our analysis using traditional techniques such as model cross validation on a new independent sample, or by randomly subdividing our current sample into a training and test samples, due to our small sample size. However we do have quasi-replication, at least in the white sample, by the multiple sampling that we performed. Finally we were unable to adjust for potential bias in pulse oximetry since this is a retrospective study. However recent work \[[@B18],[@B19]\] questions the idea that pulse oximetry does not perform as well in those with increase pigmentation as compared to those with lighter pigmentation. Therefore we feel that it is unlikely that this would systematically bias the results of our study. Conclusions =========== Despite these limitations, we believe our results, and the results of other studies, allow us to conclude that blacks are less likely than whites to have processes of care that are considered to represent superior quality of CAP care. Despite this, blacks with CAP have similar, and perhaps lower mortality than whites. Further research should both investigate how to make sure all patients receive optimal CAP care, and identify the factors responsible for the paradoxical advantage in survival that is seen among blacks. Competing interests =================== None declared. Authors\' contributions ======================= EM and JW conceived the study. EM and JC were responsible for the analysis. EM was responsible for the initial draft of the manuscript. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1472-6963/4/20/prepub> Acknowledgements ================ This work was supported by a contract from the Center for Medicare and Medicaid Services, and Dr. Mortensen was supported by AHRQ individual National Research Service Award grant F32 HS00135. The analyses upon which this publication is based were performed under Contract Number 500-99-PA01, entitled \"Utilization and Quality Control Peer Review Organization for the State of Pennsylvania,\" sponsored by the Centers for Medicare & Medicaid Services, Department of Health and Human Services. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government. The author(s) assume full responsibility for the accuracy and completeness of the ideas presented. This article is a direct result of the Health Care Quality Improvement Program initiated by the Centers for Medicare & Medicaid Services, which has encouraged identification of quality improvement projects derived from analysis of patterns of care, and therefore required no special funding on the part of the contractor. Ideas and contributions to the author concerning experience in engaging with issues presented are welcomed. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Results of regression analysis for race with the dependent variable initial antibiotics within 8 hours.**Forest plot demonstrating the results of the 10 individual conditional logistic regressions models for the dependent variable initial antibiotics within 8 hours after adjusting for severity of illness using the PSI. The presented relative risk with 95% confidence intervals is for the independent variable race (dichotomized into black versus white). :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Results of regression analysis for race with the dependent variable obtaining blood cultures prior to initial antibiotic dose.**Forest plot demonstrating the results of the 10 individual conditional logistic regressions models for the dependent variable blood cultures obtained prior to initial antibiotic dose after adjusting for severity of illness using the PSI. The presented relative risk with 95% confidence intervals is for the independent variable race (dichotomized into black versus white). :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Results of regression analysis for race with the dependent variable assessment of oxygen saturation within 24 hours.**Forest plot demonstrating the results of the 10 individual conditional logistic regressions models for the dependent variable assessing oxygen saturation within 24 hours after adjusting for severity of illness using the PSI. The presented relative risk with 95% confidence intervals is for the independent variable race (dichotomized into black versus white). :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Results of regression analysis for race with the dependent variable use of guideline-concordant antibiotics.**Forest plot demonstrating the results of the 10 individual conditional logistic regressions models for the dependent variable use of guideline-concordant antibiotics after adjusting for severity of illness using the PSI. The presented relative risk with 95% confidence intervals is for the independent variable race (dichotomized into black versus white). :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Results of regression analysis for race with the dependent variable mortality at 30-days.**Forest plot demonstrating the results of the 10 individual conditional logistic regressions models for the dependent variable all-cause mortality at 30-days after adjusting for severity of illness using the PSI. The presented relative risk with 95% confidence intervals is for the independent variable race (dichotomized into black versus white). :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Univariate characteristics of patients with community-acquired pneumonia by race ::: **Number and % of Patients** ------------------------------ ------------------------------ ------ ----- ------ ------- **Age** 1.00 \<65 132 18.3 44 18.3 65--74 189 26.3 63 26.3 75--84 255 35.4 85 35.4 \>84 144 20.0 48 20.0 **Gender** Male 312 43.3 104 43.3 Female 408 56.6 136 56.7 **Length of Stay** 1--2 days 63 8.8 23 9.6 0.663 3--6 days 375 52.2 116 48.3 7--13 days 220 30.6 76 31.7 \>13 days 60 8.4 25 10.4 Nursing home resident 148 21.0 51 22.2 0.774 **Pneumonia Severity Index** 0.001 I-III 280 39.0 68 28.8 IV 296 41.2 97 41.1 V 142 19.8 71 30.1 **Comorbid conditions** Malignancy 57 8.0 28 12.2 0.073 Chronic renal disease 75 10.6 58 25.2 0.000 Liver disease 9 1.3 5 2.2 0.039 Congestive heart failure 199 20.0 80 34.9 0.057 History of stroke 105 14.9 55 23.9 0.002 \*The average frequency over the 10 random samples :::

PubMed Central

2024-06-05T03:55:47.231819

2004-8-10

PMC514715

Background ========== Lymphatic filariasis, identified as one of the leading causes of permanent disability worldwide, has been targeted for global elimination \[[@B1],[@B2]\]. Interrupting transmission and controlling morbidity are the twin pillars of the global filariasis elimination programme \[[@B3]\]. Mass drug administration (MDA) for control of transmission began in the filarial endemic areas of Sri Lanka in 1999 and to-date 5-6 rounds of treatment have been completed. Morbidity management for those with filarial lymphoedema is still in its infancy. A regime of rigorous skin hygiene and simple self-help measures such as limb elevation, exercise, use of topical antibiotics and antifungals aimed at minimizing episodes of acute dermatolymphangioadenitis (ADLA) attacks and lymph stasis is the model recommended by the World Health Organization for management of filarial lymphoedema \[[@B4]\]. Little information is available on the methods of treatment practiced by patients affected by filarial lymphoedema in Sri Lanka. Furthermore, the frequency and the duration of debilitating acute attacks in Sri Lankan patients with filarial lymphoedema remains unclear. The objectives of this study were to determine the knowledge, practices and perceptions regarding lymphoedema management and the burden of ADLA attacks among patients with lymphoedema attending filariasis morbidity control clinics in an endemic area. Methods ======= The study was carried out in the Gampaha district in the Western province of Sri Lanka in November 2003. Ethical clearance for the study was obtained from the Ethics Committee, Faculty of Medicine, University of Kelaniya. Patients with lymphoedema attending morbidity control clinics conducted by the National Antifilariasis Campaign (AFC) were selected as the study population. These clinics are distributed in the operational areas of the AFC (Western, Northwestern and Southern provinces of Sri Lanka) and offer treatment exclusively to microfilaraemics and patients with chronic filariasis. Two weekly filariasis clinics (3.5 km apart) serve a population of 2,066,096 in a 1,387 km^2^area of the Gampaha district. A structured questionnaire was formulated to gather information on knowledge of lymphoedema management and limb care activities practiced along with frequency, treatment seeking behavior, cost and the degree of disability incurred in relation to ADLA attacks. A medical officer interviewed all patients regarding their disease history, examined and staged the limb affected by lymphoedema using World Health Organization criteria \[[@B5]\]. The presence of entry lesions, state of skin and nail hygiene of the affected part were recorded. Informed consent was obtained from the study participants prior to administering the questionnaire. The pre-tested questionnaire was administered in the local language: each patient was interviewed in depth by medical and paramedical staff (trained in lymphatic filariasis disability management) to explore their knowledge of lymphoedema management and the limb care activities practiced. The questions on lymphoedema management knowledge focused on the degree of importance attached to limb washing, elevation, exercise, nail hygiene, minimizing skin trauma, use of foot wear, intake of diethylcarbamazine citrate (DEC) and use of antiseptics/ topical antibiotics on entry lesions. The limb care activities practiced by the subjects were assessed by exploring how frequently they practiced the activities mentioned above in the knowledge component of the questionnaire. Questions were asked on prevalence of acute attacks during the preceding year, treatment seeking behavior during an acute attack, treatment cost per acute attack and the degree of disability incurred by an acute attack. An episode of ADLA was defined by the following criteria; painful swelling of the affected part with increased local warmth, redness and tenderness with or without associated constitutional symptoms such as fever, nausea and vomiting. Patient perceptions of disease reversibility and community attitudes towards their altered physical appearance were also explored. At the end of the interview all patients were educated on lymphoedema management practices recommended by the World Health Organization \[[@B5]\]. Results ======= A total of 66 patients \[male:female ratio 22 (43.3%): 44 (66.7%)\] with a mean age of 51.18 years (SD ± 13.9) were enrolled in the study. A substantial proportion of this study sample (24.3%) had received only primary school level education. Most subjects had lymphoedema of a single lower limb \[n = 56, 84.8%\]. Both lower limbs were affected in 6 (9%), a single upper limb in 3 (4.4%) and a single upper limb and a breast in one subject. The oedema was completely reversible (stage 1) in 21 (31.8%) subjects; 30 (45.5%) were in stage 2. Subjects with non-reversible lymphoedema had associated shallow skin folds 11 (16.7%) (stage 3), one (1.5%) had skin knobs (stage 4) and three (4.5%) had deep skin folds (stage 5). The common entry lesions were sole fissures (n = 14), minor injuries (n = 5) and eczema (n = 2). Knowledge --------- Knowledge and attitudes regarding lymphoedema management are summarized in Figure [1](#F1){ref-type="fig"}. Almost two thirds of the population were aware of the importance of skin and nail hygiene, limb elevation and use of foot wear. However, exercising the affected part was considered important by only 21 (31.8%) subjects. Regular treatment with DEC and avoidance of certain food items (fatty foods) were considered as important measures by 92% and 30.3% of the subjects respectively. Most patients (71.2%) had received advice regarding lymphoedema management from the medical officer at the morbidity control clinic (65.2%), from general practitioners (7.6%) and public health inspectors (6.7%). Few had also acquired their knowledge by reading leaflets / booklets on lymphatic filariasis (18.2%). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Lymphoedema management knowledge of the population.**Washing: Washing the affected part with soap and water Avoidance of foods: Avoidance of food items Elevation: Keeping the affected limb elevated Exercise: Exercising the affected limb Nail hygiene: Keeping the nails of the affected limb clean and short Trauma reduction: Minimizing trauma to the affected part Entry lesion survey: Examining the affected part for entry lesions Antiseptics: Use of antiseptics on entry lesions on the affected part Footwear: Use of footwear for lower limb lymphoedema :::  ::: Practices --------- Most of the study subjects washed the affected body part with soap and water either daily (n = 29, 40.9%) or twice daily (n = 32, 48.5%). A significant proportion of the study sample (n = 19, 29%) also used coir (coconut fibre) to rub the skin while washing. Half and one third of the study population practiced limb elevation on a regular basis while lying and sitting down respectively. Exercising the affected limb was practiced on a regular basis only by a minority (n = 4, 6%) of the patients. Footwear was used on a regular basis by 34.8% of the subjects, while 48.5% used them only outdoors. 7.6% of the sample did not use footwear at all. Almost all the subjects took DEC daily (n = 59, 89.4%) or periodically (n = 7, 10.6%). Only 14.3% required assistance to clean the affected body part and it was usually a family member who assisted them. The need for assistance was associated with advanced age and/or stage of lymphoedema and ADLA attacks. Assistance was required not only for cleaning the affected part (fetching water, washing) but also for surveying for entry lesions owing to failing sight in the elderly. ADLA attacks ------------ Thirty-one (46.97%) subjects had one or more ADLA attacks during the preceding year. Although around 50% of those who reported acute attacks suffered only one, some had experienced as many as 12 attacks during the past year (Table [1](#T1){ref-type="table"}). The mean number of ADLA attacks tended to rise with stage of lymphoedema. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Frequency of ADLA attacks according to the stage of lymphoedema ::: ADLA attacks/year\> Numbers affected according to lymphoedema stage --------------------- ------------------------------------------------- ------ ------ --- ------ ---- ---- 1 7 6 2 0 1 0 0 2 0 1 0 0 0 0 0 3 2 3 0 0 0 0 0 4 1 2 0 0 1 0 0 5 0 0 1 0 0 0 0 6 0 0 2 0 0 0 0 12 0 0 1 0 1 0 0 Mean 0.81 0.83 2.82 0 5.67 \- \- ::: Out patient departments (OPD) of Government hospitals (44.9%), private practitioners (30.6%) and filariasis clinics (14.3%) were the preferred sources of treatment for acute attacks. The treatment cost of an acute attack ranged from Rs.100 to 3500 (1 US\$ = 100 SLR) with an average cost of Rs. 737.91 per attack for those who sought private care (n = 48). The average duration of an acute attack was 3.5 days. Fifty two percent of those who had experienced an acute attack in the past (n = 48) were totally incapacitated (unable to attend to any domestic / economic activity) while 31.3% were moderately incapacitated (able to attend to some domestic / economic activity) for the duration of the acute attack. Almost all subjects (n = 60, 92.3%) believed that their lymphoedema was reversible and treatment with antiparasitic drugs was identified as the most important therapeutic option for reversing/ halting the progression of the disease. Only 18.2% of the study sample felt that they were being shunned by the society and that their altered physical appearance elicited various reactions ranging from sympathy to fear and ridicule from the community. Non-clinic attendees -------------------- During the period of this study an attempt was also made to collect comparable data from non-clinic attendees. However, it was possible to identify only 10 subjects, all seeking treatment in the OPD of the University Teaching Hospital. These 10 patients were in lymphoedema stages I (n = 2), II (n = 2) and III (n = 6) respectively. According to the data elicited by using the same questionnaire, 8 of these 10 non-clinic attendees (80%) were unaware of the currently recommended disability management measures. Further analysis was not carried out since the sample size was not comparable with that of clinic attendees. Discussion ========== The results of this survey indicate that the majority of patients attending filariasis clinics are aware of the importance of the currently recommended morbidity control measures. It was also encouraging to note that many practiced at least some of the morbidity control measures that they had learnt, especially washing the affected part with soap and clean water once or twice a day. Although many were aware of the importance of minimizing skin trauma, a sizeable proportion used coir to clean the skin of the affected part during washing. This is a harmful practice which could traumatize the skin and lead to entry lesions; therefore demonstration of suitable alternatives is clearly necessary. A community educational programme may be especially beneficial as it would also involve the family members of the affected population who are often the caregivers of patients affected with severe grades of lymphoedema. Involving the family in the care of the patient will also help to reduce the feelings of stigma, isolation and neglect experienced by some of the patients. Limb elevation was identified as an important measure by most patients but not practiced on a regular basis by many. Poor compliance with limb elevation has also been reported in Southern India \[[@B6]\]. The authors attributed this to inconvenience and the belief that it caused only a temporary reduction in lymphoedema. Movement and elevation has been shown to play an important role on the lymphatic and venous systems of the body \[[@B7]-[@B9]\]. However, only 31.8% of the study population was aware of the importance of exercise and very few practiced it on a regular basis. Patients with lymphoedema tend to become increasingly immobile and the affected limb is most often in a dependent position causing venous hypertension and resultant overloading of the failing lymphatics \[[@B9]\]. Therefore, health education emphasizing the long-term benefits of exercise and limb elevation should be provided in a comprehensive manner to patients and caregivers. It is also important to ensure that footwear is used indoors as well as outdoors, as the commonest entry lesions noted in the study population were sole fissures which are highly likely to become infected when walking barefoot. In contrast to previous reports \[[@B10]\], fungal infections of the interdigital spaces were not seen in any participants in the present study -- perhaps because most of the study population had early grades of lymphoedema. The practice of prescribing DEC on a regular basis to patients affected with chronic disease needs to be discouraged as there is no conclusive evidence regarding any beneficial effects of long term use of DEC in the management of lymphoedema \[[@B10],[@B11]\]. Misconceptions regarding the need to avoid fatty foods seem to be based on the belief that such foods may increase the volume (fatness) of the affected limb. Such beliefs may have originated from health messages of the lymphatic filariasis program with regard to management of chyluria. These misconceptions need to be corrected, as patients tend to avoid some commonly available food items. Although the data regarding acute attacks may not be highly accurate as it was based on one-year recall, there is reason to believe that recall data reliably reflect the burden imposed by ADLA attacks on the affected population \[[@B12],[@B13]\]. Where patients had sought private care for ADLA attacks, average total expenditure on treatment was about Rs. 737.91 per attack -- it is not surprising, therefore, that more than 50% of the population (59.1%) sought treatment from Government health facilities which provide free health care. Income lost by patients during an ADLA attack due to physical incapacitation was not determined in this study. The findings of this study may not be generalized to all patients with lymphoedema in the Gampaha district as the subjects surveyed were patients already enrolled in morbidity control clinics. Limited information gathered from non-clinic attendees indicated that the level of knowledge among clinic attendees might be far superior to that of non-clinic attendees. Lymphatic filariasis continues to be a major public health burden. Globally an estimated 15 million suffer disfiguring symptoms of lymphoedema and elephantiasis \[[@B14]\]. Although the national burden of lymphoedema has not been fully quantified, a prevalence of 3% has been reported from some parts of the country \[[@B15]\]. Therefore, in Sri Lanka, morbidity management needs to be strengthened. Conclusions =========== Referral of lymphoedema patients to morbidity control clinics is recommended as they appear to play an important role in the dissemination of morbidity control knowledge. The services of these clinics needs to be further improved to ensure that all clinic attendees receive proper education on disability management. Training and education of specialists in the management of lymphoedema may be a useful investment for eliminating the handicap caused by chronic lymphatic filariasis. List of abbreviations used ========================== ADLA: Acute dematolymphangioadenitis MDA: Mass drug administration programme AFC: Anti Filariasis Campaign DEC: Diethylcarbamazine citrate SD: Standard deviation OPD: Out patients Department SLR: Sri Lanka rupee US\$: U.S. dollar Competing interests =================== None declared Authors Contributions ===================== TC designed the questionnaire, conducted the survey and drafted the manuscript. RP Participated in data collection and interpretation. NS provided overall supervision of the study and preparation of the manuscript. All authors read and approved the final manuscript. Acknowledgements ================ We acknowledge the excellent support received by Drs. Jayasekera, M.A.K.T Perera, A.V.S.N. D. Ranasinghe, Mr M.Y.D. Dayanath. We also thank Mr. K.B.A.T. Bandara for assistance received in the preparation of the manuscript. Finally we would also like to thank the patients who participated in the survey.

PubMed Central

2024-06-05T03:55:47.233615

2004-8-3

PMC514716

Background ========== This study is an observational epidemiological investigation of breast cancer incidence in Massachusetts. It examines geographic variations over a ten year period using both purely spatial and space-time models to determine whether observed fluctuations in incidence rates are random or whether fluctuations represent statistically significant deviations from randomness. This study examines whether apparent excesses are stable over time, or are temporary, and also determines whether excesses, high or low, can be accounted for by risk factors such as socioeconomic status (SES) or urban/rural status. This study demonstrates how surveillance data can be analyzed to identify those geographic areas that warrant closer attention, the basis for determining the need for public health action or to aid in assessing the effectiveness of intervention programs \[[@B1]\]. Massachusetts has been included in studies of inter-region and intra-region variability of breast cancer incidence and mortality in the United States. Several of these studies aggregated breast cancer mortality data to the regional or county level \[[@B2]-[@B8]\]. Laden and colleagues studied regional variation using 3603 incident cases among nurses from eleven U.S. states \[[@B9]\]. They compared California, the Northeast, and the Midwest to the South; no significant excess incidence of breast cancer was observed in the Northeast. However, because regions and counties are large geographic areas, such studies can miss variability at smaller geographic levels, such as tracts within counties. A recent study on the geographic distribution of the proportion of late-stage breast cancer cases diagnosed in Massachusetts females between 1982 and 1986 aggregated cases to town, ZIP Code, and census tract levels \[[@B10]\]. The town-, ZIP Code-, and census tract-level analyses all identified approximately the same statistically significantly high area in western Massachusetts. The current study examines the incidence of invasive breast cancer with patients diagnosed between 1988 and 1997. Results ======= Principal component analysis ---------------------------- The principal components analysis performed on the seven SES variables revealed two components. The two components accounted for about 80% of the variance among the seven economic measures with the loadings of each variable on the two components shown in Table [1](#T1){ref-type="table"}. The first component had high positive loadings from median income, median rent, median house value, and percent with at least a high school diploma. This component explained 49.1% of the variance and will be referred to as wealth. The second component had high positive loadings from the percent unemployed, percent working class, and percent below the poverty level. This second component explains an additional 31.0% of the variance and will be referred to as poverty. Although these components are similar, wealth at one end of a spectrum and poverty at the other, they independently contribute to explain SES. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### SES Index. Rotated Component Matrix from the principal component analysis of socioeconomic status (SES) variables. ::: Component ---------------------------------------- ----------- -------- Percent Unemployed ^b^ 0.099 0.932 Median Income ^a^ 0.895 -0.294 Median Rent ^a^ 0.923 -0.037 Median House Value ^a^ 0.835 -0.264 Percent Working Class ^b^ -0.405 0.671 Percent with a High School Diploma ^a^ 0.919 -0.081 Percent Below the Poverty Level ^b^ -0.269 0.828 ^a^Represented more in the first component. ^b^Represented more in the second component. ::: Component scores for wealth and poverty were calculated for each tract. The scores were divided into quintiles and used in a Poisson regression to determine their capacity to predict breast cancer incidence. SES variables are scaled so that high wealth scores represent the most wealth and high poverty scores represent the most poverty. Poisson regression ------------------ The initial Poisson regression of the wealth component showed that category 5, the quintile of the most wealth, had a higher associated risk for breast cancer. There was no increasing or decreasing trend when the fifth category was compared to the other four categories, but they were all lower than the fifth, so categories 1 to 4 were collapsed to create a dichotomous wealth variable. A Poisson regression showed that the highest wealth category had an 8.9% increase in incidence over the combined other categories. The analysis also revealed that breast cancer incidence was inversely related to poverty level, as shown in Table [2](#T2){ref-type="table"}. Those census tracts with the highest poverty levels had an incidence rate 36.9% lower than those in the lowest poverty level; those in the second highest poverty level had an incidence rate 32.2% lower than those in the lowest poverty level, with rates of 29.3% and 19.4% lower for those in categories three and two as compared to those in the lowest poverty level. The analysis also showed that incidence is related to urban/rural status with urban tracts having an incidence rate on average 2.7% higher than rural tracts. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Wealth and poverty SES. Relative changes in breast cancer incidence associated with 2 levels of wealth and five levels of poverty. For wealth, category 2 represents the highest level of wealth. For poverty, category 5 represents the highest level of poverty. For example, the women in the highest poverty level, 5, had an incidence rate 36.9% lower than those in the lowest poverty level, 1. ::: Categories Compared \% Change ------------- --------------------- ----------- Wealth SES 1 -- 2 +8.9 Poverty SES 1 -- 5 -36.9 2 -- 5 -32.2 3 -- 5 -29.3 4 -- 5 -19.4 ::: Purely spatial analysis, adjusted for age ----------------------------------------- The purely spatial analysis ignores the time of diagnosis. Figure [1](#F1){ref-type="fig"} summarizes the purely spatial age-adjusted analysis of breast cancer incidence from 1988 to 1997. Fourteen areas have been identified as excessive in their variation, ten significantly higher and four significantly lower than expected under the null hypothesis. The ten areas of high breast cancer incidence are numbered and the four areas of low incidence are lettered in order of significance, with \"1\" and \"A\" having the lowest p-values. For each area of high or low incidence shown in Figure [1](#F1){ref-type="fig"}, the left hand side of Table [3](#T3){ref-type="table"} contains the relative risks (RR), number of observed cases, and p-value for the purely spatial analysis of the age-adjusted cases. Expected case counts can be calculated from Table [3](#T3){ref-type="table"} by dividing the observed count by the RR. The most statistically significant area of excess, High 1, was found west of Boston. It had 3344 observed cases and a relative risk of 1.15, indicating 15% more cases than would have been expected under the null hypothesis. There are two high areas, labeled \"4\" and \"5\", east of High 1. To the west of High 1 is High 7, a single tract with five times more cases than were expected. The second most statistically significant area of excess is High 2 located in the southeast, which includes one tract on Cape Cod, with 2.73 times more cases than expected. High 6 in the northeast shows an elevated risk of 17%. Highs 8 and 9, south and north of High 1, show elevated risks of 15% and 18%, respectively. High 3 in the western part of the state has 2.73 times more cases based on 57 observed and 20.9 expected cases. High 10 in the southeast is elevated by 17%. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Purely spatial, age-adjusted.**Purely spatial analysis results for age-adjusted Massachusetts female invasive breast cancer incidence, 1988--1997. :::  ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Purely spatial analyses. Age-adjusted female invasive breast cancer statistics for the purely spatial analyses, Massachusetts, 1988--1997. ^a^Relative Risk. ^b^Area covers more geographic area than in the age-adjusted analysis. ^c^Area covers more geographic area than in the analysis adjusted for SES. ^d^Area covers less geographic area than in the age-adjusted analysis. ^e^High 11 is a combination of Highs 2 and 10. \*High or low area was not significant for this analysis. ::: Age-adjusted only Adjusted for Urban/Rural Status ------- ------------------- --------------------------------------- ---------- ---------- ------- ---------- Area Observed RR^a^ p-value Observed RR^a^ p-value High 1 3344 1.15 \<0.0001 3344 1.15 \<0.0001 2 94 2.73 \<0.0001 94 2.40 \<0.0001 3 57 2.73 \<0.0001 57 2.72 \<0.0001 4 126 1.75 0.0002 126 1.74 0.0007 5 276 1.44 0.0003 276 1.43 0.0008 6 1258 1.17 0.0008 1258 1.17 0.0013 7 16 5.05 0.0054 16 5.03 0.0062 8 1354 1.15 0.0060 1354 1.15 0.0097 9 874 1.18 0.0163 874 1.18 0.0190 10 941 1.17 0.0228 997 1.17 0.0120 11^e^ \* \* \* \* \* \* Low A 1068 0.74 \<0.0001 1068 0.74 \<0.0001 B 521 0.77 \<0.0001 521 0.76 \<0.0001 C 2276 0.89 \<0.0001 \* \* \* D 588 0.81 0.0012 \* \* \* E \* \* \* 3600 0.91 0.0004 F \* \* \* \* \* \*   Adjusted for SES Adjusted for SES & Urban/Rural Status   High 1 \* \* \* \* \* \* 2 \* \* \* \* \* \* 3 57 3.20 \<0.0001 57 3.21 \<0.0001 4 \* \* \* \* \* \* 5 \* \* \* \* \* \* 6 1418^b^ 1.16 0.0015 1418^b^ 1.15 0.0021 7 16 5.93 0.0013 16 5.93 0.0006 8 \* \* \* \* \* \* 9 \* \* \* \* \* \* 10 \* \* \* \* \* \* 11^e^ 1763 1.21 \<0.0001 1994^c^ 1.21 \<0.0001 Low A 498^d^ 0.73 \<0.0001 498^d^ 0.73 \<0.0001 B 528^b^ 0.82 0.0204 528^b^ 0.82 0.0226 C 2196 0.91 0.0221 \* \* \* D 588 0.83 0.0243 588 0.83 0.255 E \* \* \* \* \* \* F 15 0.36 0.0335 \* \* \* ::: Four areas had significantly fewer cases than expected for the ten year period. Low A had 1068 observed cases and a RR of 0.74, indicating about 26% fewer cases than expected. Low B is west of Cape Cod with 521 observed cases and a RR of 0.77. The other two low areas, \"C\" and \"D,\" in western Massachusetts had 11% and 19% fewer cases than expected, respectively. Purely spatial analysis, adjusted for multiple covariates --------------------------------------------------------- The remainder of Table [3](#T3){ref-type="table"} shows how the results of the purely spatial analysis of age-adjusted incidence rates within tracts are changed by the inclusion of urban/rural status and SES in the model. When urban/rural status is added to the model, low areas \"C\" and \"D\" in rural parts of the state are no longer statistically significant. An area including part of Low D and southward, labeled \"E\" in Table [3](#T3){ref-type="table"} but not shown in the figures, becomes statistically significant with 9% fewer cases than expected. All other previous findings from the age-adjusted purely spatial analysis are unaffected by the adjustment for urban/rural status. When the purely spatial model is adjusted for the two SES variables representing wealth and poverty, seven of the high areas are no longer significant. High areas \"3,\" \"6\" and \"7\" remain significant. High 6 has expanded to include more tracts and more cases but the RR is about the same. High areas \"2\" and \"10\" have merged into a larger area, High 11. Low A covers fewer tracts compared to the age-adjusted analysis, while Low B includes more tracts. A low area, labeled \"F\" in Table [3](#T3){ref-type="table"} but not shown in the figures, covering Nantucket appears; it includes 15 cases, 64% fewer cases than expected. As compared to the analysis adjusting for SES alone, when the two SES variables and urban/rural status are all included in the purely spatial model, the results show little change, except that High 11 is slightly larger in geographic area and low areas \"C\" and \"F\" are no longer significant. Figure [2](#F2){ref-type="fig"} maps the results of the purely spatial analysis adjusted for urban/rural status and the two SES variables. While three of the original ten high areas remain statistically significant, the total geographic area covered by the high areas has been reduced considerably. One of the four original low areas is no longer statistically significant and despite the slight expansion of Low B, the entire geographic area represented by the low areas has also diminished. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Purely spatial, multiple adjustments.**Purely spatial analysis adjusted for socioeconomic status and urban/rural status Massachusetts female invasive breast cancer incidence, 1988--1997. :::  ::: Space-time analysis ------------------- The results of the space-time age-adjusted analysis are somewhat similar to the purely spatial analysis in that seven of the ten areas of excess incidence from the purely spatial analysis were also statistically significantly high in the space-time analysis, along with all four of the low areas. However, as shown in the left side of Table [4](#T4){ref-type="table"}, only two of the high areas, \"4\" and \"5,\" and one of the low areas, Low A, remained statistically significant for the ten-year period. The RRs of these areas were not as elevated in the purely spatial analysis because the space-time analysis captures only the most statistically significantly elevated time periods. However, those areas that were statistically significantly high or low during the whole time period in the space-time analysis had the same RRs in the purely spatial analysis, as they should. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Space-time analyses. Age-adjusted female invasive breast cancer statistics for space-time analyses of Massachusetts, 1988--1997. ^a^Observed count. ^b^Relative Risk. ^c^Area is shifted east compared to the purely spatial analyses. ^d^Area is significant for 1993--1997 and has dramatically increased geographic area. ^e^Area is significant for 1988--1993 only. ^f^Geographic area increased from the age-adjusted analysis. ^g^Geographic area slightly decreased from the age-adjusted analysis. \*High or low area was not significant for this analysis. ::: Age-adjusted only Adjusted for Urban/Rural Status ------ ------------ ------------------- --------------------------------------- ---------- --------- ------- ---------- Time Frame Obs^a^ RR^b^ p-value Obs^a^ RR^b^ p-value High 1 91--97 3639 1.16 \<0.0001 3639 1.16 \<0.0001 3 89--96 51 3.05 \<0.0001 51 3.04 0.0002 4 88--97 126 1.75 0.0078 126 1.74 0.0127 5 88--97 276 1.44 0.0091 276 1.43 0.0154 6 92--95 492 1.35 0.0004 492 1.35 0.0007 7 88--95 \* \* \* \* \* \* 8 92--97 841^c^ 1.23 0.0034 841^c^ 1.23 0.0085 11 93--97 1178 1.28 \<0.0001 1178 1.28 \<0.0001 Low A 88--97 1068 0.74 \<0.0001 1068 0.74 \<0.0001 B 88--94 1119 0.81 \<0.0001 1119 0.81 \<0.0001 C 88--96 1951 0.86 \<0.0001 1951 0.87 0.0002 D 88--95 450 0.77 0.0069 450 0.78 0.0246 Adjusted for SES Adjusted for SES & Urban/Rural Status High 1 91--97 \* \* \* \* \* \* 3 89--96 51 3.58 \<0.0001 51 3.59 \<0.0001 4 88--97 \* \* \* \* \* \* 5 88--97 \* \* \* \* \* \* 6 92--95 492 1.35 0.0008 492 1.34 0.0005 7 88--95 15 6.95 0.0130 15 6.95 0.0102 8 92--97 2991^d^ 1.12 0.0009 2991^d^ 1.12 0.0011 11 93--97 1178 1.29 \<0.0001 1178 1.31 \<0.0001 Low A 88--97 600^e^ 0.75 \<0.0001 600^e^ 0.74 \<0.0001 B 88--94 1219^f^ 0.86 0.0303 1219^f^ 0.86 0.0301 C 88--96 1927^g^ 0.88 0.0134 \* \* \* D 88--95 \* \* \* \* \* \* ::: Figure [3](#F3){ref-type="fig"} displays the results of the space-time age-adjusted analysis. Table [4](#T4){ref-type="table"} shows the time frames, the observed number of cases diagnosed in that time period, the RRs and the p-values. The numbers and letters correspond to approximately the same geographic areas as before. However, while all of the identified areas are still statistically significant at p \< .05, they are not necessarily listed in order of statistical significance in Table [4](#T4){ref-type="table"}. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Space-time analysis, age-adjusted.**Space-time analysis results for age-adjusted Massachusetts female invasive breast cancer incidence, 1988--1997. :::  ::: The two high areas, \"4\" and \"5\", around Boston that were statistically significant in the purely spatial analysis are also statistically significantly high for the entire study period. High 1 was elevated 16% from 1991 to 1997. In northeast Massachusetts, High 6 is statistically significant from 1992 to 1995 with a RR of 1.35. Another high area, High 3, from the purely spatial analysis is also statistically significant in the space-time analysis and represents a single tract in western MA. It is significant from 1989 to 1996 with 51 cases diagnosed and a RR of 3.05. High 9 from the purely spatial analysis was not significant in the age-adjusted space-time analysis. High 11, a large area in the southeast, includes most of High 2 and High 10 from the purely spatial analysis plus some additional tracts that had not been statistically significant before. Low A in Boston is statistically significant for the entire 10-year period. Low B, in southeastern MA, and Low C, in western MA, remain statistically significantly low from 1988 to 1994, and 1988 to 1996, respectively. Low D, east of Low C, is also statistically significantly low from 1988 to 1995. Space-time analysis, adjusted for multiple covariates ----------------------------------------------------- Adjusting the space-time analysis for urban/rural status has little effect, except to change the RR and p-values slightly. When the model includes both the wealth and poverty components as covariates, five areas are found to be excessively high, and one of these, High 7, is new to the space-time analyses. This small area had been identified in the purely spatial analysis and remained statistically significant after adjustment for SES. Low D is no longer statistically significant. When adjusting for SES and urban/rural status together, all the high and low areas from the SES and age-analysis, except for Low C, remain significant. These results are displayed in Figure [4](#F4){ref-type="fig"}. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Space-time analysis, multiple adjustments.**Space-time analysis adjusted for socioeconomic status and urban/rural status Massachusetts female invasive breast cancer incidence 1988--1997. :::  ::: Discussion ========== We examined the possibility of a geocoding bias. Of 46,333 total cases, 4440 were randomly assigned to a census tract located within the reported town. These cases could not be assigned directly to a census tract because their records did not have a residential address. Cancer registries like the MCR collect a patient\'s usual residence at the time of diagnosis. For patients having only postal box addresses, if the MCR staff were unable to obtain a residential address at the time of diagnosis, we assigned these cases to tracts within the town of the mailing addresses, as described in Methods below. This should not be problematic in larger areas of excess or deficit since error would be confined to a town. In the smaller areas identified by the analyses, all had higher than expected case numbers. We examined the smaller areas of excess to determine how many of the observed cases had been randomly placed into a tract. To estimate how the RRs would be affected, the RRs were re-computed from the purely spatial age-adjusted analysis, while excluding these cases. High areas \"2\" and \"3\" did not contain any cases assigned to the tract randomly. In High 4, eight of the 126 cases observed were assigned at random; without these cases, the RR would be 1.63 rather than 1.75. High 5 had 31 of the 276 observed cases assigned randomly; without these cases, the RR would be 1.28 rather than 1.44. Thus, even if none of the randomly assigned cases were assigned to these smaller areas, the relative risks would not be altered substantially. However, MCR has studied High 7 and determined that several large apartment complexes were geocoded as being part of the census tract which makes up High 7 instead of placing it in the correct neighboring tract. Since High 7 does not have many females as part of the population, these geocoding errors may have made a large difference and therefore, this tract is not likely to have high breast cancer incidence. This study adjusted for age, SES, and urban/rural status. Other known risk factors could possibly explain the high areas uncovered. The following attributable risk percentages for such factors have been reported in the literature: 10.7% due to high alcohol intake, 15.0% due to low beta-carotene intake, 8.6% due to low vitamin E intake, 11.6% due to low levels of physical activity \[[@B11]\], 5.0% due to BRCA1 and BRCA2 mutations \[[@B12]\], 29.5% due to late age at first birth and nulliparity, 9.1% due to family history of breast cancer \[[@B13]\], and 2.5% due to smoking \[[@B14]\]. Patient smoking history data were available but were considered unreliable. These other risk factors were not included in this study since the information was not available in the cancer registry. Also, data on town urban/rural status and the SES variables from the 2000 Decennial Census were not available at the time of the analyses, so we were not able to determine if there were any significant changes over time or how these changes might alter the results. For Massachusetts, we would not expect any significant changes. This study used aggregated data to compensate for the lack of individual level data and is therefore exposed to the ecological fallacy. Krieger has shown that this is not a serious problem \[[@B15]\]. Also, it would be impossible to perform this study of breast cancer incidence at the point level since no denominator data exist for population counts at the individual address level. Conclusions =========== The current study is part of the surveillance process, an observational epidemiologic investigation to provide reliable statistical modeling of the raw surveillance data so that program evaluation or planning can be focused on those variations in incidence rates with a low likelihood of being random. In the current study it is especially interesting to observe whether geographical areas, high or low, are affected by SES adjustment or not. It is also interesting to observe which geographical areas seem to be high or low for the entire study period, and which are high or low on a temporary basis. Recommendations --------------- The first step when studying these high and low areas is to look closely at the geocoded data in these areas to ensure that there were not any geocoding-related errors. An example of a geocoding-related error was High 7. However, this problem should only be an issue in areas that only contain a small number of tracts. Roche et al. have made other suggestions as to what can be done when these areas are found to be truly elevated \[[@B16]\]. Several areas (Highs \"1,\" \"4,\" \"5,\" and \"9\" and Low C) are no longer significant with the adjustment for SES and urban/rural status. What is it about the high SES in these areas that affects the RR of these areas? It could be that late age at first birth has a substantial contribution since women of high SES delay childbearing to establish their careers. Note that the p-values and RRs are not correlated and both need to be studied. For ease of comparing statistics, refer to Tables [3](#T3){ref-type="table"} and [4](#T4){ref-type="table"}. High areas \"3\" and \"6,\" as well as Low areas \"A\" and \"B\" do not change a great deal with SES adjustment, so these covariates are not having the same effects as they had above. Perhaps a case-control study could be designed on these areas to determine what other risk factors are driving this elevated incidence rate, such as genetic or behavioral factors. High 6 remained unaffected by adjustment for SES and urban/rural status. However, since it was only temporarily elevated for 4 years, perhaps this area should also be monitored to see if its incidence rate significantly elevates. However, High 11 should be investigated for other factors that may be elevating the risk since it remained elevated for 5 years near the end of the study period. The low areas should be assessed since cases might be going undetected if sufficient screening programs are not in place. If screening programs are in place, they may need to be altered and improved. Finally, it is possible for the populations in two different areas to experience the same amounts of cancer; yet if some cancers remain undetected in one area and are diagnosed in the second area, the second area will appear to have a higher incidence. This is why when breast cancer screening is aggressively promoted in an area with worrisome incidence, the cancer rate goes up for a period of time. Furthermore, not all areas have equal access to the latest in diagnostic equipment capable of detecting the most minute cancers or even pre-cancerous cells. A patient\'s cancer might be diagnosed if she went to a facility with the latest diagnostic equipment, but might not be diagnosed if she had gone to a different place. Incidence may also reflect not just differences in diagnosis rates, but also differences in treatment patterns. If women or physicians in one area tend to opt for more lumpectomies without radiation, the patients still have a chance to develop later cancers in that breast; if women or physicians in another area tend to opt for lumpectomies with radiation or mastectomies for the same stage of disease, they will not be having any subsequent cancers in that breast \[[@B17]\]. Methods ======= Data ---- Cases are from the Massachusetts Cancer Registry (MCR), and include all invasive breast cancer cases diagnosed in female residents between 1988 and 1997 (n = 46,333), using the standard definition of invasive breast cancer\[[@B18]\]. Annual reports on MCR data completeness, methods and other issues can be found at the cited website \[[@B19]\]. Each case record included the town, ZIP Code, and census tract of the patient\'s usual residence at the time of diagnosis, as well as the age at diagnosis, date of diagnosis, race, and stage of disease at diagnosis. Of the 46,333 patients, 43,529 were white females, 1071 black, 6 American Indian, Aleutian, or Eskimo, 115 Chinese, 17 Japanese, 129 other Asian and Pacific Islander, 119 other non-white, and 1347 of unknown race. Race was not included as a covariate in this study due to the large number of unknown. Age was included as a covariate and was divided into fourteen 5-year categories starting with age 15 years up through age 84, with all ages 85 years and above as the fifteenth age group. Aggregation unit ---------------- Census tracts were used to aggregate cases because they are more uniform than towns or Zip Codes in population size, provide a more sensitive analysis of densely populated areas, and are more homogeneous in their resident characteristics. However, 12.6% (n = 5832) of the cases diagnosed in 1988--1997 could not be assigned a reliable residential census tract because of inaccuracies or omissions in the address information provided to the MCR. In most of these cases, a mailing address had been provided and, even after extensive research, MCR staff could not assign a reliable residential address for the patient at the time of diagnosis. To assign the unassigned cases to census tracts, we compared town and census tract boundaries for 1196 four-digit census tracts. About 90 of the census tract codes included 2-digit suffixes to designate 2, 3, 4, or 5 distinct tracts. In this study, only the first four digits of such tract codes were used, thus combining some separate tracts into one. For example, census tracts 6002.01 and 6002.02 were combined into the 4-digit tract 6002. 1990 Census block numbering areas in Massachusetts are treated here as if they were census tracts. Nearly all of the 1196 four-digit 1990 tracts could be related to town boundaries, with one or more tracts located entirely within a town, or one or more towns entirely within a census tract. Two census tracts overlapped parts of multiple towns and for these it was possible to estimate the proportion of the total tract population living in each town. Unassigned cases were then assigned as follows: for towns completely contained in one tract, cases were assigned to that tract, accounting for 1392 cases. For a town containing two or more census tracts, cases were randomly assigned to tracts proportionate to the town\'s female population, accounting for 4440 cases. The allocation of cases should therefore be free from systematic error and any error should be localized to a particular town, while the broader state patterns remain correct. Statistical analyses -------------------- The spatial scan statistic \[[@B20]\] was used to perform purely spatial and space-time analyses. Under the null hypothesis, the incidence of breast cancer follows a Poisson distribution and the probability of a case being diagnosed in a particular location is proportional to the covariate-adjusted population in that location. Tract population data are from the 1990\[[@B21]\] and 2000 Decennial Censuses \[[@B22]\]. This population data is an excellent match to the MCR data since both the Census and the MCR ask patients for their \"usual residence.\" Therefore the numerator and denominator are created in the same way. Properties that make the spatial scan statistic suitable for geographic surveillance are: 1) it takes into account the uneven geographic distribution of cases and population densities, 2) it does not make assumptions about cluster size or location, 3) it adjusts for multiple testing -- a common problem in testing multiple combinations of cluster locations and sizes, 4) it identifies the spatial or space-time locations where the null hypothesis is rejected, and 5) it can detect multiple clusters \[[@B10]\]. Purely spatial analyses were performed first, ignoring time. The maximum spatial cluster size was first set to include up to 50% of the population for both excesses and deficits and then set at 10%, to test for excesses and deficits separately because testing at the 10% level can identify smaller, more defined areas. However, each area identified in an analysis had a likelihood associated with it that was compared to the 9,999 likelihoods from the initial 50% maximum spatial size test, thus accounting the inference for a multiplicity of statistical tests. Space-time analyses were then performed to determine whether the clusters from the purely spatial analysis were long term or temporary. The maximum temporal cluster size was set at 90% and also included purely spatial clusters with a temporal size of 100% for all space-time analyses. This allows for the entire time period all the way down to the smallest amount of time to be considered as the time window, ten years down to one year. All SaTScan analyses were performed on age-adjusted population counts. To calculate the age-adjusted expected counts, the 1990 and 2000 female population counts were combined into a weighted average of the two based on the years being analyzed, 1988 through 1997. This was done for each age group within each tract. The natural log of this weighted average was entered as the offset variable in a Poisson regression in SAS \[[@B23]\]. There were a few tracts with a zero population for a certain age group; for these, a population of one was entered so a log could be taken. The Poisson regression included age as a class variable and the number of cases within each tract and age group as the dependent variable to calculate the age-adjusted expected counts. The expected counts were aggregated across age for all SaTScan analyses. This method was verified by entering age as a covariate in a SaTScan analysis with the same results. Although SaTScan accurately estimates age-adjusted incidence rates, when other covariates enter the model, SaTScan computes the interaction of each of the 15 age categories with each covariate. In the current study, that would mean estimating 14 × 1 × 1 × 4 = 56 interaction terms because three risk factors were used as covariates. By first computing the age-adjusted counts in SAS, the number of interactions estimated by SaTScan is reduced to 8. After identifying statistically significant geographic areas and determining whether they were long term or temporary, the next step was to determine whether these areas would change when the model was adjusted for known risk factors. Socioeconomic factors have long been identified as risk factors for breast cancer \[[@B9],[@B24],[@B25]\]. Urban/rural status has also been used as a covariate when studying breast cancer \[[@B4],[@B26]-[@B29]\]. An SES index was created following the method of Yost et al \[[@B30]\]. Yost and colleagues included the following variables from the U.S. Decennial Census in a principal component analysis (PCA): percent unemployed, percent working class, percent below the Federal poverty level, median income, median rent, median house value, and percent with at least a high school diploma. We used the same variables from the 1990 Census \[[@B21]\] in a PCA, described in the results section. A PCA factors the matrix of correlations between all pairs of variables into seven uncorrelated components with the hope that a smaller number of these components will account for 70 to 80% of the variation contained in the matrix of correlations among the original seven variables. Instead of using seven highly correlated SES variables as covariates, PCA can reduce the number of covariates to one or two without loss of information. A PCA was also performed on a similar group of variable from the 1990 Census based on what Krieger \[[@B24]\] used to create a SES index, however it did not account for as much of the variance as the variables used by Yost \[[@B30]\]. A variable indicating the urban or rural status of each tract was derived from data available on the Massachusetts Institute for Social and Economic Research (MISER) \[[@B31]\]. An urban area consists of one or more contiguous census blocks with a population density of at least 1,000 persons per square mile, while all other areas are designated as rural \[[@B32]\]. We assigned all tracts entirely within a town the same classification as that town, regardless of their individual populations. Tracts that included several towns were classified as rural since all towns within such tracts were also classified as rural. The capacity of SES and urban/rural status to predict incident cases within census tracts was tested in Poisson regression models, using the GENMOD procedure in SAS \[[@B23]\]. SaTScan \[[@B33]\] created files listing census tracts and the number of the high and low areas they belonged to. These files were brought into Maptitude \[[@B34]\] to create the figures. Authors\' contributions ======================= TJS: PI, responsible for design, funding, of project with overall responsibility for implementing the project, including the final paper. LMD: Principal data analysis, responsible for final checks on accuracy of data and all analyses, including their written interpretation. MK: Co-PI, assisted in the overall design of the project and close collaborator on all analyses. MK also assisted a great deal in editorial decisions of drafts and of the final paper. DG: Co-PI, assisted in the design of the project, including funding and implementation. DG also consulted on editorial and technical issues. SG: Director of MCR, she was responsible for the overall integrity of the data, and ensuring that use of the data corresponded to the policies of the Massachusetts Dept of Public Health. She also assisted with editorial help on all drafts. MM is responsible for the geocoding of the data in the MCR. She was a constant help to this project as we tried multiple approaches to our analyses. MM was also very helpful in clarifying limitations of the work and also provided many editorial suggestions. All authors read and approved the final manuscript. Acknowledgements ================ The study was supported by a grant from the National Cancer Institute, CA81763-02. We also acknowledge many substantive and editorial improvements suggested by the three anonymous reviewers working through the Massachusetts Department of Health.

PubMed Central

2024-06-05T03:55:47.235661

2004-8-3

PMC514717

Background ========== Livers of pig are used in a wide range of experimental settings related to problems encountered in human medicine such as fulminant hepatic failure \[[@B1],[@B2]\] and graft survival in liver transplantation \[[@B3]\]. Bioartificial liver support systems, which combine living cells of the liver in an extracorporeal circuit, have been successfully used in first clinical trials \[[@B4]\]. The shortage of human organs to be used for bioreactors and the lack of safe and effective human liver cell lines have resulted in pigs becoming an important hepatic cell source. Although much work based on the use of pig liver has been carried out in the past, no studies have so far focused on isolated cell populations of the liver. To overcome this lack of knowledge, we recently published a study on high yield isolation and characterization of pig liver cells \[[@B5]\]. This work represents a basis for *in vitro*studies of the cells constituting the hepatic reticulo-endothelial system: the Kupffer cells (KC) and the liver sinusoidal endothelial cells (LSEC) \[[@B6]\]. Together these cells represent the most powerful scavenger system of the body. Most soluble or insoluble macromolecular harmful material entering the body from the gut is transported to the liver via the portal vein and eliminated by uptake in KC and LSEC. Studies in rats have revealed that LSEC are actively engaged in blood clearance of an array of waste macromolecules of both soluble and colloidal nature. Major components of connective tissue as well as other potentially hazardous products are eliminated rapidly and almost exclusively from the blood by receptor-mediated endocytosis in LSEC \[[@B7],[@B8]\]. The endocytic receptors of rat LSEC can be listed in five categories: (i) collagen α-chain receptor (COLLAR), (ii) hyaluronan receptor (HAR), (iii) IgG Fc receptor, (iv) mannose receptor, and (v) scavenger receptor (SR). With specificity for free α-chains of types I, II, III, IV, V, XI collagens, but not native triple helical collagen, the COLLAR function has so far only been described in LSEC \[[@B9],[@B10]\]. The HAR is responsible for the clearance of circulating hyaluran (HA) \[[@B11],[@B12]\]. Chondroitin sulphate, dermatan sulphate, as well as some ligands for the SR are also endocytosed via this receptor, which has been purified and characterized \[[@B13],[@B14]\]. In both KC and LSEC the IgG Fc receptor has been shown to mediate the elimination of circulating IgG immune complexes in the liver \[[@B15]-[@B18]\]. The mannose receptor recognizes the terminal non-reducing sugar residue of the oligosaccharide (fucose, mannose or N-acetyl-glucosamine) moiety of glycoproteins \[[@B19],[@B20]\]. First found in alveolar macrophages \[[@B21]\], this receptor was later functionally demonstrated on LSEC as well \[[@B22]\], where it performs an extremely rapid endocytosis of mannosylated glycoproteins \[[@B23]\]. Sharing the feature of recognizing negatively charged ligands, the SR is a rapidly growing family of receptors that have been identified in macrophages and other cell types \[[@B24],[@B25]\]. An array of physiological and foreign ligands has been reported to be eliminated from the circulation predominantly via endocytosis in the rat LSEC SR \[[@B8],[@B26]-[@B29]\], using the receptors described above. Recent observations have resulted in new insight into the LSEC SRs and a relationship between SR and the HAR. First, it was published that a rat LSEC surface protein has affinity for both HA and SR-ligands; Antibodies generated against this protein inhibited endocytosis of both HA and various SR-ligands \[[@B13]\]. Therefore, the name hyaluronan/scavenger receptor (HA/S-R) was coined to describe its activity \[[@B14],[@B30]\], but it is also known as stabilin-2. It was recently reported that LSEC in knock-out mice lacking SR class A, remove SR-ligands both *in vivo*and *in vitro*as avidly as do wild-type mice \[[@B31]\]. These results show that LSEC, employing this unique receptor, represents the most important cellular site of uptake of blood-borne SR-ligands. We have previously shown that pig LSEC express a functionally active receptor that both *in vivo*and *in vitro*clear a SR ligand \[[@B5]\]. Until now this represents the only functional study that has been done on pig LSEC. Accordingly, in the present study we wanted to study if pig LSEC carry out the same avid clearance based on the use of the same endocytosis mechanisms that has been already shown for the rat. We also compared the technique of isolation of pig LSEC with the established method used to prepare isolated rat LSEC \[[@B32]\]. Apart from the difference in organ size, the most striking difference from rat LSEC isolation was the high number of stellate cells within the pig non-parenchymal cells (NPC) which hampered the purification of the pig LSEC. In addition to represent a functional study of pig LSEC, this study also describes a method for obtaining a high number of purified pig LSEC which is necessary for the further purification and characterisation of molecules unique for LSEC. We here show that pig LSEC, but not stellate cells, parenchymal cells (PC) or KC, express the same five categories of endocytosis receptors for soluble and colloidal ligands as reported for rat LSEC. Ligands endocytosed via these receptors are transported along the endocytic pathway and degraded. Assays carried out to measure activity of lysosomal enzymes in cells showed that several enzymes are present at higher specific activities in LSEC than in either KC or PC. Results ======= Isolation and cultivation of liver cells ---------------------------------------- The fraction of cells collected after density centrifugation in OptiPrep consisted of LSEC, stellate cells, KC, PC and a minor proportion of unidentified cells, possibly of epithelial nature. In this fraction, LSEC were most numerous, closely followed by stellate cells. Further purification, using either elutriation centrifugation or selective adherence, resulted in 80--95 % purity of LSEC. One cycle of elutriation centrifugation resulted in approximately 10^8^LSEC. Contaminants in LSEC cultures were epithelial-like cells, stellate cells, KC, and PC. Compared with rat and mice PC, the population of PC from pig are much more heterogeneous in size; consequently, pig LSEC cultures may be contaminated by small PC. Cultures of KC prepared by selective adherence were prepared by incubating the non-PC fraction, for 15 min, on glass dishes coated with glutaraldehyde-treated BSA This procedure yielded 70--80 % pure KC. The use of tissue culture plastic dishes resulted in a lower purity of KC (40--50%), with a higher proportion of LSEC. Specificity of endocytosis via different receptors in cultured LSEC ------------------------------------------------------------------- Experiments were carried out to establish whether ligands known to be taken up by rat LSEC via the different categories of receptors are endocytosed with the same specificities in pig LSEC. To this end, trace amounts of ^125^I-labelled AGG, FSA, collagen α-chain (COLLA), mannan and FITC-bHA were incubated with purified cultures of pig LSEC with or without the presence of excess amounts of unlabelled substances known to compete with the radiolabelled ligands for endocytosis in rat LSEC. Control experiments showed that endocytosis was substantial following 2 h incubation at 37°C with only radiolabelled ligands. Expressed as % endocytosis of total added ligand, the results were as follows: AGG, 7.1%; FSA, 42.7%; COLLA, 59.7%; mannan, 38.2%; FITC-bHA, 12.1% (Fig. [1](#F1){ref-type="fig"}). Approximately 40--50% of the endocytosed COLLA and FSA and 30% of endocytosed AGG were recovered as acid-soluble degradation products in the medium after 2 h of incubation. No acid soluble radioactive degradation products of ^125^I-mannan or ^125^I-FITC-bHA were released to the medium. Competition experiments showed that AGG (300 μg/ml) inhibited uptake of trace amounts of ^125^I-AGG by 73% (Fig. [1A](#F1){ref-type="fig"}). AGG at a concentration of 100 μg/ml was only 40% inhibitory (results not shown). FSA (100 μg/ml) inhibited uptake of trace amounts of ^125^I-FSA and ^125^I-FITC-bHA by 95% and 14%, respectively (Figs. [1B](#F1){ref-type="fig"} and [1E](#F1){ref-type="fig"}); COLLA (100 μg/ml) inhibited uptake of trace amounts of ^125^I-COLLA by 84% (Fig. [1C](#F1){ref-type="fig"}); mannose (50 mM) inhibited uptake of trace amounts of ^125^I-mannan by 96% (Fig. [1D](#F1){ref-type="fig"}); and hyaluronan (100 μg/ml) inhibited uptake of trace amounts of ^125^I-FITC-bHA by 62% (Fig. [1E](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Receptor specificity.**Specificity of uptake of ^125^I-AGG (A), ^125^I-FSA (B), ^125^I-collagen (C), ^125^I-mannan (D), or ^125^I-FITC-bHA (E) in LSEC in the presence of mannose (50 mM), FSA (0.1 mg/ml), COLLA (0.1 mg/ml), hyaluronan (0.1 mg/ml) and AGG (0.3 mg/ml). Uptake (cell-associated radioactivity (solid bars) plus acid soluble radioactivity in spent medium \[open bars\]) was measured after 120 min incubation at 37°C. Results, given as % of control, are means of three experiments, each consisting of three parallels. Error bars represent standard deviation (+SD). :::  ::: Endocytosis of fluorochrome-labelled ligands -------------------------------------------- To study the transport from early endosomes to later endocytic compartments, late endosomes and lysosomes were prelabelled with TRITC by administering TRITC-FSA to pigs via the jugular vein, 1.5 h prior to preparation of cultures \[[@B33]\]. Following incubation for 6.5 h at 37°C to allow the LSEC to adapt to the *in vitro*conditions, cultures were pulsed for 1 h at 4°C with fluorochrome-labelled ligands, and chased after an additional 10 min or 2 h incubation at 37°C. Due to weak intensity, FITC-bHA needed to be pulsed for 20 min at 37°C in cells that were not prelabelled with TRITC-FSA, and then chased for 20 min or 2 h to obtain a detectable image of the uptake. Observation of these cultures in the fluorescence microscope revealed that following a 10 min chase, FITC-labelled FSA (Fig. [2D](#F2){ref-type="fig"}), COLLA (Fig. [2F](#F2){ref-type="fig"}) and mannan (Fig. [2H](#F2){ref-type="fig"}) were distributed in small (green) vesicles scattered throughout the cytoplasm. Co-localization of these green vesicles with perinuclear organelles prelabelled with red TRITC-FSA could not be observed at this early chase period. FITC-bHA chased for 20 min was also found in similar small vesicles and also in larger vesicles spread throughout the cell (Fig [2J](#F2){ref-type="fig"}). Furthermore, following 10 min chase, most FITC-labelled AGG and TRITC-FSA co-localized (yellow color indicates co-localization) in large ring shaped vesicles throughout the cell, but some vesicles with only FITC-AGG were observed (Fig. [2A](#F2){ref-type="fig"}). After 2 h, the co-localization with TRITC-FSA was significant for FSA (Fig. [2E](#F2){ref-type="fig"}), COLLA (Fig. [2G](#F2){ref-type="fig"}), and mannan (Fig. [2I](#F2){ref-type="fig"}), whereas FITC-bHA was found in vesicles similar to those observed after 20 min (Fig. [2K](#F2){ref-type="fig"}). After this chase-period FITC-AGG and TRITC-FSA, only partly co-localized as seen in a single cell (Fig. [2B](#F2){ref-type="fig"}). FITC-AGG was still present in large ring shaped vesicles, of which many still were spread throughout the cell, whereas TRITC-FSA was found in small red vesicles but also in some big ring shaped vesicles together with FITC-AGG. Interestingly, control cells chased for 2 h with only TRITC-FSA (Fig. [2C](#F2){ref-type="fig"}) had all ligand concentrated in the perinuclear region in lysosomes, and no ring shaped vesicles were observed. This indicates that FITC-AGG influences the transport of TRITC-FSA to the lysosomes. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Intracellular transport of endocytosed ligands.**Cultures of LSEC were pulsed with both TRITC-FSA and FITC-AGG for 1 h at 4°C. Chasing was performed after removal of unbound ligand by washing, and transferring of the cultures to 37°C. The cultures were fixed after chase periods of 10 min or 2 h and examined in fluorescence microscope. At 10 min (A) all TRITC-FSA co-localized with FITC-AGG (yellow colour indicates co-localization) in large ring-shaped vesicles (large arrowheads), and some vesicles with only FITC-AGG were observed (small arrowheads). After 2 h (B), co-localization of FITC-AGG and TRITC-FSA in large vesicles (arrows) was observed together with big vesicles with only FITC-AGG (large arrowheads) and small vesicles with only TRITC-FSA (small arrowheads). Controls show a more perinuclear appearance of TRITC-FSA when pulsed and chased for 2 h alone (C). In other experiments, TRITC-FSA was injected intravenously 1.5 h before isolation of the cells. Following an additional 6.5 h of cultivation at 37°C cultures of LSEC were pulsed with FITC-FSA (D-E), FITC-collagen (F-G), or FITC-mannan (H-I) for 1 h at 4°C. Following a 10 min chase, the FITC-ligands were observed to appear in small vesicles (arrowheads), and did not co-localize with TRITC-FSA (D, F and H). After 2 h, the FITC-ligands were transported to perinuclear compartments that co-localized almost completely with TRITC-FSA (small arrows in E, G and I). Other cultures of LSEC were pulsed for 10 min at 37°C with FITC-bHA. Following a 20 min chase, the FITC-bHA was observed in vesicles distributed all over the cell (arrowheads in J), and a similar appearance was observed after 2 h (arrowheads in K). Occasionally, cells that did not take up TRITC-FSA *in vivo*, but endocytosed FITC-ligands *in vitro*(big arrow in I), were observed. Scale bars: 10 μm. :::  ::: The FITC-ligands were observed only in cells that endocytosed TRITC-FSA *in vivo*. In spite of the fact that nearly all cells judged as LSEC in the cultures accumulated FITC-ligands *in vitro*, approximately 5% of the cells had not taken up TRITC-FSA *in vivo*. Lysosomal enzyme activities --------------------------- By using sensitive fluorometric assays, we measured the activities of six different lysosomal hydrolases, and compared their specific activities in three major liver cell types LSEC, KC and PC. The results, listed in Table [2](#T2){ref-type="table"}, reveal that all the enzymes measured were present in significantly higher specific activities in LSEC as compared to PC, with LSEC:PC ratios as high as 7.5, 6.8, 4.9, 3.3 and 2.3 for hexosaminidase, glucuronidase, aryl sulphatase, acid lipase and acid phosphatase, respectively. The specific activities of α-mannosidase and glucuronidase in LSEC were significantly higher also when compared to KC. Compared to PC, all the enzymes except for acid lipase were present in significantly higher activities in KC. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Specific activities of lysosomal enzymes in parenchymal (PC), sinusoidal endothelial cells (LSEC) and Kupffer cells (KC) isolated from pig liver. ::: **LSEC** **KC** **PC** -------------------------- --------------- ---------------- --------------- α-Mannosidase (= 7) 0.76^a^(0.14) 0.59^b^(0,16) 0.40^c^(0.08) Hexosaminidase (n = 7) 30.3^a^(7.86) 29.2^a^(1.70) 4.02^b^(2.06) Acid lipase (n = 6) 6.66^a^(2.40) 4.33^ab^(2.60) 2.03^b^(0.46) Acid phosphatase (n = 6) 11.2^a^(3.90) 8.89^ab^(2.72) 4.97^b^(1.96) Glucuronidase (n = 5) 3.55^a^(0.77) 2.35^b^(0.53) 0.52^c^(0.28) Aryl sulphatase (n = 4) 0.27^a^(0.11) 0.20^a^(0.05) 0.05^b^(0.02) Each experiment consisted of three parallels. Values are means (SD). The letters (a, b, and c) indicate significant statistical differences between the values of the different cell types. Enzyme activities are expressed as 10^6^4-methylumbelliferyl (-mannopyranoside, -N-acetyl-glucosaminide, -oleate, -phosphate, -glucuronide, -sulfate) molecules released per min per g cell solubilisate at 37°C. (p \< 0.05, ANOVA, LSD post hoc test). ::: Discussion ========== In spite of the fact that most researchers in the field would assume that pig LSEC perform the same important scavenger function as rat LSEC \[[@B34]\], no studies have been published so far to actually prove this supposition. The present study was undertaken to establish whether LSEC from pig liver have the same high scavenger capacity as has been found for LSEC from rat liver \[[@B7]\]. To this end pig liver LSEC were isolated and cultivated, and studied with respect to endocytic and degradative activity. Since LSEC make up just a few percent of the total liver volume \[[@B35]\], it is important to employ a cell marker that readily and specifically distinguishes LSEC from other types of liver cells. We found that soluble FITC- and TRITC-FSA serve this function. Although other fluorescently marked probes can be used for the same specific distinction of LSEC, FITC-FSA stands out as the optimal marker, since it is inexpensive, easy to prepare and very stable. To label LSEC *in vivo*, 20 mg FITC/TRITC-FSA in 20 ml physiological saline was administered via the left external jugular vein 90 min prior to perfusion of the liver with collagenase. The selective adherence steps used for the separation of KC from the NPC fraction in mice (10 min on uncoated tissue culture plastic dishes) \[[@B31]\] and rats (30 min on tissue culture plastic dishes coated with glutaraldehyde-fixed BSA) \[[@B36]\] are not useful for separation of KC from pig NPC. Glass dishes coated with GA-fixed BSA were found to be the best substrate for selective adherence of KC. After KC depletion by selective adherence on this substrate, the non-adherent NPC fraction was transferred to fibronectin-coated plastic culture dishes, incubated for 15 min to allow attachment of LSEC, and washed extensively. This procedure resulted in 90% pure LSEC. Longer incubation time or too weak washing were found to yield a higher relative number of contaminating stellate cells and small PC. Furthermore, in contrast to liver cell isolation from rat and mice, it is not feasible to remove all PC from the pig NPC suspension by neither isopycnic- or elutriation centrifugation due to the large heterogenity in both density and size of the cells. Heterogenity in density was also observed in LSEC, since the cells were found in all tested OptiPrep-layers, with densities varying from 1.038--1.086 g/ml. Moreover, LSEC were elutriated at all flow rates varying from 20--50 ml/min, indicating heterogeneity in cell size. For mass isolation of LSEC by elutriation centrifugation, we obtained purities between 80--95%, very similar to purities obtained by selective adherence. Taken together, LSEC purified by selective adherence is faster as long as a limited number of cells are needed. Elutriation centrifugation is more time-consuming, but yields a higher cell number, with up to 1.1 × 10^9^purified LSEC from one pig. The addition of the detergent Pluronic acid F-68 in the elutriation buffer eliminated clotting of cells in the elutriation chamber \[[@B37]\], thereby allowing a higher number of cells to be loaded per cycle. The ligands COLLA, FITC-bHA, mannan, FSA, and AGG were used to probe the endocytic activity in pig LSEC via the COLLAR, HAR, SR, mannose receptor and IgG Fc receptors, respectively. We found that all these ligands were avidly endocytosed in LSEC. Competition experiments showed that the ligands were taken up in a specific manner, via the five different categories of receptors. Moreover, morphologic pulse-chase experiments using FITC-labelled ligands to study the intracellular transport of endocytosed ligand suggested that all ligands studied, with the exception of AGG, reached lysosomal compartments within a time span of 2 h. At that time, AGG was still found in ring shaped structures which is a typical feature of early and late endosomes \[[@B33],[@B38]\]. The finding that endocytosed AGG was degraded (albeit not as efficiently as other ligands), without reaching the lysosomal compartment, indicates that this ligand is processed differently than the other ligands studied. This phenomenon has also been observed by Løvdal et al. \[[@B15]\] who noted that rat LSEC and KC *in vitro*degraded endocytosed IgG-complexed antigen much slower than ligands for the mannose- and scavenger receptor, and that the delay was due to slow departure from early endosomes. We also observed that FITC-AGG delayed the transport of TRITC-FSA to the lysosomes when the ligands were given simultaneously. This is consistent with the observation that AGG reduces the amount of ^125^I-FSA degraded even if the total amount of endocytosed FSA was not changed. Interestingly, we observed no uptake of FITC-AGG in KC. FITC-labelled ligands were seen concentrated in small spherical vesicles scattered over the entire cell body after a 10 min chase. These small vesicles are probably early endosomes reminiscent of small bristle coated vesicles that have been previously observed electron microscopically as vesicles with a diameter of 180 nm, located directly below the cell surface \[[@B39]\]. The vesicles containing FITC-bHA after a 10 min pulse at 37°C followed by a 20 min chase, are probably late endosomes (with diameter ranging between 800--1500 nm) as reported in similar studies in rat LSEC \[[@B38],[@B39]\]. After a 2 h chase, the stain partly co-localized with TRITC-FSA, indicating further transport to late endosomes and lysosomes. FITC-bHA was observed in small perinuclear vesicles after a 2 h chase, and almost 30% of the endocytosed ^125^I-FITC-bHA was found as low molecular weight material, demonstrating intracellular degradation (results not shown). We stress that, although practically all cultured LSEC accumulated FITC-labelled ligands *in vitro*, not all LSEC had taken up TRITC-FSA *in vivo*. We speculate that the explanation for the heterogeneous uptake *in vivo*was due to circulatory regulation: not all sinusoids may have allowed entrance of blood at the time of injection, thus preventing LSEC from being exposed to the injected ligand in those sinusoids. This is not an unreasonable explanation, since it is known that hepatic sinusoids may regulate blood flow by a sphincter mechanism \[[@B40]\]. If LSEC are an important part of the reticulo-endothelial system in the body they also need a high activity of lysosomal enzymes to degrade waste material endocytosed from the circulation. Therefore, we compared the lysosomal activity in LSEC with the metabolically very active PC and the phagocytic KC. Earlier studies in rat have revealed that KC and LSEC, as compared to PC, contain higher specific lysosomal enzyme activities \[[@B41]-[@B43]\], and we found that all enzymes measured were present in considerably higher specific activities in LSEC than in PC. The specific activities of α-mannosidase and glucuronidase in LSEC were also higher than in KC. The high specific activities of lysosomal enzymes in LSEC are compatible with the notion that these cells are true professional scavenger cells. Moreover, studies in rat have shown that LSEC may recruit lysosomal enzymes from the circulation by endocytosis via the mannose receptor \[[@B44],[@B45]\]. This may partly explain the very high specific activity of such enzymes in these cells in the pig as well. In all our experiments we used freshly isolated cells because we believe that when cells have just been isolated from the intact organ, their *in vitro*scavenger properties resemble their *in vivo*properties. Once outside their micro-environment in the liver, rat LSEC dedifferentiate and eventually die after 2--4 days on culture dishes. During the first hours in culture only small changes in endocytic capacity of LSEC occur in so far as experiments in our laboratory have shown that pig LSEC cultivated in RPMI for 2 days retain 80% of the endocytic capacity when compared to freshly isolated cells (results not shown). Because of the short cultivation time used in our experiments, the endocytic capacity of the cells would not be influenced by mediators in the medium. But in other experiments with cultures of rat LSEC that were incubated for 18 h or more, it has been shown that inflammatory mediators like tumor necrosis factor-α and interleukin-1β enhance 2--3-fold endocytosis via the SR and mannose receptor, while COLLAR mediated endocytosis remained unaffected \[[@B46]\]. Also lipopolysaccharide can increase endocytosis in LSEC by stimulating the cells to release autocrine interleukin-1β. Another mediator, the nitric oxide, decreases endocytosis via the mannose receptor in rat LSEC \[[@B47]\]. Using interleukin-10, Knolle et al. \[[@B48]\] found a similar effect as that reported with nitric oxide, namely down-regulation of mannose receptor mediated endocytosis in mouse LSEC. Other potentially mediators of endocytic capacity are VEGF \[[@B49],[@B50]\] and phorbol ester \[[@B51]\] which at least have been shown to improve maintenance of rat LSEC in culture. Conclusion ========== Our results suggest that pig LSEC are functionally very similar to rat LSEC, at least in terms of clearance activity. It is therefore highly likely that LSEC in pig (as in rat) represent the major site of elimination of an array of soluble waste molecules from the circulation. Several if not all of these waste macromolecules are harmful if allowed to accumulate in the blood. The very active endocytosis in LSEC ensures that these soluble waste macromolecules are never allowed to increase above trace levels in the circulation. Methods ======= Chemicals and animals --------------------- 1,3,4,6-tetrachloro-3α,6α-diphenylglycoluril (Iodogen), carrier-free Na^125^I, and TRITC were from Pierce, Rockford, IL, USA, Institute for Energiteknikk, Norway and ICN Biomedicals Inc., OH, USA. FITC, BSA, 4-methylumbelliferyl-substrates for fluorometric assays of lysosomal enzymes, Triton X-100, mannose, and mannan was from Sigma Chemical Co, St.Louis, MO, USA. Collagenase P was from Boehringer Mannheim, Germany. Fibronectin was kindly donated by Dr. B. Hansen, University of Tromsø, Norway. Human IgG and high molecular weight hyaluronan (Healon) were from Pharmacia, Sweden. OptiPrep was from Nycomed, Norway. Human serum albumin was from Octapharma, Ziegelbrucke, Switzerland. Monoclonal mouse anti-human desmin, clone D33, was from DAKO A/S, Denmark. Monoclonal goat anti-mouse IgG, TRITC-conjugate, was from Zymed, CA USA. Two mouse monoclonal antibodies (clones 2G6 and 2B10) against porcine macrophages were kindly provided by Dr. A. Berndt, Institute of Pathology, Friedrich Schiller University, Jena, Germany. Castrated male piglets (*Sus scrofa domesticus*, Norwegian strain), weighing 7--8 kg, were fasted for 18 h, drinking water *ad libitum*, prior to sacrifice. Animals received care according to \"Guide for the Care and Use of Laboratory Animals\" prepared by the National Academy of Sciences and published by the National Institutes of Health, NIH publication 86-23 revised 1985. Isolation and cultivation of liver cells ---------------------------------------- The procedure for isolation of functionally intact LSEC from a single pig liver was as described \[[@B5]\]. Briefly, the liver was perfused with a physiological saline buffer to wash out blood cells before perfusion with a collagenase buffer to disperse the liver cells. The resulting single cell suspension was subjected to 2 × 3 min velocity centrifugation (50 *g*) to pellet PC, and the resulting supernatant was concentrated by centrifugation (850 *g*) for 10 min, mixed into a 21% OptiPrep density solution, and centrifuged for 30 min (3300 *g*). The PC and RBC were pelleted, and a layer consisting of 6 × 10^8^-3.5 × 10^9^non-PC were recovered. For further purification, 2.0 × 10^8^non-PC in HBBS solution containing 0.3 % BSA, 0.4 mM EDTA, 200 μg/ml Pluronic acid F-68, and antibiotics were introduced into a standard elutriation chamber of a JE-6B rotor (Beckman Instruments) in a J-21-type Beckman centrifuge at a flow rate of 22 ml/min and a rotor speed of 2500 rpm. The first fraction of 100 ml was enriched in stellate cells as judged by strong staining with anti-desmin antibody. The flow rate was then increased to 35 ml/min, and 150 ml was collected. This fraction contained purified LSEC characterized by their specific accumulation of FITC-labelled formaldehyde-treated serum albumin (FSA). The cells remaining in the chamber were pumped out and discharged after the centrifuge was stopped. These were mainly KC and PC. KC were identified by the immunostaining with two anti-porcine macrophage antibodies. The average numbers of LSEC grown per cm^2^were 2.5 × 10^5^in Falcon dishes (Becton Dickinson, France). An alternative separation technique was also used: the cell suspension of non-PC after the gradient centrifugation was diluted to 4 millions/ml and seeded on glass dishes at a concentration of 5 × 10^5^/cm^2^and allowed to attach for 15 min, at 37°C. Prior to use, the glass dishes were washed in 96% ethanol, coated with bovine serum albumin and fixed in 1% glutaraldehyde for 30 min., before being extensively rinsed in distilled water. It was mainly KC that attached well to these dishes. Poorly attached cells that were detached by gentle washing together with non-adherent cells were transferred to fibronectin-coated culture dishes, incubated for another 15 min followed by thorough washing, and supplied with fresh medium to enable attachment and spreading of viable SEC. The purity of these LSEC cultures was between 80--95%. Ligands ------- FSA was prepared by treating BSA with 10% formaldehyde in 0.2 M carbonate buffer, pH 10, for 3 days as described \[[@B52]\]. Aggregated gamma-globulin (AGG) was prepared by heating purified human IgG (10 mg/ml) for 30 min at 63°C \[[@B53]\]. Insoluble AGG was removed by centrifugation for 30 min (3300 *g*). Native triple helical collagen ((Nutacon, Leimunden, The Netherlands) was denatured to single collagen α-chains (COLLA) by incubation at 60°C for 60 min. Biotinylated hyaluronan, bHA, was prepared by incubating HA with Biotin-LC-Hydrazide (Pierce, Rockford, IL, USA) and 1-ethyl-3-\[3-dimethylaminopropyl\] carbodiimide hydrochloride (Sigma Chemical) in a ratio allowing a maximum of 1 out of 10 carboxyl groups per HA molecule to be labelled \[[@B54]\]. Labelling of ligands with fluorochromes --------------------------------------- FSA, AGG, collagen, bHA and mannan dissolved in sodium carbonate buffer (0.1 M, pH 9.5) were incubated with FITC or TRITC in a ligand/dye weight ratio of 5:1, at 4°C overnight. To remove unbound dye, the solutions were dialysed against PBS. Radiolabelling procedures ------------------------- Macromolecular ligands (FSA, mannan, collagen, AGG and FITC-bHA) in PBS were labelled with Na^125^I employing Iodogen as the oxidizing agent \[[@B55]\]. Radiolabelled proteins and free iodine were separated by gel filtration on a PD-10 column (prepacked Sephadex G-25, Pharmacia, Uppsala, Sweden). The resulting specific radioactivity was 1--3 × 10^6^cpm/μg protein. To radiolabel bHA, FITC was first attached to the biotin, thereby providing an ^125^I acceptor. For separation of free iodine from radiolabelled FITC-bHA, the reaction solution was dialysed against PBS, giving a final radioactivity of 0,3 × 10^6^cpm/μg HA. Receptor specificity of ^125^I-labelled ligands ----------------------------------------------- After seeding and cultivation for 2--3 h in 24 well dishes, purified cultures of LSEC were washed and supplied with fresh RPMI 1640 medium containing 1% human serum albumin and trace amounts of one of the six ^125^I-ligands (10,000--30,000 cpm per culture) and excess cold ligands. All endocytosis experiments were terminated after an incubation-period of 2 h at 37°C, by transferring the conditioned medium (200 μl), along with 500 μl PBS used for washing of the cells, to tubes containing 500 μl 20% trichloroacetic acid. Following centrifugation of the tubes, the extent of degradation was determined by measuring the radioactivity in the pellets and the supernatants, except for mannan or FITC-bHA where the trichloroacetic acid-precipitation step was omitted due to lack of degradation products being released from the cells. The ^125^I was attached to the protein core in mannan and to the FITC adduct in FITC-bHA of which both accumulate in the lysosomes, since mammalian cells do not carry degradative hydrolases for these molecules. Cell-associated radioactivity was determined by measuring the amount of ^125^I released by solubilizing washed cultures in 1% SDS. All experiments were carried out in triplicate. Accumulation of fluorochrome-labelled ligands --------------------------------------------- TRITC-FSA (20 mg) in 20 ml physiological saline was administered via the left external jugular vein 90 min prior to liver perfusion. Prior to use, TRITC-FSA was centrifuged at high speed and sterile-filtered in order to remove aggregates which would otherwise be taken up in KC by phagocytosis. Cultures of LSEC were established on fibronectin-coated 14 mm diameter glass coverslips for 3--4 h in serum-free growth medium. The cultures were then washed in PBS, and pulsed in fresh medium with 0.1 mg/ml FITC-ligands for 1 h at 4°C. Because FITC-labelling of proteins remove positive charges, and thus may turn proteins into negatively charged ligands for the SR \[[@B56]\], FITC-mannan and FITC-collagen were pulsed in the presence of 0.5 mg/ml FSA to avoid binding to SR. Endocytosis of FITC-AGG was studied in cells that had not been prelabelled with TRITC-FSA. Instead, the cells were pulsed in fresh medium with both 0.1 mg/ml FITC-AGG and TRITC-FSA for 1 h, at 4°C. After removal of unbound ligand by washing with PBS, bound ligands were chased for 10 min and 120 min in fresh pre-warmed medium at 37°C. As the only fluorochrome in the cells, FITC-bHA (0.2 mg/ml) was pulsed for 10 min at 37°C before medium change, and then chased for 20 min and 2 h. Incubations were terminated by fixation in 4% formaldehyde, and the specimens examined in a Zeiss Axioplan fluorescence microscope. Micrographs were taken with a Nikon Coolpix 4500 digital camera. Assay of lysosomal enzymes -------------------------- Samples of PC were taken from the pellet resulting from the density centrifugation, and solubilized in 0.1 % Triton, whereas KC and LSEC samples were obtained from the solubilizates of cultures seeded on 7.5 cm^2^Falcon culture dishes. The assay conditions of the six enzymes are given in Table [1](#T1){ref-type="table"}. In all the enzyme assays, except for acid lipase (see below) aliquots of 100 μl substrate were incubated at 37°C with the appropriate amount of cell-solubilizates, and the appropriate length of time after which 2.0 ml of 0.5 M glycine/sodium hydroxide buffer pH 10.4 were added to stop the reaction and to develop the fluorochrome. To assay for acid lipase 10 μl of substrate was used, and 2.0 ml of 0.5 M Tris buffer pH 8.5 was added to stop the reaction. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Conditions of enzyme assays. ::: **Enzyme** **Conc. (mM)** **Substrate** **Buffer** **pH** **Incubation time (min)** **Solubilisate volume (μl)** ------------------ ---------------- ------------------------------------------------- --------------------------- -------- --------------------------- ------------------------------ α-Mannosidase 2.5 4-methylumbelliferyl-α-D-mannopyranoside Phosphate/citrate (0.1 M) 4.0 120 40 Hexosaminidase 5.0 4-methylumbelliferyl-N-acetyl-β-D-glucosaminide Phosphate/citrate (0.1 M) 4.5 30 5 10 (PC) Glucuronidase 2.5 4-methylumbelliferyl-β-D-glucuronide Acetate (0.1 M) 4.5 120 20 5 (LSEC) Acid phosphatase 1.0 4-methylumbelliferyl-phosphate Acetate (0.1 M) 4.5 30 10 Aryl sulphatase 10.0 4-methylumbelliferyl-sulfate Acetate (0.5 M) 5.5 120 40 Acid lipase 0.3 4-methylumbelliferyl-oleate Acetate (0.1 M)\* 4.0 60 20 \*Acetate (0.1 M) + 0.1% Triton X-100 + phosphatidylcholine (100 μM) + taurodeoxycholic acid (300 μM). ::: The fluorescence of 4-methylumbelliferone, resulting from the action of the enzymes on the various substrates, was measured in a Shimadzu RF 5000 spectrofluorometer with excitation set at 360 nm and emission at 450 nm. Enzyme activities are given in number of substrate molecules transformed min^-1^·gram protein^-1^under the conditions stated above. Protein content in the samples was measured according to Lowry \[[@B57]\], and BSA in 0.1 % Triton was used as standard. Statistics ---------- The values are expressed as: mean (standard deviation) unless otherwise noted. We used SPSS 10.0 software package (SPSS, Chicago, IL) for statistical analysis. Analysis of variance (ANOVA) was used to test whether any statistical significance existed between the three cell populations enzyme activity followed by the LSD post hoc comparison test. Probability values of p ≤ 0.05 were considered significant for all tests applied. Authors\' contributions ======================= KHE and GIN designed and carried out the experiments. KHE drafted the manuscript. AR and BS coordinated the study and contributed to the text of the manuscript. Acknowledgements ================ The excellent technical assistance by Hege Hagerup and Marna-Lill Kjæreng is highly appreciated. This work was supported by the Norwegian Research Council.

PubMed Central

2024-06-05T03:55:47.240308

2004-8-12

PMC514718

Background ========== Primary cardiac tumors are rare disorders, with an incidence of 0.02% at autopsy. Three quarters of the primary tumors of the heart are benign, half of which are myxomas \[[@B1]\]. As noninvasive cardiac imaging becomes widely available, with increasing resolution provided by echocardiography, computed tomography and magnetic resonance imaging, cardiac tumors are being diagnosed more often. Angiography, apart from its preoperative role to rule out concomitant coronary artery disease, is rarely needed for the diagnostic work-up of cardiac tumors. This report describes the delayed presentation of a left atrial myxoma which was not depicted in an initial coronary angiography performed 51 months earlier in a woman with chest pain. Case Report =========== A 62-year-old woman with known metabolic syndrome was referred to our clinic to exclude coronary artery disease invasively. She has been experiencing chest pain for four months which has not increased in frequency or duration since it started. She denied pain at rest, nocturnal pain, difficulty breathing, or palpitations. An echocardiographic stress examination revealed significant ischemia with anteroseptal hypokinesia. Cardiac chambers were morphologically normal. The \"baseline\" echocardiography and stress echo was unsuspicious of a left atrial myxoma (Fig. [1](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Stress echocardiography: long axis view (baseline). No sign of myxoma in left chamber or in left atrium. LA = left atrium, LV = left ventricle, Ao = Ascending aorta (with kind permission of Dr. Herbst, Potsdam) :::  ::: Coronary angiography revealed coronary artery disease with a stenosis of the proximal left anterior descending coronary artery (LAD) near the left main stem. A stent was implanted in the proximal LAD with no residual stenosis. The ventricle was morphologically normal. A coronary angiography performed 13 months after stent implantation showed no re-stenosis and a normal left ventricle. No echocardiogram was performed at this time 32 months after the control coronary angiography, the patient was readmitted because of increasing dyspnoea and palpitations. A transthoracic echocardiography disclosed a big (70 × 30 mm) mass in the left atrium attached to the interatrial septum. The tumor prolapsed into the left ventricle obstructing the mitral valve orifice (Fig [2](#F2){ref-type="fig"}). The mean pressure gradient across the mitral valve was 8 mm Hg. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Echocardiography (45 months) long axis view. Myxoma in the left atrium prolapsing into the left ventricle. LV = left ventricle, My = myxoma :::  ::: A subsequent coronary angiography and LV and RV catheterization detected a mean diastolic pressure gradient of 12 mm Hg between the pulmonary capillary wedge and the left ventricular enddiastolic pressure, and no re-stenosis of the LAD stent. The angiography was notable for a large area with small atypical, tortuous vessels in the region of the interatrial septum. These vessels were shown to originate from branches of the right coronary artery (RCA) and the circumflex coronary artery (RCX). (Fig. [5](#F5){ref-type="fig"}) ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Right coronary artery (RCA) in 60 degree LAO position (45 months, pre-operative coronary angiography). White arrow = atypical vessels in the interatrial septum :::  ::: Surgery was promptly performed and the tumor was successfully excised. Histology confirmed the diagnosis of a cardiac myxoma. A retrospective analysis of the initial coronary angiographies (baseline, 13 and 45 months) disclosed the atypical vessels in a small area of interatrial septum (Fig. [3](#F3){ref-type="fig"}, [4](#F4){ref-type="fig"}, [5](#F5){ref-type="fig"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Right coronary artery (baseline) in 90 degree LAO projection. White arrow = atypical vessels in the interatrial septum :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Left coronary artery (13 months) in 45 degree LAO and 30 degree CRAN projection. White arrow = atypical vessels in the interatrial septum :::  ::: Discussion ========== Myxomas are benign but potentially dangerous with disturbance of rhythm, peripheral embolization or mechanical valvular obstruction, of the atrial or ventricular cavity \[[@B2],[@B3]\]. The site, mobility, and size of the myxoma determine the clinical course. Some authors found no correlation between the size of the tumor and the clinical picture \[[@B4]\], others reported symptoms with left atrial myxomas weighting more than 70 g \[[@B5]\]. The rate of growth of myxomas is not exactly known \[[@B5],[@B6]\]. An increase in size of 1.8 -- 5.8 cm/year and in weight of up to 14 g/year was reported \[[@B7],[@B8]\]. The myxoma of our patient reached 70 × 30 mm before being symptomatic. Myxomas presenting with systemic embolism or intracavitary obstruction can be easily detected non-invasively. The early depiction of small intracardiac tumors by means of angiography relies on the detection of atypical vessels supplied by branches of the left or right coronary artery. Our case demonstrates that this early angiographic sign is difficult to find. However, vascular malformations are not pathognomonic for myxomas \[[@B9]\] and stromal tumors, such as myxomas, have generally poor blood supply, hence a coronary angiographic finding with a neo-vascularized or highly vascularized intracardiac area is more suggestive for other type of cardiac neoplasm. A cluster of small, tortuous and dilated vessels is also seen in old and organised thrombi, haemangiomas and venous malfomations \[[@B10]\]. The size of malformation has no relation to the size of the tumor \[[@B11]\]. It has been reported in the literature that myxomas can be induced by radiation \[[@B12],[@B13]\]. Between the first presentation to our clinic and the diagnosis of myxoma, the patient underwent two coronary angiograms including stenting, with a cumulative radiological exposure of about 30 mSv (corresponding to at least 1500 chest X-rays). To our knowledge, the patient did not undergo any other relevant radiation exposure in the past. In the present case, a retrospective analysis of the patient angiographies disclosed the atypical vessels which were initially overseen. These vessels could have probably been interpreted as an early sign of myxoma. Author\'s contributions ======================= HPD and FK have written the manuscript and have equally contributed to this publication. HPD, VG and WR have performed the coronary angiographies. WK has performed cardiac surgery. HPD, ACB, FK and GB participated in the design and coordination of the final manuscript. All authors have read and approved the final manuscript.

PubMed Central

2024-06-05T03:55:47.243282

2004-8-13

PMC514719

Background ========== Anti-ischemic therapy, in particular beta-blockers, is the most commonly employed drug for the control of myocardial ischemia in patients with stable coronary artery disease. Its widespread use also in patients with suspected coronary artery disease has important practical, clinical diagnostic and prognostic implications because diagnostic tests are heavily influenced by its effects. The diagnostic and prognostic impact of anti-ischemic therapy on stress testing is largely ignored but not negligible. The issue raises several questions: How to evaluate patients at time of testing for myocardial ischemia? How to interpret a stress test performed on anti-ischemic therapy? Are the stressors employed for the detection of myocardial ischemia created equal in relation to the different classes of drugs used in clinical practice? Is stress testing able to assess the efficacy of medical therapy in patients with known coronary artery disease? Has the protection of anti-ischemic therapy on inducible myocardial ischemia any impact on long-term survival? Pathophysiologic implications of anti-ischemic therapy during stress testing ============================================================================ The answer to all these issues relies on the mechanism through which myocardial ischemia is induced by the different stressors (exercise or pharmacologic such as dipyridamole and dobutamine) employed during stress testing. Test exploring organic coronary artery stenosis can induce ischemia by two basic mechanisms: 1. an increase in oxygen demand, exceeding the fixed supply and 2. flow maldistribution due to inappropriate coronary arteriolar triggered by a metabolic/pharmacologic stimulus \[[@B1]\]. The mechanism of increased demand can be easily fitted into the familiar concept framework of ischemia as a supply-demand mismatch, deriving from an increase in oxygen requirements in the presence of a fixed reduction in coronary flow reserve. The different stresses can determine increases in demand through different mechanisms (Fig. [1](#F1){ref-type="fig"}). In resting conditions, myocardial oxygen consumption is dependent mainly upon heart rate, inotropic state, and the left ventricular wall stress (which is proportional to the systolic blood pressure) \[[@B2]\]. Following dipyridamole or adenosine administration, a slight increase in myocardial function, a modest decrease in blood pressure, and mild tachycardia can be observed, overall determining only a trivial increase in myocardial oxygen demand \[[@B3]\]. During exercise, the increase in heart rate, blood pressure, and inotropic state accounts for the overall increase in myocardial oxygen consumption \[[@B4]\]. Pacing and dobutamine also increase -- to a lesser degree -- myocardial oxygen demand \[[@B5]\]. During pacing, the increase is mainly due to the increased heart rate. Dobutamine markedly increases contractility and heart rate. Further augment in myocardial oxygen consumption for heart rate increase occurs with the co-administration of atropine with dobutamine \[[@B6]\]. and dipyridamole \[[@B7]\]. (Fig. [2](#F2){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Major determinants of myocardial oxygen consumption in resting conditions (left) and during stress commonly employed with echocardiography. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Conceptual allocation of tests employed in combination with echocardiography to detect coronary artery disease stenosis inducing ischemia via steal effect (left) or increased myocardial oxygen demand (right), or both mechanisms. :::  ::: In the presence of coronary atherosclerosis, appropriate arteriolar dilation can paradoxically exert detrimental effects on regional myocardial perfusion, causing overperfusion of myocardial layers or regions already well perfused in resting conditions at the expense of regions or layers with a precarious flow balance in resting conditions \[[@B8]\]. Anti-ischemic therapy can interfere with all the above mentioned mechanisms of ischemia induction, although in a very different fashion. The mechanism of action of anti-ischemic drugs is easily fitted within the familiar framework of supply-demand mismatch. In particular, beta-blockers are credited with reducing exercise-induced ischemia by decreasing myocardial oxygen demand and possibly by increasing supply through a reduction in extravascular forces \[[@B9]\]. Experimental studies demonstrated that beta-blockers reduce dipyridamole-induced ischemia by an alteration of regional myocardial blood flow \[[@B10]\]. However, this straightforward explanation seems inadequate in justifying the protective effects of beta-blockers on dipyridamole-induced ischemia. Experimental data show that beta-blockers do not affect the dipyiridamole-induced increase in flow \[[@B12]\]. On the other hand, the increase in myocardial oxygen consumption does not play any significant role in the induction of dipyridamole-induced ischemia, which is due to an absolute reduction in subendocardial flow (tightly linked to regional wall thickening) mostly for \"vertical\" and \"horizontal\" steal phenomena. However, experimental studies on the model of the exercising dog have shown that beta-blockers protect myocardium from stress-induced myocardial blood flow-function relation: for a given transmural flow, there is a rise of subendocardial and a fall of subepicardial flow, with an improved regional performance \[[@B10]\]. This same mechanism has also been documented with some calcium antagonists, such as diltiazem and may explain, in part, the beneficial effects of this class of drugs on dipyridamole-induced ischemia. Calcium antagonists can effectively prevent ischemia provoked by dipyridamole also through other mechanisms, which they share with nitrates, and they tend to increase the coronary flow supply during stress. In this case, the prevention of steal phenomena may be due to the increase in collateral flow (which has been shown with nitrates and, to a much lesser extent, with some calcium antagonists) \[[@B12]\]. and to the dilation of epicardial coronary lumen size. The pronounced increase in collateral flow can prevent horizontal steal phenomena due to dipyridamole, wheras even a small increase of the coronary diameter can dramatically reduce the blood pressure drop across the stenosis, therefore preventing vertical steal phenomena. Beta-blockers exert a direct and competitive action on beta-1 receptors, as they are employed as specific antagonists of dobutamine-induced ischemia. Dobutamine, through its beta-1 agonist action determines the increase in oxygen consumption, but it induces flow maldistribution through beta-2 arteriolar receptors. Diagnostic implications of the use of anti-ischemic therapy during stress testing ================================================================================= On the basis of these premises and taking into consideration that the markers of inducible myocardial ischemia (electrocardiogram, perfusion, wall motion) are very different therefore expressing a different sensitivity to the action of anti-ischemic therapy at time of testing, it is clear that medical therapy affects test results (see table [1](#T1){ref-type="table"}. In fact, the AHA/ACC Guidelines on Chronic stable angina state that \[[@B13]\]. \"whenever possible, it is recommended that beta-blockers (and other anti-ischemic drugs) be withheld for four to five half-lives (usually about 48 h) before exercise imaging studies for the diagnosis and initial risk stratification of patients with suspected CAD\". Ideally, these drugs should be withdrawn gradually to avoid a withdrawal phenomenon that may precipitate events. When beta-blockers cannot be stopped, stress testing may detect myocardial ischemia less reliably, but it usually will still be positive in patients at the highest risk. The same recommendations apply to imaging stress testing. Nonetheless, in patients who exercise to a submaximal level because of the effect of drugs, perfusion or echocardiographic imaging still affords higher sensitivity than the exercise ECG alone \[[@B14]\]. On the basis of these recommendations patients undergoing a stress testing for diagnostic purposes should be evaluated off therapy not to offset test results. Exercise imaging stress (nuclear perfusion or ultrasound) testing have a lower sensitivity when performed on anti-ischemic therapy \[[@B14]-[@B19]\]. due to the limited increase in heart rate and blood pressure which determine oxygen consumption. Antianginal therapy lowers the sensitivity of exercise echocardiography as it does with vasodilator stress testing \[[@B19],[@B20]\]. Antianginal therapy with beta-blockers, calcium-antagonists or nitrates in various combinations prevent dipyridamole-induced ischemia by delaying the appearance of the transient dyssynergy \[[@B21]\]. this variation on dipyridamole time parallels variations in exercise time at exercise stress testing \[[@B21]\]. (Fig. [3](#F3){ref-type="fig"}). Dipyridamole stress sensitivity was 91% off therapy and fell to 65% under therapy in various combinations (beta-blockers and/or calcium antagonists and/or nitrates). The same reduction of dipyridamole test sensitivity is obtained with monotherapy with beta-blockers at time of testing (100% off therapy vs. 38% on therapy) \[[@B22],[@B23]\]. (Fig. [4](#F4){ref-type="fig"}). Interestingly, the positive effects of beta-blockers on dipyridamole stress are largely independent of the effect on heart rate, possibly involving a direct anti-steal effect \[[@B21]-[@B23]\]. Angiotensin-converting enzyme inhibitors have no effect on dipyridamole stress echocardiography results \[[@B24]\]. The sensitivity of dobutamine is heavily affected by concomitant beta-blocker therapy. Beta-blockers effect a rightward shift in the dose-response curve to dobutamine and sharply lower test sensitivity, unless atropine is used \[[@B25]\]. (Fig. [5](#F5){ref-type="fig"}). Calcium antagonists and/or nitrates only mildly reduce dobutamine stress sensitivity (100% off therapy vs.88% on therapy, p=ns) (Fig. [6](#F6){ref-type="fig"}). Non-beta-blocker antianginal therapy reduces the severity of dobutamine-induced ischemia by reducing the value of peak wall motion score index and time of ischemia appearance. However, these changes are not correlated to variations in exercise tolerance \[[@B26]\]. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Correlation between the therapy-induced variations in dipyridamole and exercise time in the 38 patients with positivity of both tests off treatment (Modified from 21). :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Dipyridamole test sensitivity on and off beta-blocking therapy (Modified from 23). Test sensitivity is significantly reduced in patients studied on beta-blocking therapy. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Heart rate during dobutamine-atropine stress testing on and off beta-blockers (Modified from 25). :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### Dobutamine-atropine stress echocardiography test sensitivity in patients on and off non-beta blocking therapy (Modified from 26). :::  ::: Dipyridamole stress nuclear imaging techniques do not seem to be influenced by anti-ischemic therapy \[[@B27]\]. However, it has been recently demonstrated that acute administration of beta-blockers in patients with known coronary artery disease reduces dipyridamole SPECT sensitivity from 69% with placebo to 52% (p = 0.039) with 10 or 20 mg of metoprolol in a per-vessel analysis, but not overall sensitivity \[[@B28]\]. The reason for this difference between stress echocardiography and nuclear imaging is likely to be related to the different markers of ischemia, i.e. wall motion abnormalities vs. perfusion, in the face of the same pathophysiologic mechanism of ischemia: the reduction of coronary reserve. In fact, in the presence of a coronary stenosis, during stress, flow remains elevated in the subepicardial layer but falls in the subepicardium. This selective stress-induced hypoperfusion is important for stress echocardiography, since the regional systolic thickening is linearly and closely related to subendocardial perfusion and only loosely related to subepicardial perfusion \[[@B29],[@B30]\]. (Fig. [7](#F7){ref-type="fig"}). ::: {#F7 .fig} Figure 7 ::: {.caption} ###### Schematic illustration of the principle underlying the impact of antianginal therapy on different markers of ischemia: regional function and perfusion imaging. At rest, perfusion is homogeneously distributed between endocardial and epicardial layers. In the presence of a significant coronary stenosis, vasodilation induced by pharmacologic stress, provokes a subendocardial underperfusion with a relative epicardial overperfusion which is translated into an impairment of function (echocardiographic dyssynergy) and perfusion (reversible defect at scintigraphy). Medical therapy at time of testing re-equilibrates the imbalance between subendocardial and subepicardial layers, but it affects only function. :::  ::: ::: {#F8 .fig} Figure 8 ::: {.caption} ###### Kaplan-Meier survival curves (considering total mortality as an endpoint) in patients stratified according to presence (DET +) or absence (DET -) of myocardial ischemia at pharmacological stress echocardiography on and off antianginal medical therapy. The best survival is observed in patients with no inducible ischemia off therapy; the worst survival in patients with inducible ischemia on therapy (Positive DET vs. Negative DET off antianginal medical therapy, p \< 0.000; Positive DET vs. Negative DET on antianginal medical therapy, p \< 0.074) (Modified from 31). :::  ::: Prognostic implications of anti-ischemic therapy during stress testing ====================================================================== The protective effect of anti-ischemic therapy on inducible myocardial ischemia might exert a powerful impact on prognosis. From the EPIC-EDIC Data bank, it has been analyzed the prognostic impact of antianginal therapy at time of testing in 7333 patients with suspected or known coronary artery disease undergoing pharmacologic stress echocardiography with either dipyridamole or dobutamine. The results show that a positive test on medical therapy is an additional marker of ischemia severity at stress testing whereas a negative test on medical therapy is less prognostically benign, being a false negative result \[[@B31]\]. (Fig. [8](#F8){ref-type="fig"}). No prognostic difference was found among the various forms of anti-ischemic drugs at time of testing, but the presence per se of antinaginal therapy at time of testing is an independent predictor of death. It is worth noting that in the study only a very low percentage of patients was taking beta-blockers: if on one side this aspect represent a clear lack of adherence to recommendations \[[@B13]\]., on the other it is an observed pattern of prescription in our data base, which simply reflected the clinical practice and the lack of a universally accepted policy of testing regarding concomitant therapy \[[@B32],[@B33]\]. Marwick et al. \[[@B34]\]. have demonstrated a protective effect on mortality of beta-blocker therapy in patients with a negative exercise echocardiography whereas specificity and negative predictive value is increased for the prediction of cardiac events (cardiac death, myocardial infarction and unstable angina) during exercise scintigraphy in patients evaluated off medical therapy at time of testing \[[@B35]\]. The clinical implications of these results are far-reaching. Inducible myocardial ischemia during pharmacological stress testing on medical therapy identifies the subset of patients at highest risk of death. On these patients an aggressive approach has to be undertaken in order to change the natural history of coronary artery disease. On the far opposite end the incidence of death in patients with a negative pharmacologic test off therapy is so low that no intervention could lower the spontaneous rate of death any further. At intermediate risk are those patients with a negative test on medical therapy or with a positive test off medical therapy. Different clinical scenarios can be foreseen on the basis of the present results: 1) A negative test on medical therapy might represent a false negative result, therefore it is advisable to repeat the test off therapy in order to assess the real ischemic burden through the conventional stress echocardiographic parameters \[[@B36],[@B37]\]. -- i.e. number of ischemic segments, severity of induced dysfunction, (both expressed by peak wall motion score index), pharmacologic load and time of onset of ischemia. This is in line with the recommendations of the American Heart Association in patients with stable angina \[[@B13]\].; 2) In the case of a positive test off medical therapy, the effect of therapy can be assessed with the advantage of using an objective, primary ischemic end point such as changes in wall motion during stress. Conclusions =========== Patients may be undergoing various forms of antianginal therapy at the time of testing, both an advantage and a disadvantage for stress echocardiography testing. The disadvantage is that antianginal therapy reduces sensitivity, since stress-induced wall motion abnormalities are caused by the development of obligatory myocardial ischemia. The advantage is that the effect of therapy can be assessed using an objective, primary ischemic end-point such as changes in stress-induced wall motion abnormalities. The presence of ischemia can be titrated on the basis of the ischemic-free stress time and the extent and severity of the induced dyssynergy. The various forms of stress are differently affected by various forms of therapy. In patients with known or suspected coronary artery disease, ongoing anti-ischemic therapy at the time of testing heavily modulates the prognostic value of pharmacological stress echo. In presence of concomitant anti-ischemic therapy, a positive test is more prognostically malignant and a negative test less prognostically benign. However, the decision to remove a patient from beta-blocker therapy for stress testing should be made on an individual basis and should be done carefully to avoid a potential hemodynamic rebound effect, which can lead to accelerated angina or hypertension \[[@B38]\]. No major side effects were recognized when medical therapy was withdrawn in large series of consecutive patients undergoing pharmacologic stress echocardiography \[[@B39]\]. In practical terms, when a test is performed for diagnostic purposes it should be done off medical therapy in order to avoid the influence of medical therapy (in case of hypertensive patients it is possible to prescribe ACE-inhibitors or Angiotensin II receptor blockers that do not exert any protective effect on myocardial ischemia). In patients with known coronary artery disease the decision to suspend medical therapy should be taken on an individual basis in view also of the fact that pharmacologic stress echocardiography is a versatile tool that can assess medical therapy efficacy in the long term prognosis \[[@B31]\]. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Effects of oral therapy on stress testing sensitivity ::: **STRESS** -------------------------- ------------ ---- ---- Beta-blockers ↓ ↓ ↓↓ Calcium channel blockers ↓ ↓ ↓↔ Nitrates ↓ ↓ ↓↔ ACE-inhibitors ↔ ↔ ↔ Aminophylline ↓↔ ↓↓ ↔ ACE, angiotensin-converting enzyme; ↓, decreased sensitivity; ↓↓ markedly decreased sensitivity; ↔, no effect on sensitivity; ↓↔, mild decrease in sensitivity. :::

PubMed Central

2024-06-05T03:55:47.244136

2004-8-20

PMC514720

Review ====== Dendritic cells (DC) are professional antigen presenting cells that induce primary antigen specific T cell responses \[[@B1]\] and exhibit all functional properties required to present exogenous antigen (Ag) to immunologically naïve T cells. These properties include: a) uptake of exogenous Ag via receptor-mediated endocytoses, b) processing of complex proteins into antigenic peptides, c) assembly of these peptides with MHC molecules, d) surface expression of MHC molecules as well as costimulatory molecules, including CD80, CD86, and CD40, e) secretion of T cell stimulatory cytokines, including IL-1β, IL-6, IL-8, TNF-α, and macrophage inflammatory protein (MIP)-1α and f) migration into draining lymph nodes \[[@B2]\]. In the present study, we sought to characterize the Ag processing capacity of DC, as well as the enzymes previously involved in this process. In this regard, several groups have previously reported that epidermal LC and splenic DC, both of which contain small numbers of non-DC contaminants, exhibit significant Ag processing capacities \[[@B3]-[@B12]\]. LC freshly obtained from skin are quite potent in their Ag processing capacity, but the majority of these LC lose this capacity as they \"mature\" during subsequent culture \[[@B3]-[@B6],[@B12]\]. On the other hand, other reports have shown that DC are less efficient than macrophages in Ag processing, with each employing different pathways for Ag processing \[[@B10],[@B13]-[@B16]\]. These differences suggest the possibility of unique pathways and requirements for Ag presentation by DC. With respect to the mechanisms by which DC process complex protein Ags, chloroquine has been shown to inhibit this process; this suggests that Ag processing primarily occurs within acidic compartments \[[@B6]-[@B8]\], \[[@B10]-[@B12]\]. Macrophages and B cells have been reported to employ cathepsins B, D, and/or E for digesting protein Ag, including ovalbumin (OVA), hen egg white lysozyme (HEL), myoglobin, exogenous IgG, and *Staphylococcus aureus*nuclease \[[@B17]-[@B35]\]. These proteases may each exhibit differential pathways for activity; for example, macrophages appear to employ cathepsin D for the initial cleavage of myoglobin and cathepsin B for C-terminal trimming of resulting fragments \[[@B17]\]. Little information, however, has been available with respect to proteases that are employed by DC for Ag processing. Thus, in the present study we sought to define the protease profiles produced by DC and then to identify which protease(s) would primarily mediate Ag processing in DC. Materials and Methods ===================== Cells ----- The XS52 DC cell line (a gift of Dr. Takashima, Dallas, Texas), a long-term DC line established from the epidermis of newborn BALB/c mice \[[@B23]\], were expanded in complete RPMI in the presence of 1 ng/ml murine rGM-CSF and 10% culture supernatants collected from the NS stromal cell line as described previously \[[@B23]\]. Other phenotypic and functional features of this line are descibed elsewhere \[[@B23]-[@B25]\]. As responding T cells, we used the protein purified derivative (PPD)-reactive Th1 clone LNC.2F1 and the Th2 clone LNC.4K1 \[[@B26]\], both of which were kindly provided by Dr. E. Schmitt (Institute for Immunology, Mainz, Germany). As control cells, we also employed Pam 212 keratinocytes \[[@B27]\], 7--17 dendritic epidermal T cells (DETC) \[[@B28]\], J774 macrophages (ATCC, Rockville, MD), and BW5147 thymoma cells (ATCC). Splenic DC were purified from BALB/c mice (Jackson Laboratories, Bar Harbor, ME) by a series of magnetic bead separations as before \[[@B24],[@B25]\]. Briefly, spleen cell suspensions were first depleted of B cells using Dynabeads conjugated with anti-mouse IgG. Subsequently, T cells were removed using beads coated with anti-CD4 (GK1.5) and anti-CD8 mAbs (3.155), and then macrophages were depleted using beads conjugated with F4/80 mAb. Finally, DC were positively sorted using beads coated with anti-DC mAb 4F7 \[[@B29]\]. The resulting preparations routinely contained \> 95% DC, as assessed by flow cytometry. DCs were propagated from bone marrow as described by Inaba et al. \[[@B30]\]. Using magnetic beads, bone marrow cell suspensions were first depleted of B cells (with anti-mouse IgG), I-A^+^cells (with 2G9 mAb, Pharmingen, San Diego, CA), and T cells (with GK1.5 and 3.155 mAbs). The remaining I-A^-^cells were then cultured in the presence of GM-CSF (10 ng/ml). The purity of bone marrow derived DC was more than 95% as determined by flow cytometry using anti-CD11c and anti-I-A antibody (not shown). Determination of protease activities ------------------------------------ Cells were lysed in 0.1% Triton X-100 in 0.9% NaCl; extracts were then examined for protease activities using the following substrates: a) Z-Arg-Arg-βNA (for cathepsin B, at pH 6.0), b) denatured hemoglobin (cathepsin D/E, pH 3.0), c) Arg-βNA (cathepsin H, pH 6.8), d) Z-Phe-Arg-MCA (cathepsin J, pH 7.5), e) Z-Phe-Arg-MCA (cathepsin L, pH 5.5), f) Gly-Phe-βNA (DPPI or cathepsin C, pH 5.5), g) BLT ester (BLT esterase, pH 7.5), and h) Suc-Ala-Ala-Pro-Phe-SBz and Suc-Phe-Leu-Phe-SBz (chymotrypsin-like proteases, pH 7.5). Samples were incubated at the indicated pH and enzymatic activities were assessed by colorimetric or fluorogenic changes \[[@B31]\]. Enzymatic activities were expressed as nmol/min/mg soluble protein, in which protein concentrations were measured by the bicinchoninic acid method using bovine serum albumin as a standard \[[@B32]\]. Ag presentation and T cell stimulation assays --------------------------------------------- XS52 DC were γ-irradiated (2000 rad) and then pulsed for 8 hr with 100 μg/ml of PPD (kindly provided by Dr. E. Schmitt, Mainz, Germany) in the presence of each of the following inhibitors (or vehicle controls): a) pepstatin A (100 μg/ml, Sigma, St. Louis, MO), b) DCI 100 μM, Sigma), c) E-64 (100 μM, Sigma), d) DMSO (1%), and e) NH~4~CL (15 mM). Subsequently, the XS52 cells were washed 3 times with PBS to remove unbound PPD and then cultured in 96 round-bottom well-plates (10^4^cells/well) with either the PPD-reactive Th1 or Th2 clone (10^5^cells/well) in the presence of the same inhibitor at the above concentration. In some experiments, XS52 DC were pulsed overnight with PPD in the presence of an inhibitor and then fixed with 0.05% glutaraldehyde in PBS for 30 seconds at 4°C; the fixation reaction was stopped by adding 0.1 M L-lysine. These XS52 cells were then washed with PBS and examined for their ability to activate Th1 or Th2 clones in the absence of protease inhibitors. In order to determine the mechanism of action for pepstatin A, XS52 cells were pulsed in its presence with PPD either in a native form or following digestion with trypsin-conjugated sepharose beads (Pierce, Rockford, IL) for 15 minutes at 37°C. We also examined the effect of added pepstatin A on the capacity of XS52 cells to activated allogeneic T cells isolated from CBA mice (Jackson Laboratories). Samples were pulsed for 18 hr with 1 μCi of ^3^H-thymidine and then harvested using an automated cell harvestor. RT-PCR Analysis --------------- mRNA expression for cathepsin D was examined by RT-PCR. RNA isolation, reverse-transcription, and cDNA amplification were carried out as previously described \[[@B33]\]. The following primers were designed based on the published sequence of murine cathepsin D \[[@B34]\]: 5\'-GGTCAGAGCAGGTTTCTGGG-3\' and 5\'-GCTTTAAGCTTTGCTCTCTTCGGG-3\'. After 25 cycles of amplification, PCR products were analyzed in 1% agarose gel electrophoresis containing 2 μg/ml ethidium bromide. Other experimental conditions, including primer sequences for the β-actin control, are described elsewhere \[[@B33]\]. Results ======= DC exhibit several different protease activities and they process the complex protein Ag PPD into antigenic peptides -------------------------------------------------------------------------------------------------------------------- In the first set of experiments we sought to identify which protease activities were expressed by DC. A panel of synthetic peptide and protein substrates was incubated with extracts prepared from three DC populations: the XS52 DC line, 4F7^+^splenic DC, and GM-CSF-propagated bone marrow DC. As noted in Table [1](#T1){ref-type="table"}, each DC population exhibited all tested protease activities, including cathepsins B, C, D/E, H, J, and L, BLT esterase, and chymotrypsin-like proteases. Each protease activity in DC was substantially higher (up to 20 fold) than that detected in the BW5147 thymoma cell line, a line that expresses relatively low levels of protease activities. Moreover, cathepsin D/E activity was undetectable (\<1 nmol/min/mg) in Pam 212 keratinocytes and 7--17 DETC (data not shown), indicating further cell type-specificity. These results demonstrate that DC produce a variety of protease activities and at relatively high levels. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Protease Profiles Expressed by Several DC Populations ::: **Protease** **XS52 DC** **Splenic DC** **Bone Marrow DC** **BW5147 Thymoma Cells** ---------------------- -------------------- ------------------ -------------------- -------------------------- Cathepsin B^1^ 133 ± 39^2^ 123 ± 3.3 121 ± 3.3 1.5 ± 0.08 Cathepsin C 34 ± 7 16 ± 4 0.4 ± 0.1 \<0.01 Cathepsin D/E 34 ± 6 30 ± 14 22 ± 2 1.6 ± 0.1 Cathepsin H 2.8 ± 0.8 3.5 ± 0.7 1 ± 0.2 0.9 ± 0.2 Cathepsin J 26 ± 0.3 0.7 ± 0.3 3.0 ± 0.1 0.2 ± 0.09 Cathepsin L 14 ± 0.6 26 ± 11 19 ± 0.6 0.6 ± 0.08 BLT esterase 25,000 ± 400 58,000 ± 4,000 21,300 ± 100 \<100 Suc-FLF-SBz esterase 1,900,000 ± 61,000 310,000 ± 10,800 1,150,000 ± 10,200 12,000 ± 100 Suc-AAPFSBZ esterase 433,000 ± 2,900 134,000 ± 7,300 360,000 ± 6,100 3,400 ± 600 ^1^Extracts prepared from the indicated cell types were examined for protease activities using a panel of standard substrates. ^2^Enzymatic activities are expressed as nmol/min/mg soluble protein. Data shown represent the mean ± SD from three independent preparations. ::: We next asked whether DC are capable of digesting a complex protein Ag into antigenic products. In this regard, it has been reported previously that the original XS52 DC, as well as clones derived from this line, are capable of presenting KLH to the KLH-specific Th1 clone HDK-1 \[[@B23]\]. These results, however, did not fully test the processing capacity because it remained uncertain whether the conventional KLH preparation, which also contained many small molecular weight species, was indeed \"processed\" before effective presentation. For this reason, we developed a new experimental system using two PPD reactive T cell clones, a Th1 clone LNC.2F1 and a Th2 clone LNC.4K1. The advantage of PPD lies in the relative certainty of its purity. When pulsed with native PPD for 8 hr, XS52 DC were capable of stimulating both T Cell clones effectively. In dose-response experiments (Fig [1A](#F1){ref-type="fig"}), XS52 DC induced maximal activation of both T cell clones at 25--100 μg/ml of PPD, whereas no significant activation was observed, even at higher concentrations, in the absence of XS52 DC (data not shown). Importantly, chloroquine (100 μM) inhibited completely the capacity of XS52 cells to activate both Th1 and Th2, T cell clones (Fig. [1B](#F1){ref-type="fig"}), indicating the requirement for processing of PPD in an acidic environment. These observations indicate that XS52 DC do possess the capacity to digest a complex protein Ag into an immunogenic Ag. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **XS52 DC are capable of presenting native PPD effectively to T cells:**(A) XS52 DC were γ-irradiated and then pulsed for 8 hr with the indicated concentrations of PPD. The PPD-reactive Th1 clone (diamonds) or Th2 clone (squares) (10^5^cells/well) was cultured for 2 days with PPD-pulsed XS52 cells (10^4^cells/well). (B) Following a 3 hr incubation with or without chloroquine (100 μM), XS52 DC were pulsed with PPD (100 μg/ml) in the presence or absence of chloroquine (100 μM) and then examined for their capacity to activate the PPD-specific Th1 and Th2 clones. Data shown are the mean ± SD (n = 3) of ^3^H-thymidine uptake. Baseline proliferation of γ-irradiated XS52 DC alone was \<300 cpm. :::  ::: Pepstatin A inhibits the capacity of XS52 DC to present native PPD to T cells ----------------------------------------------------------------------------- To identify the proteases responsible for processing PPD, we employed three inhibitors: pepstatin A (aspartic acid protease inhibitor), DCI (serine protease inhibitor), and E-64 (cysteine protease inhibitor). XS52 DC were pulsed for 8 hr with native PPD in the presence of each inhibitor and then examined for the capacity to activate PPD-reactive Th1 and Th2 clones. To ensure a maximal effect, inhibitors were also added to cocultures of XS52 DC and T cells. As noted in Figure [2A](#F2){ref-type="fig"}, pepstatin A (100 μg/ml) almost completely blocked the capacity of XS52 cells to stimulate both T cell clones. When XS52 DC were pretreated with pepstatin A only during the 8 hr of Ag pulsing (but not during the subsequent coculture with T cells), we also observed significant, albeit less effective, inhibition (data not shown). By contrast, neither DCI nor E-64 caused any significant inhibition (Figure [2A](#F2){ref-type="fig"}). No inhibition was observed after treatment with 1% DMSO or 15 mM ammonium chloride alone, which was used to dissolve the above inhibitors. With respect to the mechanism of pepstatin A inhibition, the XS52 DC remained fully viable after 8 hr pre-incubation with pepstatin A (Figure [2B](#F2){ref-type="fig"}), thus excluding the possibility that pepstatin A had simply killed the XS52 DC. When pepstatin A was added to XS52 DC that had been pulsed with PPD and then fixed with paraformaldehyde, no inhibition was observed (Figure [3A](#F3){ref-type="fig"}). Moreover, pepstatin A failed to affect the capacity of XS52 DC to stimulate allogeneic T cells in a primary mixed lymphocyte reaction (Figure [3B](#F3){ref-type="fig"}); making it unlikely that pepstatin A had impaired the T cell-stimulatory capacity of XS52 DC. Finally, pepstatin A treatment was only effective when the native form of PPD was used as complex Ag, whereas it caused no inhibition when trysin-digested PPD fragments were employed (Figure [3C](#F3){ref-type="fig"}). Based on these observations, we concluded that pepstatin A had primarily inhibited the processing events. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Pepstatin A inhibits the capacity of XS52 DC to present native PPD:**(A) γ-irradiated XS52 DC were pulsed with PPD (100 μg/ml) in the presence or absence of each protease inhibitor (100 μg/ml pepstatin A, 100 μM DCI, or 100 μM E-64) or vehicle alone (1% DMSO or 15 mM NH~4~Cl). XS52 DC were then cultured for 2 days with the PPD-reactive Th1 or Th2 clone in the continuous presence of the same inhibitor or vehicle alone. Data shown are the mean ± SD (n = 3) of ^3^H-thymidine uptake in three representative experiments. (B) XS52 DC were incubated with each of protease inhibitor (100 μg/ml pepstatin A, 100 μM DCI, or 100 μM E-64) or vehicle alone (1% DMSO or 15 mM NH~4~Cl) for 16 hrs. Subsequently, cells were harvested and their viability was measured by trypan blue. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Failure of pepstatin A to inhibit the Ag presenting capacity of PPD-pulsed and fixed XS52 DC:**(A) γ-irradiated XS52 DC were pulsed with PPD and then fixed with paraformaldehyde (left panels). Alternatively, XS52 DC were first fixed and then pulsed with PPD. Subsequently, the XS52 DC were cultured with the PPD-specific Th1 or Th2 clone in the presence or absence of pepstatin A. Data shown are the mean ± SD (n = 3) of ^3^H-thymidine uptake. (B): Allogeneic splenic T cells isolated from CBA mice (5 × 10^5^cells/well) were cultured for 4 days with the indicated numbers of γ-irradiated XS52 DC in the presence or absence of pepstatin A. Data shown are the mean ± SD (n = 3) of ^3^H-thymidine uptake. (C): γ-irradiated XS52 DC were pulsed for 8 hr with either native PPD or trypsin-digested PPD in the presence or absence of pepstatin A. XS52 DC were then cocultured for 4 days with PPD-reactive Th1 or Th2 clones in the presence or absence of pepstatin A. Cocultures were then pulsed for 18 hr with ^3^H-thymidine and then harvested using a β-counter. :::  ::: Functional role of cathepsin D/E in the processing of PPD by XS52 DC -------------------------------------------------------------------- To identify the protease(s) inhibited by pepstatin A, XS52 DC were pretreated for 1 hr with pepstatin A (100 μg/ml), and extracts prepared from these cells were then examined for enzymatic activities. As noted in Figure [4](#F4){ref-type="fig"}, [1](#F1){ref-type="fig"} hr pretreatment with pepstatin A was sufficient to block cathepsin D/E activity significantly (\>70%). Pepstatin A also inhibited, albeit less effectively, cathepsin J activity and it had no significant effect on other tested protease activities. On the other hand, DCI and E64 were highly inhibitory of the chymotrypsin-like activities as well as cathepsin B, J, and/or L activities, but they did not inhibit cathepsin D/E. These results corroborate previous reports that pepstatin A inhibits cathepsin D/E activity relatively selectively \[[@B35]\]. Thus, it appears that cathepsin D/E is the primary target of pepstatin A, with the implication that these proteases play important roles in processing PPD by XS52 DC. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Pepstatin A Inhibits selectively the cathepsins D/E.**XS52 DC were pretreated for 60 min with each of protease inhibitors or vehicles. After extensive washing, the cells were extracted and subsequently examined for protease activities. Data shown are % inhibition compared with untreated control cells. :::  ::: Cathepsins D and E are prototypic aspartic acid proteases, which exhibit maximal enzymatic activities at acidic pH. Because both digest denatured hemoglobin effectively, the substrate used to measure cathepsin D/E activity, and because both are equally susceptible to pepstatin A treatment, it remained uncertain where processing of PPD in XS52 DC was mediated by cathepsin D, or cathepsin E, or both. As a first step to answer this question, we detected cathepsin D mRNA by RT-PCR in the XS52 DC line, as well as in 4F7^+^splenic DC and a bone marrow derived DC line, indicating that DC do possess the capacity to produce cathepsin D (Figure [5](#F5){ref-type="fig"}). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **DC constitutively express cathepsin D mRNA.**Total RNA isolated from the indicated cell types were subjected to RT-PCR analysis for cathepsin D and β-actin. Data are shown, including bone marrow DC and macrophages, as well as 4F7^+^splenic DC (splDC), products after 25 cycles of amplification. :::  ::: Conclusion ========== The experiments reported in this study provide new information with respect to complex Ag processing by DC. First, the long-term DC line, XS52 DC, was capable of processing PPD into immunogenic peptides, in the complete absence of other cell types. Although previous studies using several different DC preparations have documented similar results (3--12), this is the first report validating the Ag processing capacity of DC, in the absence of contaminating cells. Second, we have characterized the protease profiles expressed by DC. XS52 DC, 4F7^+^splenic DC, and bone marrow-derived DC, all exhibited significant protease activities for cathepsins B, C, D/E, H, J, and L, BLT esterase, and chymotrypsin. Thus, DC possess the capacity to produce a family of protease activities. Finally, pepstatin A, but not other protease inhibitors, abrogated almost completely the ability of XS52 DC to digest native PPD into an antigenic product, suggesting an important role for pepstatin A-sensitive proteases (most likely cathepsin D and/or E) during Ag processing by DC. Taken together, these results reinforce the concept that DC are fully capable of processing complex protein Ag into antigenic peptides. As described before, macrophages and B cells have been reported to employ cathepsins B, D, and E primarily to digest complex protein Ag, such as ovalbumin (OVA), hen egg white lysozyme (HEL), myoglobin, exogenous IgG, and *Staphylococcus aureus*nuclease (17--22). Here we report that DC also employ cathepsin D and/or E to digest PPD into an immunogenic Ag-product. This conclusion is supported by several lines of evidence: a) pepstatin A, but not other protease inhibitors, completely blocked the presentation of intact PPD by XS52 DC to PPD-reactive Th1 and Th2 clones, whereas it did not affect the presentation of PPD fragments; b) pepstatin A pretreatment inhibited cathepsin D/E activity selectively among the DC-associated protease activities; and c) all tested DC preparations expressed cathepsin D mRNA constitutively. In this regard, DC isolated from the mouse thoracic duct have been reported to produce neglible, if any, cathespin D immunoreactivity (assessed by immunofluorescence staining), whereas peritoneal macrophages produced relatively large amounts \[[@B14]\]. Also comparable levels of cathepsin D/E activity were detected in extracts from bone marrow-derived DC and from bone marrow-derived macrophages (data not shown). This discordance may reflect differences in the DC preparations tested and/or in the assays employed to detect cathepsin D. Nevertheless, our observations indicate that DC employ cathepsin D/E to degrade some protein Ag, with the implication that pepstatin A and other cathepsin D/E inhibitors \[[@B36]\] may be useful to prevent and even to treat unwanted hypersensitivity reactions against such protein Ag. It is important to emphasize that different protein Ag may be degraded by different proteases in DC. Moreover, DC isolated from different tissues or in different maturational states may employ different proteases. For example, murine DC isolated from the thoracic are unable to digest human serum albumin effectively \[[@B14]\], and murine splenic DC purified following overnight culture have failed to degrade KLH significantly into a TCA-soluble form \[[@B13]\]. Moreover, several reports document that LC lose their Ag processing capacity as they mature in culture \[[@B3]-[@B6],[@B12]\]. Thus, it will be interesting to compare DC from different tissues and in different states of maturation for their protease profiles and susceptibilities to pepstatin A treatment. We believe that the experimental system described in this report will provide unique opportunities to study the function of proteases and the regulation of their production in DC. Competing Interests =================== None declared. Author\'s Contributions ======================= Dr. Mohamadzadeh is the major contributor (15%) of the experimental data and a rough draft of the paper. The next three intermediate authors\' contributed remaining data and advice. Dr. Luftig was the overall individual who directed the several drafts and contributed to providing a new set of references to the manuscript. Acknowledgements ================ This work was supported by NIH grant DA016029 (MM) and Tulane base grant RR00164 (MM). The authors would like to thank Dr. M. J. McGuire (UTSMC, Dallas, Texas) for his support and the fruitful discussions.

PubMed Central

2024-06-05T03:55:47.246014

2004-8-2

PMC514721

Background ========== The natural history of Hürthle cell carcinoma (HCC) is not well understood. It accounts for \<5% of all differentiated thyroid malignancies. Hürthle cells are characterised by eosinophilic cytoplasm with trabecular/follicular growth pattern. \[[@B1]\]. Oncocytes are seen in follicular cell carcinoma but in HCC oncocytes represent more than 75% of cells, which exhibit a rather more trabecular growth pattern \[[@B2]\]. There is much debate regarding its clinical behaviour and little is known about the long-term survival of patients with HCC. Some studies have reported a relatively benign course while others have found the tumour to behave aggressively \[[@B3]-[@B6]\]. Most studies show that advanced age (\>45), male sex, size of primary tumour (\>4 cm), degree of invasion and recurrence are poor prognostic indicators \[[@B6]-[@B8]\]. Fine needle aspiration cytology is a good predictor of Hürthle cell neoplasm but is of little diagnostic value in evaluating HCC, since for a tumour to be deemed malignant one has to show vascular or capsular invasion \[[@B9]\]. Intraoperative frozen sections have a low predictive value. Udelsman *et al* found that in 96.4% cases with follicular neoplasm of thyroid, frozen section was neither informative nor cost-effective \[[@B10]\]. Well-encapsulated HCC run a favourable course while locally advanced HCC are associated with higher mortality and should be treated aggressively \[[@B4],[@B11]\]. In a well-differentiated thyroid carcinoma death resulting from local disease is unusual and most die of distant metastases \[[@B12]\]. We report a case of a Hürthle cell carcinoma presenting in the left lobe of thyroid following a right hemithyroidectomy for a colloid goitre 24 years ago. Case presentation ================= A 79-year-old male was referred in March 2003 with a lump on the left side of his neck. The patient had noted a sudden increase in the size of the lump over the preceding two months. He did not report any neck pressure symptoms, weight loss or anorexia. His past history included right partial thyroidectomy for a solitary nodule (colloid goitre) in 1978 and repair of abdominal aortic aneurysm in 1994. He had suffered myocardial infarction in 1995 and had an episode of acute coronary insufficiency in January 2003. His recent coronary angiograms showed an occluded left anterior descending artery and echocardiogram revealed good left ventricular function. He was a non-smoker and consumed alcohol in moderation. He had been taking warfarin, diltiazem MR, lisinopril, uniphyllin, glyceryl trinitrate tablets and buccal suscard. On examination he had left sided goitre extending superiorly into the posterior triangle and inferiorly into the retrosternal space, with variable consistency. The trachea was deviated to the right and there was cervical lymphadenopathy on the left side. Systemic examination was unremarkable and fine needle aspiration of thyroid gland showed presence of Hürthle cells. Computerised tomographic (CT) scan with contrast enhancement (figure [1](#F1){ref-type="fig"} &[2](#F2){ref-type="fig"}) of the neck and thorax revealed large left sided thyroid goitre with significant mediastinal extension. It showed mixed attenuation with foci of calcification peripherally. There was a 3 cm complex mass on the left side of the neck, posterior to the carotid sheath structures and deep to the sternomastoid, indicative of lymph node metastases. Thyroid profile and routine blood investigations were unremarkable. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Superior extension of left goitre with 3 cm diameter complex mass deep to sternomastoid, posterior to carotid sheath. Note the displacement of larynx to the right. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Mediastinal extension of left goitre. :::  ::: Based on the above findings radical surgery was planned. On exploration of the neck we confirmed left goitre with intrathoracic extension and enlarged lymph nodes under the sternocleidomastoid close to the jugulodigastric muscle and surrounding the carotid sheath. There was no remnant thyroid tissue seen on the right side following the previous thyroid surgery. Left hemithyroidectomy with modified neck dissection (lymphadenectomy, preserving all vessels and nerves) was performed. Macroscopic examination of the thyroid lobe showed a well defined solid pale brown mass approximately 8 cm in maximum dimension, surrounded by a narrow rim of preserved thyroid tissue. The lymph node specimen comprised of several nodules of partly necrotic tissue. Microscopic examination showed the thyroid lobe containing a Hürthle cell neoplasm, which was mostly encapsulated, with foci of capsular and vascular invasion. The two lymph nodes revealed metastatic Hürthle cell carcinoma. \[pT3, N1a, Mx\], (Figure [3](#F3){ref-type="fig"} &[4](#F4){ref-type="fig"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Photomicrograph showing capsular invasion (Haematoxylin and Eosin ×200) :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### photomicrograph showing Hürthle cell note the eosinophilic cytoplasm and prominent nucleoli (Haematoxylin and Eosin ×500). :::  ::: The patient had adjuvant therapy with oral radioiodine 131 (3060 MBq Sodium Iodine). He was put on a daily dose of 100 mcg of thyroxine. This was to be followed by a second dose of 5911 MBq of radioactive iodine six months from the time of the first dose. Discussion ========== Hürthle cell carcinomas are heterogeneous neoplasms that display a wide range of biological behaviour and accounts for less than 5% of all differentiated thyroid malignancies. The term HCC should be restricted to tumours with more than 75% of oncocytic cells \[[@B2]\]. Oncocytes are seen in follicular thyroid cell carcinoma and in papillary thyroid cell carcinoma \[[@B13],[@B14]\]. On one hand patients with HCC live for years with slow growing tumour and lymphatic metastases and on the other hand, patients die of highly aggressive tumour with haematogenous spread. Our patient had several indicators for poor prognosis such as his advanced age, male gender, large tumour size (8 cm), extra thyroid extension and nodal metastasis. Interestingly enough the patient had no pressure symptoms despite marked deviation of larynx, trachea and oesophagus, which may be due to previous right hemithyroidectomy. In elderly patients with sudden enlargement of neck mass and pre-existing thyroid conditions such as benign thyroid nodule, goitre (as in our case), Grave\'s disease or differentiated thyroid carcinoma, one has to bear in mind anaplastic thyroid carcinoma (ATC). In ATC local compression symptoms such as hoarseness, strider, dyspnoea and dysphagia occur as a rule \[[@B15]-[@B17]\]. In aggressive type of HCC haematogenous spread has been noted, but in ATC, at presentation patients are quite likely to have distant metastases involving lung, bone, brain and soft tissues \[[@B15],[@B16]\]. Our patient had undergone fine needle aspiration cytology, which revealed Hürthle cells. Since the lesion was rapidly growing with mediastinal extension and nodal involvement, the patient underwent staging and left hemithyroidectomy with modified neck dissection. Histology confirmed HCC based on vascular and capsular invasion. Intraoperative frozen sections have low predictive value and are particularly not a sensitive test for diagnosing HCC therefore this was not carried out \[[@B10]\]. McIvor et al have clearly shown that FNAC can easily recognise the tumour as Hürthle cell lesion \[[@B9]\]. Cases with suspicious histology and over 50 years of age carry a high risk of cancer \[[@B18]\]. In the management of HCC the primary mode of treatment is surgical, ranging from hemithyroidectomy to total thyroidectomy. Larger tumours (\>T2) require total thyroidectomy and lymphadenectomy if lymph nodes are involved \[[@B8]\]. Adjuvant radioiodine treatment or external beam radiotherapy is used for widely invasive carcinoma or locally advanced disease \[[@B8]\]. Several reports in literature have identified contra lateral foci of carcinoma in 40--70% of cases of HCC \[[@B11],[@B19]\]. HCC is less responsive to radioactive iodine therapy \[[@B20]\] and taking into account the aggressive behaviour, it has been suggested that every Hürthle cell tumour greater than 2 cm should be treated by total thyroidectomy \[[@B21]\]. In 1990 they showed that recurrent disease was noted in 17% of patients treated with total thyroidectomy compared to 59% in cases where a more limited procedure was carried out \[[@B21],[@B22]\]. Other authors support the role of total thyroidectomy as there is 15 to 35% incidence of multiple foci in HCC \[[@B23]\]. There are several reasons favouring the use of ^131^I remnant ablation after near-total thyroidectomy \[[@B24]\]. First, presence of thyroid remnant can obscure ^131^I uptake in cervical or lung metastases \[[@B25],[@B26]\]. Second, distant (lung) metastases may be seen only on the post treatment whole body scan after remnant ablation \[[@B27]\]. Finally, remnant ablation may destroy residual normal follicular cells, which may become malignant \[[@B28]\] and any occult cancer that may recur years later. Radioiodine therapy has no overall effect on mortality but subgroup analysis has shown that those patients who receive radioactive iodine for adjuvant ablation of remnant thyroid tissue have lower mortality rate compared with patients who either did not receive treatment or in whom the indication was the presence of residual disease \[[@B29]\]. Radioiodine uptake in the elderly is much lower. Schlumberger and colleagues noted ^131^I uptake at metastatic sites in only 53% of patients over 40 years of age, compared to 90% in patients below the age of 40 \[[@B30]\]. Univariate analysis indicated that older age and large tumour size predicted worse survival rates due to aggressive nature of the tumour (extra glandular invasion and multifocal disease). One recent series reviewed medical records of patients between the years 1944 and 1995. Of the 89 HCC cases studied, 29% had only undergone lobectomy as initial treatment and 50% had undergone partial resection. Of the three quarters of the patients in this series who received radioactive iodine, only 38% of patients with known metastases showed positive uptake \[[@B29]\]. Another study clearly suggested that treatment with ^131^I to ablate the thyroid remnant and to treat residual disease were independent prognostic variables that favourably influenced recurrence, distant recurrence, and cancer death rates \[[@B24]\]. Our patient received radioactive iodine treatment in the postoperative period. He has been followed up with whole body scans (Fig [5](#F5){ref-type="fig"} and [6](#F6){ref-type="fig"}), which indicate his response to adjuvant radioactive iodine therapy. He is on 125 mcg thyroxine in order to maintain a TSH level of less than 0.01 mIU/L and FT4 at the upper limit of normal (8--28 pmol/L). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Whole body scan on November 3, 2003 following ^131^I ablation therapy on 28^th^October 2003, with 3060 MBq Sodium Iodine (^131^I). Increased uptake is seen in the region of the thyroid bed. No abnormal accumulation was noted elsewhere. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### Whole body scan on 19^th^April 2004 following ^131^I ablation therapy on 13^th^April 2004 with 5911 MBq Sodium Iodine (^131^I). Two small focal area of uptake are seen in the thyroid bed. Low uptake focal area in the left lateral aspect of the neck, could possibly represent activity in a cervical node. :::  ::: Stojdinovic *et al* have treated 56 patients with HCC between the years 1940 and 2000 \[[@B8]\]. Of these patients 23(41%) had minimally invasive disease with no evidence of extra thyroid invasion (T2 N0 M0) and 33(56%) had widely invasive HCC. Primary mode of treatment was surgery ranging from lobectomy and isthumusectomy to total thyroidectomy with cervical lymphadenectomy in presence of lymph node involvement. Some patients received adjuvant radioiodine or external beam radiotherapy for widely invasive carcinoma. Study end points were relapse free survival and disease specific survival. They reported 8 years survival rate of 100% and 58% for low and high-risk cancers respectively. In their entire study cohort age was not found to predict the outcome but the most significant factor was widely invasive carcinoma. Khafif *et al* in their series (42 patients with HCC between 1957--1997) used radioiodine in patients with distant metastases; none had thyroid remnant ablation with radioactive iodine \[[@B4]\]. They reported an overall survival rate of 90.5% and noted that age, size of tumour and extent of resection adversely affected the prognosis. Hürthle cell lesion can be easily picked up on FNAC but to make a diagnosis of HCC one has to demonstrate vascular or capsular invasion. Intraoperative frozen sections have low predictive value and cases with advanced age (over 50), rapid enlargement of lump and palpable nodes should be regarded with high index of suspicion for presence of HCC. HCC or other differentiated carcinomas of thyroid in the elderly patients are generally more aggressive with less favourable prognosis compared to younger patients. They should be offered total thyroidectomy and selective lymph node dissection (when lymph nodes are involved) followed by ablative radioiodine therapy, provided they can withstand the above treatment. Coexisting medical disorders should be recognized and managed effectively prior to surgery \[[@B31]\]. Further research is needed to clarify the role of adjuvant radioiodine therapy in the management of HCC. Competing interests =================== None declared. Authors\' contribution ====================== **MRH**managed the patient, searched the literature and drafted the manuscript **LI:**did the histological study, and contributed to pathological aspects in the present study **DG:**Managed the patient, conceptualise the present report, edited the manuscript and coordinated after reviewing the manuscript Acknowledgement =============== Patients consent was obtained for publication of his case records.

PubMed Central

2024-06-05T03:55:47.247890

2004-8-11

PMC514883

Introduction {#s1} ============ Fatty acids (FAs) belong to a physiologically important class of molecules involved in energy storage, membrane structure, and various signaling pathways. Different FAs have different physical properties that determine their unique functions. Among the most abundant in animal cells as well as the most studied are those of long-chain even-numbered saturated and unsaturated FAs. C15ISO and C17ISO are saturated tetradecanoic and hexadecanoic FAs with a single methyl group appended on the carbon next to the terminal carbon ([Figure 1](#pbio-0020257-g001){ref-type="fig"}). Monomethyl branched-chain FAs (mmBCFAs) in ISO configuration as well as in anteISO configuration (methyl group appended on the second to the terminal carbon) also seem to be ubiquitous in nature. They are present in particularly large quantities in various bacterial genera, including cold-tolerating and thermophilic species ([@pbio-0020257-Merkel1]; [@pbio-0020257-Annous1]; [@pbio-0020257-Ferreira1]; [@pbio-0020257-Batrakov1]; [@pbio-0020257-Jahnke1]; [@pbio-0020257-Groth1]; [@pbio-0020257-Nichols1]). There, mmBCFAs contribute to the membrane function, regulating fluidity ([@pbio-0020257-Rilfors1]; [@pbio-0020257-Suutari1]; [@pbio-0020257-Cropp1]; [@pbio-0020257-Jones3]) and proton permeability ([@pbio-0020257-van1]). ::: {#pbio-0020257-g001 .fig} Figure 1 ::: {.caption} ###### Structure of mmBCFAs of 15 and 17 Carbons C15ISO, 13-methyl myristic acid; C17ISO, 15-methyl hexadecanoic acid; C17anteISO, 14-methyl hexadecanoic acid. Other mmBCFAs mentioned in the text are the following: C13ISO, 11-methyl lauric acid; C15anteISO, 12-methyl tetradecanoic acid. C15ISO and C17ISO are readily detectable in *C. elegans.* :::  ::: Although comprehensive reports on mmBCFAs in eukaryotes are lacking, sporadic data indicate that they are present in the fungi, plant, and animal kingdoms ([@pbio-0020257-Garton1]; [@pbio-0020257-Seyama1]; [@pbio-0020257-Martinez1]; [@pbio-0020257-Cropp1]; [@pbio-0020257-Wolff1]; [@pbio-0020257-Destaillats1]). In mammals, mmBCFAs have been detected in several tissues, including skin ([@pbio-0020257-Aungst1]), Vernix caseosa ([@pbio-0020257-Nicolaides1]), harderian and sebaceous glands ([@pbio-0020257-Nordstrom1]), hair ([@pbio-0020257-Jones1]), brain ([@pbio-0020257-Ramsey1]), blood ([@pbio-0020257-Holman1]), and cancer cells ([@pbio-0020257-Hradec1]). The fact that mmBCFAs are present in a wide variety of organisms implies a conservation of the related metabolic enzymes and consequently important and perhaps unique functions for these molecules ([@pbio-0020257-Jones1]). Nevertheless, their physiological roles and metabolic regulations have not been systematically studied and thus remain fragmentary. It was found that C21anteISO is the major covalently bound FA in mammalian hair fibers. A removal of this FA from its protein counterparts results in a loss of hydrophobicity ([@pbio-0020257-Jones1]). Other studies indicated that C17anteISO esterified to cholesterol binds to and activates enzymes of protein biosynthesis ([@pbio-0020257-Tuhackova1]; [@pbio-0020257-Hradec1]). A potential significance of mmBCFAs for human health is associated with a long-observed correlation between amounts of these FAs and disease conditions such as brain deficiency ([@pbio-0020257-Ramsey1]) and cancer ([@pbio-0020257-Hradec1]). More recent studies have revealed a role of another mmBCFA, C15ISO, as a growth inhibitor of human cancer where it selectively induces apoptosis ([@pbio-0020257-Yang1]). Given how important these FA molecules may be and how little is known about their biosynthesis and functions in eukaryotes, it is an opportune problem to study. De novo synthesis of long-chain mmBCFAs described for bacteria is quite different from the biosynthesis of straight-chain FAs ([@pbio-0020257-Smith1]; [@pbio-0020257-Oku1]; [@pbio-0020257-Toal1]). While the latter uses acetyl-coenzyme A (acetyl-CoA) as a primer condensing with a malonyl-CoA extender, branched-chain FA synthesis starts with the branched-chain CoA primers derived from the branched-chain amino acids leucine, isoleucine, and valine. To synthesize branched-chain FAs, organisms must have a system for supplying branched-chain primers along with the enzymes utilizing them ([@pbio-0020257-Smith1]). No such enzymes have been previously characterized in vivo for any eukaryotic organisms. Here we describe our approach to characterize the biosynthesis and function of mmBCFAs using the free-living nematode *Caenorhabditis elegans.* Combining genetic, molecular, and biochemical analyses, we show that the worm is not only able to synthesize mmBCFAs de novo but is also absolutely dependent on these FA species for its growth and development. Results/Discussion {#s2} ================== C. elegans Synthesizes Branched-Chain FAs De Novo and Uses Two FA Elongation Enzymes to Produce C15ISO and C17ISO {#s2a} ----------------------------------------------------------------------------------------------------------------- In characterizing FA elongation in *C. elegans,* we identified eight sequences homologous to the yeast long-chain FA elongation enzymes ([@pbio-0020257-Kniazeva1]). To test for their possible functions in vivo, we applied RNAi to the corresponding genes, followed by an analysis of FA composition in whole animals using gas chromatography (GC). RNAi treatment of four genes---*elo-3* (D2024.3), *elo-4* (C40H1.4), *elo-7* (F56H11.3), and *elo-8* (Y47D3A.30)---did not produce any notable phenotypes, whereas suppression of *elo-1* (F56H11.4) and *elo-2* (F11E6.5) affected the elongation of straight long-chain saturated and polyunsaturated FAs ([@pbio-0020257-Beaudoin1]; [@pbio-0020257-Kniazeva1]). Surprisingly, the RNAi treatment of the two remaining genes, *elo-5* (F41H10.7) and *elo-6* (F41H10.8), affected the levels of branched-chain FA. Transcriptional reporter constructs (*elo-5Prom*::*GFP* and *elo-6Prom*::*GFP*) indicated that both genes are expressed in the gut ([Figure 2](#pbio-0020257-g002){ref-type="fig"}). In addition, *elo-5* was expressed in unidentified head cells and *elo-6* was expressed in neurons, pharynx, and vulva muscles. ::: {#pbio-0020257-g002 .fig} Figure 2 ::: {.caption} ###### The Expression of *elo-5Prom*::*GFP* and *elo-6Prom*::*GFP* Constructs in Wild-Type Worms (A, C, E, and G) DIC images; (B, D, F and H) fluorescence images. (A--D) Strong expression of the *elo-5Prom*::*GFP* construct in the gut and in the head is shown. (E--H) The expression of the *elo-6Prom*::*GFP* construct in the gut, vulvae (white arrows), and nerve ring is shown. Scale bars, 100 μm. :::  ::: The RNAi of *elo-6* significantly reduced the amount of only C17ISO, while the RNAi of *elo-5* dramatically reduced quantities of both C15ISO and C17ISO ([Figure 3](#pbio-0020257-g003){ref-type="fig"}). These results indicate that ELO-5 might be involved in the biosynthesis of C15ISO and possibly also C17ISO, whereas ELO-6 may function in elongating C15ISO to C17ISO ([Figure 3](#pbio-0020257-g003){ref-type="fig"}C and [3](#pbio-0020257-g003){ref-type="fig"}D). To our best knowledge, these are the first enzymes that have been shown to be involved in long-chain mmBCFA biosynthesis in a nonbacterial in vivo system and the first enzymes of the long-chain FA elongation family related to mmBCFA production. ::: {#pbio-0020257-g003 .fig} Figure 3 ::: {.caption} ###### RNAi Treatment of *elo-5* and *elo-6* Significantly Alters the FA Composition (A and B) GC profiles showing the FA composition in the wild-type strain (Bristol N2) containing the RNAi feeding control vector and in the *elo-5(RNAi)* feeding strain. Arrowheads point to the peaks corresponding to C15ISO and C17ISO. \(C) Comparison of FA composition in three strains: wild type, *elo-5(RNAi),* and *elo-6(RNAi).* C17ISO is decreased in both RNAi strains, while C15ISO is only decreased in *elo-5(RNAi).* \(D) Suggested elongation reactions catalyzed by ELO-5 and ELO-6 in C15ISO and C17ISO biosynthesis. FAs are elongated by an addition of two carbon groups at a time. Combined data presented in this figure and in the text suggest that ELO-6 acts at the elongation step from C15 to C17, whereas ELO-5 may be involved in the production of both C15ISO and C17ISO. :::  ::: In bacteria, mmBCFA biosynthesis utilizes branched-chain α-keto-acids of leucine, isoleucine, and valine to produce mmBCFA acyl-CoA primers that substitute for acetyl-CoAs in conventional FA biosynthesis ([@pbio-0020257-Oku1]). Key enzymes engaged in synthesizing the mmBCFA acyl-CoA primers are branched-chain aminotransferase (BCAT) and the branched-chain α-keto-acid dehydrogenase (BCKAD) complex ([Figure 4](#pbio-0020257-g004){ref-type="fig"}A). The elongation of the mmBCFA backbone is then carried out by fatty acid synthetase (FAS). ::: {#pbio-0020257-g004 .fig} Figure 4 ::: {.caption} ###### The C. elegans BCKAD Homolog Is Involved in mmBCFA Biosynthesis \(A) Early steps of mmBCFA biosynthesis in bacteria, based on [@pbio-0020257-Smith1], [@pbio-0020257-Oku1], and [@pbio-0020257-Toal1]. IVD, isovaleryl-CoA dehydrogenase. Predicted corresponding C. elegans genes encoding predicted orthologs were identified (shown in italicized names of reading frames). \(B) GC profiles reveal differences in the FA composition in the wild-type animals and animals treated with RNAi of E1 alpha subunit of BCKAD encoded by Y39E4A.3. Black arrowheads point to C15ISO and C17ISO. \(C) A summary of several independent preparations shows a significant decrease in both mmBCFAs in the Y39E4A.3 dsRNA-treated animals (*p* = 0.001 and 0.008 for C15ISO and C17ISO, respectively). :::  ::: The ability of C. elegans to grow on the chemically defined axenic medium CbMM ([@pbio-0020257-Lu1]), which lacks the potential mmBCFA precursors, has suggested that the animals can synthesize mmBCFA de novo. If so, a disruption of the BCKAD complex could affect mmBCFA levels. We identified a predicted C. elegans protein, Y39E4A.3, with significant sequence homology to the E1 alpha subunit of BCKAD (Y39E4A.3 scores expect value 8e-50 on 57% of the length with the Bacillus subtilis BCKAD and 1.4e-134 on 88.4% of the length with the Homo sapiens BCKADs). RNAi of Y39E4A.3 led to a significant decrease in C15ISO and C17ISO production ([Figure 4](#pbio-0020257-g004){ref-type="fig"}B and [4](#pbio-0020257-g004){ref-type="fig"}C). RNAi suppression of another predicted component of the BCKAD complex, pyruvate dehydrogenase (T05H10.6), resulted in a similar decrease in C15ISO and C17ISO (unpublished data), indicating a role for the C. elegans BCKAD protein in long-chain mmBCFA biosynthesis. Thus, C. elegans appears to use the same initial reactions to produce mmBCFAs as bacterial cells. In addition, the worm uses enzymes of the FA elongation family, ELO-5 and ELO-6, to complete the pathway. A connection between BCKAD functions and mmBCFA quantities has been previously reported in humans ([@pbio-0020257-Jones2]). Normally hair fibers are densely covered with C21anteISO, which contributes about 38.2% to the total hair FAs ([@pbio-0020257-Jones1]). It was observed that patients with maple syrup urine disease, which is caused by an inherited mutation in the BCKAD gene, had a drastically reduced level of mmBCFAs in their hair. Together, these data suggest that long-chain mmBCFA biosynthesis could be similar in bacteria, *C. elegans,* and human. Blocking ELO-5 Function Causes Growth and Developmental Defects {#s2b} --------------------------------------------------------------- While the suppression of *elo-6* activity by feeding double-stranded RNA (dsRNA) to wild-type animals did not cause obvious morphological or growth defects, the suppression of *elo-5* resulted in more pronounced phenotypes ([Figure 5](#pbio-0020257-g005){ref-type="fig"}). Worms originating from wild-type eggs laid on the *elo-5(RNAi)* plates displayed no obvious growth or morphological abnormality until the second day of adulthood, when they developed an egg-laying defect ([Figure 5](#pbio-0020257-g005){ref-type="fig"}B). Eggs of the next generation hatched on time but the progeny arrested at the first of the four larval stages (L1). The small larvae maintained morphological integrity and could survive on a plate for up to 3--4 d. The arrest was only observed in progeny of parents exposed to *elo-5* RNAi at the L1 stage. ::: {#pbio-0020257-g005 .fig} Figure 5 ::: {.caption} ###### RNAi Treatment of *elo-5* Causes L1 Arrest and Other Physiological Defects (A--C) Nomarski images of worms grown from eggs placed on RNAi plates. Scale bars, 100 μm. \(A) Young adults had normal morphology and growth rates. \(B) On the second day of adulthood, these animals displayed an egg-laying defect; eggs hatched inside the worms. Arrows point to the late embryos and hatched larvae inside a worm. \(C) F1 generation arrested uniformly at the first larval stage (L1), and larvae arrested for 4--5 d died. (D--F) Images of worms derived from late larvae (L2--L4) placed on the RNAi plates. \(D) The F1 progeny of worms developed from the treated larvae had smaller size and a scrawny morphology compared to the wild type shown in (A). Scale bar, 100 μm. \(E) These animals produced very few oocytes, some of which gave rise to embryos and L1 worms. White arrows indicate embryos. Some oocytes remained unfertilized (black arrow). Scale bar, 10 μm. \(F) The proximal part of the gonads undergoes deterioration resulting in sterility. The white arrow indicates spermatica, the black arrow shows an abnormally amorphous oocyte, and the two-way arrow points to the clumsy gonad arm that is finely ordered in wild-type animals. Scale bar, 10 μm. :::  ::: When parental animals were subjected to *elo-5* RNAi at later larval stages (L2--L4), their progeny did not arrest in L1 but continued to develop into adulthood. These animals had no obvious defects in locomotion, pharyngeal pumping, intestinal contractions, chemotaxis response, touch sensitivity, or general anatomy (unpublished data). However, the growing worms became progressively sick ([Figure 5](#pbio-0020257-g005){ref-type="fig"}D--[5](#pbio-0020257-g005){ref-type="fig"}F). The gonads appeared normal at the L4 and early adult stages, but after fertilization of one to ten oocytes, oogenesis became impaired. Gonad degeneration began with a pronounced vacuolization in the midsection of the gonad followed by the appearance of disorganized clumps of nuclei in the proximal part. An egg-laying defect became apparent and only a few progeny arose from these worms, which then arrested at L1. The development of the *elo-5* RNAi phenotypes is likely due to a gradual elimination of the ELO-5--associated functions. Our data suggest that these functions are crucial for larval growth and development. We also obtained a likely null mutant of the *elo-5* gene, *elo-5(gk208),* which has a 245-bp deletion eliminating the predicted first exon (Genome Science Center, BC Cancer Research Center, Vancouver, British Columbia, Canada). This allele phenocopies the L1 arrest phenotype of the *elo-5(RNAi)* animals. A Deficiency of C15ISO and C17ISO FAs Is Solely Responsible for the Defects Caused by *elo-5(RNAi)* {#s2c} --------------------------------------------------------------------------------------------------- We reasoned that if the defects observed in the *elo-5(RNAi)* animals resulted directly from the deficiency of C15ISO and C17ISO, then feeding these worms with C15ISO and C17ISO should mask a shortage of endogenous C15ISO and C17ISO and permit the animals to grow normally. As predicted, the C17ISO and C17anteISO supplements rescued the *elo-5* RNAi defects (in 52 of 60 and 58 of 60 plates, respectively). A partial rescue was observed on the plates supplemented with C15ISO and C15anteISO (23 of 38 and 20 of 28 plates, respectively). Corroborating results were obtained when homozygous *elo-5(gk208)* animals supplied with C17ISO grew normally. In sharp contrast, neither saturated or mono- or polyunsaturated FA molecules (C16:0, C16:1 n7, C17:0, C18:3 n6), mmBCFAs with shorter or longer backbones (C13ISO, C18ISO, C19ISO), nor polymethyl branched phytanic acid were able to rescue or reduce defects (0 of 30 plates in each experiment). Therefore, we have determined that only dietary 17-carbon mmBCFAs are competent to bypass the biochemical defect caused by loss of ELO-5 function. GC analysis of FA composition in *elo-5(RNAi)* worms grown on supplemented plates revealed that only C17ISO and C17anteISO are significantly incorporated into lipids ([Figure 6](#pbio-0020257-g006){ref-type="fig"}A--[6](#pbio-0020257-g006){ref-type="fig"}C). Because the addition of C15ISO did not result in elongation to C17ISO ([Figure 6](#pbio-0020257-g006){ref-type="fig"}A), we wanted to determine whether ELO-6 was capable of extending an FA backbone in the absence of ELO-5, or whether the supplied free FA molecules could enter a different metabolic pathway, for instance, a degradation pathway. To distinguish between these two possibilities, we added mmBCFA-producing bacteria on top of the *elo-5(RNAi)* feeding Escherichia coli strain (HT115), which lacks mmBCFAs. This mmBCFA-producing strain was identified by chance; we noticed that in the presence of a certain bacterial contaminant the animals could overcome the *elo-5(RNAi)* effects. Using a rapid bacterial identification method ([@pbio-0020257-Lane1]), we determined the contaminant to be *Stenotrophomonas maltophilia.* GC analysis revealed that this bacterial strain produced a high quantity of C15ISO and C15anteISO but not 17-carbon mmBCFAs ([Figure 6](#pbio-0020257-g006){ref-type="fig"}D). GC analysis of *elo-5(RNAi)* animals fed with S. maltophilia indicated that they not only accumulated bacterial C15ISO and C15anteISO but also efficiently elongated these FA species to C17ISO and C17anteISO, which are absent in S. maltophilia ([Figure 6](#pbio-0020257-g006){ref-type="fig"}D and [6](#pbio-0020257-g006){ref-type="fig"}E). This suggested that elongation from C15ISO to C17ISO mmBCFA was not impaired in the *elo-5(RNAi)* animals. Therefore, ELO-6 function remains intact in *elo-5(RNAi).* Apparently, C15ISO added to the plates could not be utilized by ELO-6 whereas C15ISO-CoA and/or C15anteISO-CoA originating from the bacterial food could, suggesting that free and esterified mmBCFAs were likely to enter alternative pathways. ::: {#pbio-0020257-g006 .fig} Figure 6 ::: {.caption} ###### The FA Composition in Worms Maintained on *elo-5* RNAi Plates Supplemented with FA or with S. maltophilia Enriched with C15ISO and C15anteISO FA Black arrowheads indicate positions of mmBCFAs. \(A) Animals grown with C15ISO supplements were partially rescued to the wild-type phenotype; however, no accumulation of C15ISO or its elongation to C17ISO was detectable. (B and C) Animals grown with the (B) C17ISO or (C) C17anteISO supplements were fully rescued. Peaks corresponding to C17ISO and C17anteISO are prominent. \(D) The FA composition in *S. maltophilia.* Arrowheads point to the major FAs, C15ISO and C15anteISO. \(E) The *elo-5(RNAi)* animals are able to elongate dietary C15ISO and C15anteISO into C17ISO and C17anteISO. Arrowheads indicate mmBCFAs. The horizontal arrow illustrates the elongation from C15 to C17 mmBCFA. :::  ::: The essential roles of C15ISO and C17ISO were also supported through an examination of the *elo-5(gk208)* deletion mutant. The homozygous mutants grew without any obvious morphological defects when maintained on the plates supplemented with C17ISO or seeded with *S. maltophilia.* However, removal of the FA supplements or S. maltophilia by bleaching resulted in the same L1 arrest phenotype seen for the *elo-5(RNAi)* worms. L1 Arrest of the *elo-5(RNAi)* Animals Is Reversible and Related to the Variations in Levels of C17ISO during Development {#s2d} ------------------------------------------------------------------------------------------------------------------------- We then asked if *elo-5(RNAi)* animals arrested at the L1 stage could be recovered by adding the 17-carbon mmBCFA supplements. Indeed, C17ISO and C17anteISO could effectively release L1 larvae from the developmental arrest; about 50% of 2-d-arrested and 1% of 4-d-old L1 were rescued to full growth and proliferation. Since C17anteISO could not be detected in the laboratory animals under normal conditions of culturing, C17ISO appeared to be the principal molecule conveying the ELO-5 function. Therefore, the L1 arrest of the C17ISO-depleted worms is both completely penetrant and reversible, indicating that C17ISO plays a critical role in growth and development at the L1 stage. The analysis of the FA levels of staged worms revealed that the C17ISO level increases gradually from a relatively low level at L1 to its peak in gravid adults containing eggs ([Figure 7](#pbio-0020257-g007){ref-type="fig"}A). Based on the analysis of green fluorescent protein (GFP) reporter constructs (unpublished data) and in situ hybridization data (results from NextDB by Y. Kohara, Tokyo, Japan), neither *elo-5* nor *elo-6* is significantly expressed in eggs or L1. Therefore, C17ISO likely accumulates in embryos during oogenesis. It may be directly transported from gut to gonads, since both ELO-5 and ELO-6 were expressed mainly in the gut and since feeding C17ISO rescued the *elo-5* mutant phenotypes. When RNAi-mediated disruption of *elo-5* occurs at the L1 stage of a parent and consequently blocks C17ISO synthesis from that stage on, the eggs and L1 animals of the next generation are expected to contain a critically low concentration of C17ISO, halting further development. Because the arrested L1 can be rescued by a dietary supply of the mmBCFA, the deficiency is not likely to cause critical defects during the embryonic and early postembryonic periods. ::: {#pbio-0020257-g007 .fig} Figure 7 ::: {.caption} ###### A Fluctuation of the C17ISO Amounts in Development \(A) Relative amounts of C15ISO and C17ISO in the worm samples collected at different developmental stages. The amount of the mmBCFA molecule is presented as the percentage of total FA in each sample. Grey bars, C15ISO; black bars, C17ISO. \(B) Proposed relationship between the levels of mmBCFA during development and the RNAi effects. Depending on the time of RNAi onset, the amount of C17ISO in F1 eggs varies. If *elo-5* is suppressed in parental animals after they have begun to synthesize mmBCFA, then their eggs will have a reduced C17ISO level that is still above the critical low level, which permits these animals to grow but causes them to display gonadal defects. These worms produce a small number of progeny that is then arrested in L1. If parental animals are treated with *elo-5(RNAi)* right after hatching, they are unable to initiate mmBCFA biosynthesis and the levels of C15ISO and C17ISO in their eggs are reduced to below the critical low level, resulting in L1 arrest of their progeny. :::  ::: If *elo-5* RNAi is applied to the parent worms at or after the L2 larval stage, when the amount of C17ISO has already been elevated and/or the RNAi effect is less penetrant, the progeny may receive sufficient C17ISO to pass the L1 arrest stage. The resulting animals, however, become visibly unhealthy at the L4 and adult stages as mentioned earlier, suggesting that C17ISO also plays a role in late developmental stages. Based on these results, we propose a relationship between the amounts of C17ISO and developmental stages ([Figure 7](#pbio-0020257-g007){ref-type="fig"}B). In this model, the level of C17ISO is monitored at the first larval stage and the decision is made whether to proceed or pause in development. The analysis of GC data from staged animals has also indicated that the variation of the C17ISO level is correlated with only two other FA species, suggesting a potential compensatory and coregulatory mechanism. The C17ISO Level Correlates with the Levels of Two Other FAs during Development {#s2e} ------------------------------------------------------------------------------- FA homeostasis implies that relative amounts of various FA species are coordinated and balanced for optimal performance. To obtain information that may help us understand why and how numerous FAs and their specific metabolic enzymes are maintained in nature, we carried out analysis to determine a possible correlation between changes in the levels of C17ISO and other FAs detected in worms. We have analyzed a large amount of GC data (*n* = 50) obtained from mixed populations of wild-type animals where the fractions of eggs, larvae, and adults randomly varied. We also included GC data separately obtained from staged worms: eggs, L1, L2, L3, L4, and gravid adults. We found that the amounts of C17ISO significantly correlated with only two other FA molecules: linoleic acid (C18:2 n6) and vaccenic acid (C18:1 n7) ([Figure 8](#pbio-0020257-g008){ref-type="fig"}). A potential physiological significance of these correlations is intriguing. ::: {#pbio-0020257-g008 .fig} Figure 8 ::: {.caption} ###### Correlation between the Level of C17ISO and the Levels of Linoleic and Vaccenic Acids during Development Graphical illustrations of the correlation between the levels of C17ISO and (A) vaccenic acid and (B) linoleic acid. Data were obtained by GC analysis of synchronized populations of worms. Combined with the GC measurements generated from 50 additional samples (see Materials and Methods), these data were used to calculate correlation coefficients: CORREL ~C17ISO/C18:2\ n6~ = 0.82772, T-TEST = 6.54814 × 10^−7^, and CORREL ~C17ISO/C18:1\ n7~ = −0.85162, T-TEST = 4.74094 × 10^−5^. Black bars, C17ISO; white bars, vaccenic acid; grey bars, linoleic acid. :::  ::: The observed negative correlation between the levels of C17ISO and C18:1 n7 throughout development may indicate a compensatory adjustment important for physiological functions, such as retention of the cell membrane physical properties. mmBCFAs and monounsaturated straight-chain FAs have been previously implicated in regulating membrane fluidity, which depends on the ratio of saturated FA to monounsaturated and branched-chain FA content in bacterial cells ([@pbio-0020257-Rilfors1]; [@pbio-0020257-Suutari1]; [@pbio-0020257-Cropp1]). An elevation in monounsaturated FA amounts in response to the decrease of branched-chain FAs, but not vice versa, was observed in Streptomyces avermitilis ([@pbio-0020257-Cropp1]), suggesting that monounsaturated FAs may sense a state of membrane fluidity. In the *elo-5(RNAi)*-treated worms, a substantial loss of C15ISO and C17ISO is also accompanied by a change in the FA composition, most noticeably by the elevation in C18:1 n7 (see [Figure 3](#pbio-0020257-g003){ref-type="fig"}C), a result consistent with the above observation. To estimate the effect of the C15ISO and C17ISO deficiency on the membrane saturation, the saturation index (SI = \[saturated FA\]/\[mmBCFA + monounsaturated FA\]) was calculated. No significant differences were detected in *elo-5(RNAi)* worm compared to wild type (SI = 0.325 ± 0.011 \[*n* = 6\] and SI = 0.320 ± 0.032 \[*n* = 5\], respectively). Therefore, *elo-5(RNAi)* may not cause massive cell membrane dysfunction. A positive correlation between the amounts of C17ISO and C18:2 n6 may suggest a potential common function during development. In addition to the importance of linoleic acid as a substrate for polyunsaturated FA biosynthesis, its hydroxylated fatty acid derivative (HODEs) is known as a signaling molecule affecting chemotaxis ([@pbio-0020257-Kang1]), cell proliferation ([@pbio-0020257-Eling1]), and modulation of several enzymatic pathways ([@pbio-0020257-Hsi1]). A correlation between C17ISO and linoleic acid may also suggest a similar regulation of biosynthesis of the two molecules. The changes in the FA composition associated with a decrease in C15ISO and C17ISO indicate that the metabolism of straight-chain FA species is responsive to the mmBCFA levels and suggest a cross regulation. Interestingly, in the *elo-5(RNAi)* animals fed with C15ISO or C15anteISO containing bacterial supplement *(S. maltophilia),* the FA composition was significantly altered (see [Figure 6](#pbio-0020257-g006){ref-type="fig"}E). It appears that mmBCFAs become principal components in a range of 16--18-carbon FAs. This suggests that large quantities of mmBCFAs are not toxic. In contrast, because these worms grow and proliferate well, mmBCFAs seem to be efficient substitutes for saturated and monounsaturated straight-chain FAs. The Worm SREBP Homolog Controls Production of Branched-Chain FAs {#s2f} ---------------------------------------------------------------- In mammals, straight-chain FA biosynthesis depends on the 1c isoform of sterol regulatory element binding protein (SREBP-1c), which promotes the expression of FA metabolic enzymes ([@pbio-0020257-Edwards1]; [@pbio-0020257-Horton1]; [@pbio-0020257-Matsuzaka1]). There is only one protein in C. elegans that is homologous to mammalian SREBPs, Y47D3B.7 (the gene has been named *lpd-1,* for "lipid depleted 1") ([@pbio-0020257-McKay1]). McKay and coauthors have shown that worms treated with *lpd-1* RNAi display a lipid-depleted phenotype. They have also shown that *lpd-1* regulates the expression of several lipogenic enzymes, acetyl-CoA carboxilase (ACC), FAS, and glycerol 3-phosphate acyltransferase (G3PA) ([@pbio-0020257-McKay1]). Thus, similar to its mammalian homolog, *lpd-1* is involved in straight-chain FA biosynthesis. We wanted to see if *lpd-1* also plays a role in mmBCFA metabolism. We first applied RNAi to *lpd-1* and determined the FA composition of the mutant worms. As expected, the FA content of treated animals was significantly changed, but surprisingly the most reduced were the levels of C15ISO and C17ISO ([Figure 9](#pbio-0020257-g009){ref-type="fig"}). Also significantly reduced was the amount of C18:2 n6. In contrast, the C16:0 level was elevated. These data indicate that, in addition to regulating the first steps of global FA biosynthesis through the activation of the ACC and FAS transcription, the worm SREBP homolog regulates mmBCFA elongation as well as desaturation of straight-chain FAs. ::: {#pbio-0020257-g009 .fig} Figure 9 ::: {.caption} ###### RNAi of the C. elegans SREBP Homolog Alters the FA Composition (A and B) The GC profiles of (A) wild-type and (B) *lpd-1(RNAi)*-treated worms. \(C) A summary of several independent GC runs. Bars represent the percentages of total FAs. The levels of C15ISO, C17ISO, and C16:0 are significantly altered by the RNAi treatment. Black arrowheads point to differences in the C15ISO and C17ISO amounts. Grey arrowheads indicate the changes in palmitic acid, C16:0. :::  ::: As reported previously, disruption of *lpd-1* through a mutation or RNAi injection caused early larval arrest ([@pbio-0020257-McKay1]). The effect of *lpd-1* RNAi feeding in our experiments was apparently less severe. The RNAi-treated animals displayed slow growth, morphological abnormalities, and egg-laying defects but no larval arrest. Supplementing C17ISO to the plates did not significantly rescue these defects. LPD-1 and LPD-2 Diverge in Functions {#s2g} ------------------------------------ LPD-2 (C48E7.3) is another C. elegans homolog of a mammalian lipogenic transcription factor, CCAAT/enhancer-binding protein (C/EBP). McKay and coauthors have shown that the *lpd-2(RNAi)* and *lpd-1(RNAi)* phenotypes are quite similar; affected worms are defective in growth, pale and scrawny in appearance, and lacking in fat content ([@pbio-0020257-McKay1]). They have also shown that LPD-1 and LPD-2 control the expression of the same lipogenic enzymes: ACC, FAS, ATP-citrate lyase, and G3PA. We tested to see if LPD-1 and LPD-2 function similarly in the regulation of mmBCFA biosynthesis. In contrast to the result from *lpd-1*(RNAi), the FA composition in *lpd-2(RNAi)* worms was not significantly different from that of wild-type animals even though these animals had a noticeably sick appearance (unpublished data). This result suggested that, in addition to having some common targets, LPD-1 and LPD-2 have distinct functions. LPD-1 is important for production of mmBCFAs as well as other very-long-chain FAs, whereas LPD-2 has no specificity for any particular type of FA. *elo-5* and *elo-6* Are Likely Targets of LPD-1 {#s2h} ----------------------------------------------- The changes in FA composition observed in *lpd-1(RNAi)* would be consistent with downregulation of *elo-5, elo-6* (decrease in mmBCFA), *elo-2* (increase in C16:0) ([@pbio-0020257-Kniazeva1]), and Δ*9-* and/or Δ*12*-desaturase genes (decrease in C18:2 n6). The genes encoding mammalian orthologs of the *C. elegans elo-2* and Δ*9*-desaturase genes are known targets of SREBP-1c ([@pbio-0020257-Edwards1]; [@pbio-0020257-Horton1]; [@pbio-0020257-Horton2]). To examine if *elo-5* and *elo-6* are targets of *lpd-1,* we analyzed the expression of *elo-5, elo-6,* and *lpd-1.* Evaluation of the expression from an *lpd-1Prom*::*GFP* fusion construct (a gift of J. Graff) in transgenic animals revealed that, in addition to the previously reported expression in intestinal cells ([@pbio-0020257-McKay1]), the construct is strongly expressed in a subset of head neurons ([Figure 10](#pbio-0020257-g010){ref-type="fig"}A---[Figure 10](#pbio-0020257-g010){ref-type="fig"}D). Using a lipophilic dye, DiI, which highlights chemosensory ciliated neurons, we identified these neurons as amphids ([@pbio-0020257-Murphy1]). In the strains carrying *elo-5Prom*::*GFP* and *elo-6Prom*::*GFP* reporter constructs, GFP fluorescence was also detected in the gut and several head neurons, including amphid neurons ([Figure 10](#pbio-0020257-g010){ref-type="fig"}E--[10](#pbio-0020257-g010){ref-type="fig"}H and [Figure 2](#pbio-0020257-g002){ref-type="fig"}). ::: {#pbio-0020257-g010 .fig} Figure 10 ::: {.caption} ###### The Expressions of *elo-5* and *lpd-1* Reporter Constructs Are Spatially Similar (A and B) Nomarski and GFP-filtered images of an adult animal containing the *lpd-1Prom*::*GFP* construct, showing strong expression in two symmetrical head neurons, each of which has processes to the nose and around a nerve ring. Scale bars, 10 μm. \(C) DiI staining of amphid neurons in *lpd-1Prom*::*GFP* (dsRed filter). Arrows indicate neuronal nuclei shown in (D). Scale bar, 10 μm. \(D) GFP expression in the animal shown in (C). Scale bar, 10 μm. (E and F) Nomarski and GFP-filtered images of an animal containing *elo-5Prom*::*GFP,* revealing fluorescence in the similar amphid neuron. Scale bar, 7.5 μm. (G and H) The intestinal and intestinal-muscle GFP expression in (G) *lpd-1Prom*::*GFP* and (H) *elo-5Prom*::*GFP* constructs. Scale bar, 7.5 μm. :::  ::: If LPD-1 promotes *elo-5* and *elo-6* expression, then RNAi of *lpd-1* should alter GFP intensity in *elo-5Prom*::*GFP* and *elo-6Prom*::*GFP* reporter strains. The level of GFP expression driven by *elo-5* and *elo-6* promoters is high in conventionally cultured animals. In the worms maintained on the *lpd-1(RNAi)* plates, the expression was noticeably weakened, suggesting a downregulation of the promoter activities ([Figure 11](#pbio-0020257-g011){ref-type="fig"}A--[11](#pbio-0020257-g011){ref-type="fig"}D). No significant changes in GFP expression were detected in a control strain containing a *kqt-1Prom*::*GFP* construct that also expresses GFP in head neurons and the gut (unpublished data). ::: {#pbio-0020257-g011 .fig} Figure 11 ::: {.caption} ###### The Expression of GFP Fusion Constructs Suggests the Involvement of *lpd-1, acs-1,* and *pnk-1* in mmBCFA Biosynthesis (A--D) *elo-5Prom*::*GFP* expression is downregulated in the *lpd-1(RNAi)* background. Scale bars, 100 μm. (A and C) GFP-filtered images of (A) *elo-5Prom*::*GFP* and (C) *elo-6Prom*::*GFP* in wild-type animals, showing the characteristic bright intestinal fluorescence. (B and D) GFP-filtered images of (B) *elo-5Prom*::*GFP* and (D) *elo-5Prom*::*GFP* in *lpd-1(RNAi)* animals, revealing diminished fluorescence in the gut. (E--H) *lpd-1* expression is upregulated in neurons of the *elo-5(RNAi)* animals deficient for C15ISO and C17ISO. Scale bars, 15 μm. (E and F) Nomarski and GFP images of wild-type L1 larvae containing *lpd-1Prom*::*GFP.* (G and H) Nomarski and GFP images of *elo-5(RNAi)*-treated animals (L1 arrested) containing *lpd-1Prom*::*GFP,* showing a visibly brighter fluorescence than that seen in (E) and (F). Circles are centered on the pharyngeal back bulb. (I--K) *acs-1Prom*::*GFP* expression is upregulated in the *elo-5(RNAi)* animals deficient for C15ISO and C17ISO. Panels show GFP images of *acs-1Prom*::*GFP* animals grown on the (I) control, (J) *elo-5(RNAi),* and (K) *lpd-1(RNAi)* plates. The fluorescence from *acs-1Prom*::*GFP* in (J) is significantly stronger than that in (I). Scale bars, 100 μm. (L--N) *pnk-1Prom*::*GFP* expression is upregulated by *elo-5(RNAi)* but downregulated by *lpd-1(RNAi).* Panels show GFP images of *pnk-1Prom*::*GFP* animals grown on the (L) control, (M) *elo-5(RNAi),* and (N) *lpd-1(RNAi)* plates. The fluorescence of the fusion construct is stronger in (M) but weaker in (N) than that in the control (L). Scale bars, 100 μm. :::  ::: To test if the disruption of FAS, a target of LPD-1 ([@pbio-0020257-McKay1]), could contribute to the observed decrease of C15ISO and C17ISO in *lpd-1(RNAi),* we analyzed FA composition in *FAS(RNAi)* strains. There is one predicted FAS gene, F32H2.5, and its shorter homolog, F32H2.6, in the C. elegans genome. The latter can only encode the N-terminal portion of the protein. These genes share extended nucleotide identity, and RNAi of one could thus possibly affect the other. Consistent with a critical role for FAS in the first steps of FA biosynthesis, the RNAi-mediated disruption of F32H2.5 and F32H2.6 resulted in multiple defects and a lethal growth arrest (unpublished data). The FA composition (the content and relative amounts of various FA species) of the affected animals remained, however, unchanged. This suggests that disruption of FAS does not selectively alter FA biosynthesis and that neither FAS protein is specific for mmBCFA. Therefore, downregulation of FAS by loss of *lpd-1* cannot account for the severe deficiency of mmBCFA in *lpd-1(RNAi).* Thus, we have shown that disruption of *lpd-1* affects C15ISO and C17ISO biosynthesis. The fact that *lpd-1, elo-5,* and *elo-6* are expressed in the same cells concurrently and that the GFP reporter analysis indicated that *elo-5* and *elo-6* transcription is downregulated in the absence of *lpd-1* suggests that *elo-5* and *elo-6* are likely to be the targets of *lpd-1.* Since ACC and FAS catalyze the first steps in the biosynthesis of straight-chain FAs while ELO-5 and ELO-6 extend mmBCFA molecules, LPD-1 appears to integrate conventional and "unusual" FA biosyntheses. It seems reasonable to predict that in order to differentiate between these metabolic pathways and mediate compensatory or adaptive changes in FA composition, LPD-1 must interact with other factors such as nuclear receptors activated by specific FA ligands. It is thus important to screen for such interactions to better understand FA homeostasis in *C. elegans.* A Reciprocal Correlation between *lpd-1* Expression and mmBCFA Levels {#s2i} --------------------------------------------------------------------- Because mammalian SREBP-1c regulates polyunsaturated FA biosynthesis and is feedback inhibited by polyunsaturated FAs ([@pbio-0020257-Jump1]), we asked if *lpd-1* could be regulated by mmBCFAs at the transcriptional level. Our microarray data (discussed below) indicated a 1.68-fold upregulation of *lpd-1* in the *elo-5(RNAi)* animals, while no changes were detected in its levels between samples from wild-type animals at different developmental stages (see [Materials and Methods](#s3){ref-type="sec"}). To examine the influence of the mmBCFA deficiency on *lpd-1* expression, we grew the *lpd-1Prom*::*GFP*-containing strain on the *elo-5(RNAi)* and control plates to compare GFP fluorescence. No obvious difference in the GFP expression driven by the *lpd-1* promoter in intestinal cells was detected on the *elo-5(RNAi)* plates versus the control plates. A modest change in the transcription level (1.68-fold) could be masked by a variability of the expression between individual animals and even between individual cells (unpublished data). In contrast to the observation for the intestinal cells, a strong induction of GFP was detected in amphid neurons of *lpd-1Prom*::*GFP*;*elo-5(RNAi)* animals ([Figure 11](#pbio-0020257-g011){ref-type="fig"}E--[11](#pbio-0020257-g011){ref-type="fig"}H). This suggests that a chronic deficiency of mmBCFA in *elo-5(RNAi)* animals may transcriptionally stimulate LPD-1 production, at least in neuronal cells. Collectively, our results suggest that the relationship between *lpd-1* and C15ISO/C17ISO is reciprocal; while downregulation of *lpd-1* transcription results in the C17ISO deficiency, the C15ISO and C17ISO deficiency upregulates *lpd-1* transcription at least in a subset of cells. Therefore, the worm SREBP homolog, LPD-1, may play an important role in mmBCFA homeostasis. Screening for Additional Genes Involved in mmBCFA Homeostasis {#s2j} ------------------------------------------------------------- Because C15ISO and C17ISO play critical roles in animal development and growth, we suspected mechanisms might exist to respond to and regulate their levels. Regulation of mmBCFA homeostasis may involve transcription factors, metabolic enzymes, and transport and binding proteins. It is reasonable to suggest that a deficiency of mmBCFA triggers a compensatory alteration in the expression of these genes. It is also possible that a comparative analysis of global gene expressions between wild-type and mmBCFA-deficient animals may reveal these potential changes and the changes underlying developmental and growth functions of mmBCFA. We used DNA microarray analysis to compare the total gene expression in *elo-5(RNAi)* and wild-type animals. To select candidate genes, we applied restrictive criteria and excluded genes of which the expression was also changed in the *spt-1(RNAi)* strain ([Protocol S1](#sd002){ref-type="supplementary-material"} and [Table S1](#st001){ref-type="supplementary-material"}). The *spt-1* (C23H3.4) gene encodes a predicted C. elegans homolog of serine-palmitoyl transferase subunit 1. RNAi of *spt-1* strongly affects the FA composition without reducing the C15ISO or C17ISO levels (unpublished data). The F1 generation of *spt-1(RNAi)* animals developed gonadal and egg-laying defects that were similar to the phenotype of F1 animals from parents treated with *elo-5(RNAi)* at a late larval stage (described earlier; see [Figure 5](#pbio-0020257-g005){ref-type="fig"}B and [5](#pbio-0020257-g005){ref-type="fig"}F). We thought that by deselecting genes that have altered expressions in *spt-1(RNAi),* we would be able to eliminate variations in gene expressions unrelated to the mmBCFA deficiency. Such variations might emerge from altered straight-chain FA metabolism and from general sickness. Here, we discuss the analysis of the first set of candidate genes that are differentially expressed in *elo-5(RNAi)* and may relate to the C15ISO and C17ISO homeostasis. Twenty-five genes were selected in the screen ([Table 1](#pbio-0020257-t001){ref-type="table"}) and each was functionally tested by RNAi and GC analysis for its role in C15ISO and C17ISO metabolism. RNAi of four of these genes (*pnk-1* \[C10G11.5\], *nhr-49* \[K10C3.6\], *acs-1* \[F46E10.1\], and C27H6.2) significantly affected the FA composition ([Figure 12](#pbio-0020257-g012){ref-type="fig"}). All four genes encoded products structurally homologous to the known proteins (PNK-1, human pantothenate kinase; NHR-49*,* nuclear hormone receptor; ACS-1, very-long-chain FA CoA ligase; and C27H6.2, RuvB-like DNA binding protein). ::: {#pbio-0020257-g012 .fig} Figure 12 ::: {.caption} ###### RNAi of Four Candidate Genes with Altered Expression in *elo-5(RNAi)* Worms Affects the FA Composition \(A) GC profile of the wild type. (B--E) GC profiles of the RNAi-treated worms. (B--D) RNAi of the three genes resulted in a decrease of the C17ISO or both C15ISO and C17ISO levels, indicated by black arrowheads. In addition, a significant elevation in straight-chain saturated FA levels, indicated by grey arrowheads, is observed in K10C3.6(RNAi). \(E) C27H6.2(RNAi) does not cause significant changes in mmBCFA but results in an elevation of straight-chain monounsaturated FA and C18:1 n7, indicated by white arrowheads. Statistical analysis of several GC runs on each of the samples was also carried out (unpublished data). :::  ::: ::: {#pbio-0020257-t001 .table-wrap} Table 1 ::: {.caption} ###### Candidate Genes and Their Encoded Proteins Selected from Microarray Data for Functional Tests (RNAi and GC Analysis) :::  ^a^ Open reading frames (ORFs) predicted by the C. elegans Genome Project (WormBase.org) ^b^ Data from comparing arrays from the experimental sample with that from a baseline control sample (Sample I; see Protocol S1) ::: Analysis of the Candidate Genes {#s2k} ------------------------------- Circumstantial evidence suggests that these four candidate genes may be involved in feedback regulation of mmBCFA biosynthesis. First, the expression of these genes is not variable in nature, as judged by a comparison of the microarray data obtained from developmentally different populations of N2 ([Protocol S1](#sd002){ref-type="supplementary-material"}) as well as for vulval development pathway mutants (data obtained for an unrelated project, J. Chen, personal communication). Second, the direction of the changes for three of the genes is in concordance with the proposed feedback regulation: *pnk-1, nhr-49,* and *acs-1* were upregulated in C17ISO-deficient *elo-5(RNAi)*. Lastly, a functional analysis shows that these three candidate genes are required for the normal level of mmBCFA production (RNAi of the genes affects mmBCFA production). The fourth candidate gene, C27H6.2, affects the level of vaccenic acid (C18:1 n7), which is related to the levels of mmBCFA (see [Figure 8](#pbio-0020257-g008){ref-type="fig"}), suggesting cross-talk. To detect potential feedback regulation involving *acs-1* and *pnk-1,* we made reporter strains with GFP expression driven by *acs-1* and *pnk-1* promoters, *acs-1Prom*::*GFP* and *pnk-1Prom*::*GFP,* respectively. These two genes showed a higher degree of upregulation than the other candidates according to the microarray data. In addition, RNAi of these two genes resulted in a significant loss in the mmBCFA fraction. The GFP fluorescence from *acs-1Prom*::*GFP* and *pnk-1Prom*::*GFP* was readily detectable in the gut. Expression of *acs-1Prom*::*GFP* was also detected in the canal-associated neurons in the head neurons and vulval cells. A comparison of synchronized animals grown on the control and *elo-5(RNAi)* plates indicated a significantly brighter fluorescence in the RNAi worms (see [Figure 11](#pbio-0020257-g011){ref-type="fig"}I, [11](#pbio-0020257-g011){ref-type="fig"}J, [11](#pbio-0020257-g011){ref-type="fig"}L, and [11](#pbio-0020257-g011){ref-type="fig"}M), suggesting upregulation of *acs-1*and *pnk-1* under C15ISO or C17ISO deficiency. These results were in concordance with the microarray data. Moreover, *pnk-1,* but not *acs-1,* seemed to be regulated by LPD-1 because *pnk-1Prom*::*GFP* expression was significantly reduced on *lpd-1(RNAi)* (see [Figure 11](#pbio-0020257-g011){ref-type="fig"}L and [11](#pbio-0020257-g011){ref-type="fig"}N). It was interesting to note that the *pnk-1* and *acs-1* genes were previously selected in two different screens as potential targets of the *daf-2/daf-16* (Y55D5A.5 and R13H8.1, respectively) pathway. *pnk-1* had been identified in a screen for genes affecting the life span and metabolism of C. elegans through analysis of promoter regions, and it was confirmed as a direct target of DAF-16, a forkhead transcriptional factor ([@pbio-0020257-Lee1]). *acs-1* had been identified in a microarray screen for DAF-16 targets that influence life span ([@pbio-0020257-Murphy1]). A third gene, *nhr-49,* had been previously selected in a screen for fat regulatory genes ([@pbio-0020257-Ashrafi1]). It was shown that RNAi of this gene leads to an increase in fat accumulation in affected animals. Our analysis of *nhr-49(RNAi)* animals showed that reduction of *nhr-49* activity results in upregulation of saturated FA biosynthesis that may contribute to fat accumulation. Although the regulatory path for this process remains unknown, the involvement of *daf-2* has not been ruled out. A potential link between the candidate genes and DAF-2/insulin signaling is very intriguing. The C. elegans insulin-signaling pathway is involved in sensing nutritional state and metabolic conditions as well as controlling growth and diapause ([@pbio-0020257-Kimura1]; [@pbio-0020257-Ailion1]). As we report in this paper, a mmBCFA deficiency causes transient L1 arrest. This phenotype strikingly resembles L1 arrest of worms hatched in the absence of food (a method commonly used to obtain synchronized animals). An investigation of possible roles for mmBCFA in food sensation and insulin signaling pathways is underway. Downregulation of the fourth candidate gene, C27H6.2, may result in a significant increase in monounsaturated FA levels ([Figure 12](#pbio-0020257-g012){ref-type="fig"}). This is consistent with the enlarged fraction of monounsaturated FAs observed in the *elo-5(RNAi)* animals (see [Figure 3](#pbio-0020257-g003){ref-type="fig"}C). Downregulation of C27H6.2 may have an adaptive effect to compensate for the loss of mmBCFAs in cell membranes. If so, C27H6.2 may be part of a mechanism that senses and tunes physical properties of membranes. C27H6.2 is homologous to RuvB/TIP49a/Pontin52, an evolutionarily conserved protein essential for growth and proliferation ([@pbio-0020257-Kanemaki1]; [@pbio-0020257-Bauer1]; [@pbio-0020257-Qiu1]). Its mammalian ortholog acts as a transcriptional cofactor that binds to β-catenin, TATA-box binding protein, and likely to a number of other diverse transcription factors ([@pbio-0020257-Bauer1]). Concluding Remarks {#s2l} ------------------ Two mmBCFAs are normally detected in *C. elegans:* C15ISO and C17ISO. A deficiency of these FAs is lethal and cannot be compensated by any other FA present, indicating their crucial importance for growth and development. There are two sources of C15ISO and C17ISO available for worms. First, they possess a system for mmBCFA biosynthesis that includes two FA elongation enzymes, ELO-5 and ELO-6, which are regulated at least in part by the nematode homolog of SREBP-1c *(lpd-1).* Second, worms may obtain mmBCFAs from their diet (bacteria). Therefore, C. elegans is able to produce, activate, transport, and utilize mmBCFAs and is vitally dependent on this system. The level of C15ISO and C17ISO in eggs appears to be critical for growth and development, as animals depleted of C15ISO or C17ISO completely arrest at the L1 stage. The uniformity and reversibility of the arrest would be consistent with a regulatory role in growth and development for these mmBCFAs or for more complex lipid molecules containing them. However, it cannot be ruled out that the arrest is due to the failure of a metabolic or structural function that is essential for growth and development at the first larval stage. In addition, C15ISO and C17ISO may directly or indirectly regulate genes involved in FA homeostasis. Consistent with this, their deficiency triggers a large alteration in gene expression that may reflect a complex feedback mechanism. Among the potentially responsive genes are transcription factors and metabolic genes. Ubiquitous and unattended mmBCFAs come forth as physiologically important molecules that regulate essential functions in eukaryotes. Many interesting questions regarding mmBCFAs remain to be addressed. What are the other components of the mmBCFA biosynthetic machinery? What are the components of their transport system? Does an organism have a mechanism by which the mmBCFA level is measured? What are the signaling pathways involved in the mmBCFA responses? How do mmBCFAs exert their physiological function? Do mmBCFAs act alone or as parts of more complex lipids? How are mmBCFAs synthesized in mammals? Lastly, what are the specific physiological functions of mmBCFAs in mammals? Both genetic and biochemical approaches will be taken to address these questions. Materials and Methods {#s3} ===================== {#s3a} ### RNA interference by feeding {#s3a1} The RNAi feeding vectors were either made in our laboratory using Taq PCR and cloning genomic fragments into a double T7 vector, pPD129.36 (gift of A. Fire), or obtained from the C. elegans whole genome RNAi feeding library (J. Ahringer, MRC Geneservice, Cambridge, United Kingdom). The RNAi feeding strain was E. coli HT115 transformed with either empty pPD129.36 vector (controls) or with dsRNA-producing constructs. Plates were prepared as described in [@pbio-0020257-Kamath1]. Unless stated differently, wild-type N2 Bristol animals were plated as synchronized adults. To obtain synchronized worms of various stages, a large quantity of N2 gravid adults were collected, bleached, and grown to the required stage on HT115 that had been transformed with pPD129.36 (control). ### GC analysis {#s3a2} A mixed population of well-fed worms were washed off the plates with water, rinsed 3--4 times, and, after aspirating away water, frozen at −80 °C. FA methyl esters and lipid extraction were performed as described in [@pbio-0020257-Miquel1]. GC was performed on the HP6890N (Agilent, Palo Alto, California, United States) equipped with a DB-23 column (30 m × 250 μm × 0.25 μm) ([@pbio-0020257-Kniazeva1]). Each experiment was repeated at least five times. Average values and standard deviations were then calculated for each of the compounds in the experiments. ### Staging worms to test for FA composition {#s3a3} After bleaching gravid adults, an aliquot of the eggs was set apart, and the rest was incubated overnight in M9 at room temperature. On the next day, an aliquot of L1 was frozen for GC analysis. The rest of L1 was plated on agar plates. Subsequently, L2, L3, L4, young adults, and adults along with hatched L1 were collected as the separate samples. Mixed populations of worms starved for 24--100 h were included in the experiment to monitor a possible effect of the starvation. ### Phenotype rescue using FA supplements. {#s3a4} Ninety microliters of the 4 mM solution of FA (Sigma, St. Louis, Missouri, United States) in 1% NP40 was dropped on the side of the bacterial lawn that contained either *elo-5* dsRNA-producing plasmid or the control HT115 vector. Two synchronized young adults were plated and their progeny was scored 4 and 5 d later. Each experiment was performed in at least 30 replicates. For recovering *elo-5(RNAi)* worms from L1 arrest, wild-type adults were placed on the *elo-5(RNAi)* plates. Four days later, their progeny was removed and eggs of the next generation were left on the plates. Hatched L1 were kept for 2 or 4 d before transferring as agar chunks to new *elo-5(RNAi)* plates. FA supplements were added to spots next to the chunks. Ten plates were prepared for each FA supplement. Control plates contained no supplements. To verify that an addition of supplements did not affect RNA interference per se, we used *let-418(RNAi)* animals, which have a sterile phenotype, as a control. Neither C15 nor C17 mmBCFA added to *let-418(RNAi)* plates modified the expected phenotype. ### Designing of GFP reporter constructs. {#s3a5} To prepare the GFP fusion constructs, genomic fragments were PCR amplified and cloned in frame into one of the GFP fusion vectors (gift of A. Fire). The locations of the genomic fragments and PCR primers used are listed below: \(1) *elo-5Prom*::*GFP,* starting at 3.894 kb genomic upstream of the first codon and ending 4 bp into the first exon; primers, F-BamHI-TTTAGGTCATTTTTTGAGTCGCCA and R-BamHI-TAGTCTGGAATTTTGAAATTGAACGG; vector, pPD95.69; \(2) *elo-6Prom*::*GFP,* a 4.764-kb fragment covering 3,104 bp upstream and 1,660 bp downstream of the predicted start codon and ending 14 bp into the third exon; primers, F-Sph1-GCCCTTGGAAACCATCTACGACGAATC and R-Sma1-TCCGAACAGAACGACATAAGAGATTTCC; vector, pPD95.77; \(3) *acs-1Prom*::*GFP,* a 3.142-kb genomic fragment containing 3,048 kb upstream of the first predicted ATG and ending 24 bp into the second predicted exon; primers, F-SphI-CATAATTACTATTGCGTCACATG and R-SphI-CTCTTCCAAACTGGCGATGTCGA; vector, pPD95.69; \(4) *pnk-1Prom*::*GFP,* a 1.14-kb fragment that includes 937 bp upstream of the first predicted codon of the C10G11.5 and 203 bp downstream, ending 24 bp into the second exon; primers, F-SphI-TCGTACGATCGGACCATAGGCTAA and R-SphI-CTGATCCTCTGTAGCAGCGGCCCT; vector, pPD95.69. These constructs were injected into C. elegans at 10--50 ng/μl to form extrachromosomal arrays. In the case of *acs-1,* the extrachromosomal array had been integrated into the C. elegans genome. ### Staining chemosensory ciliated neuron with DiI {#s3a6} Worms were soaked in a 5-μg/ml solution of DiI (Molecular Probes, Eugene, Oregon, United States) in M9 buffer for 1 h. They were then rinsed three times with M9 and visualized by fluorescence using the Texas Red filter. ### Correlation analysis {#s3a7} The FA quantities obtained by GC were expressed as a percentage of the total. *t* test (two-tailed distribution) and correlation analysis were performed using the Microsoft Excel program. ### Visualization and scoring of the GFP expression in promoter::GFP lines {#s3a8} Synchronized adults were placed on control (HT115 bacterial strain transformed with empty vector, pPD129.36) and RNAi (HT115 bacterial strain transformed with dsRNA construct) plates. Several worms of the next generation were picked from the control and RNAi plates and mounted on the same microscopic slide. GFP images were obtained with the fixed settings and exposure. ### Microarray analysis {#s3a9} One young adult of the N2 Bristol strain was plated on each control and RNAi feeding plate. Control plates contained the E. coli HT115 strain transformed with empty pPD129.36 vector. Experimental RNAi plates contained E. coli HT115 transformed with corresponding dsRNA constructs. The growth conditions, RNA preparations, and data analyses are described in detail in [Protocol S1](#sd002){ref-type="supplementary-material"}. Expression data are presented in [Dataset S1](#sd001){ref-type="supplementary-material"}. Supporting Information {#s4} ====================== Dataset S1 ::: {.caption} ###### Microarray Expression Data (1.8 MB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S1 ::: {.caption} ###### Expression of Collagen Genes as an Indicator of Developmental Differences in Mixed Populations of Worms Samples I, II, and III represent mixed populations of wild-type animals started simultaneously from one young adult. Each was harvested at three time points, when mostly adults represented the F1 generation and the embryos and larvae in different proportions represented the F2 generation (see [Protocol S1](#sd002){ref-type="supplementary-material"}). Sample III corresponds to the most diverse mixture of worms. Numbers of collagen genes that were differentially expressed between pairs of samples are shown above or bellow the arrow brackets. Sample I and Sample III, which originated from the most distal mixed populations, have the largest number of differentially expressed collagens. Sample I and an experimental sample corresponding to the *elo-5(RNAi)* phenotype have a lower number of the changed collagen genes, suggesting that populations on these experimental and control plates are similar. (24 KB PPT). ::: ::: {.caption} ###### Click here for additional data file. ::: Protocol S1 ::: {.caption} ###### Microarray Data Analysis (37 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S1 ::: {.caption} ###### Filtering Candidate Genes by Comparing Different Mutant and Wild-Type Samples (28 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: We thank J. Graff for the *lpd-1*::*GFP* strain, A. Fire for vectors, J. Ahringer and MRC Geneservice for the RNAi library, and A. Kniazev for technical help with microarray data processing. We thank T. Kharchenko, E. Kim, and K. Hirono for assistance, W. Wood and members of our laboratory for discussions, and E. Seaman, D. Eastburn, E. Kim, and reviewers for comments on the manuscript. This project is supported by the Howard Hughes Medical Institute, of which MK is a specialist and MH is an Associate Investigator, and by National Institutes of Health grants R01DE13848 and R01DE13335 to CYW. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** MK and MH conceived and designed the experiments. MK, QTC, and MS performed experiments. MK analyzed the data. CYW contributed reagents/materials/analysis tools. MK and MH wrote the paper. Academic Editor: Paul W. Sternberg, California Institute of Technology Citation: Kniazeva M, Crawford QT, Seiber M, Wang CY, Han M (2004) Monomethyl branched-chain fatty acids play an essential role in Caenorhabditis elegans development. PLoS Biol 2(9): e257. ACC : acetyl-CoA carboxilase BCAT : branched-chain aminotransferase BCKAD : branched-chain α-keto-acid dehydrogenase CoA : coenzyme A dsRNA : double-stranded RNA FA : fatty acid FAS : fatty acid synthetase G3PA : glycerol 3-phosphate acyltransferase GC : gas chromatography GFP : green fluorescent protein mmBCFA : monomethyl branched-chain fatty acid SI : saturation index SREBP : sterol regulatory element binding protein

PubMed Central

2024-06-05T03:55:47.249529

2004-8-31

PMC514884

Introduction {#s1} ============ Exonic splicing enhancers (ESEs) were identified about a decade ago as short oligonucleotide sequences that enhance exon recognition by the splicing machinery (reviewed in [@pbio-0020268-Blencowe1] and [@pbio-0020268-Cartegni1]). Sequences with ESE activity have been identified in both plants and animals and have been found to occur frequently in constitutively spliced exons as well as alternatively spliced exons ([@pbio-0020268-Tian1]; [@pbio-0020268-Coulter1]; [@pbio-0020268-Liu1]; [@pbio-0020268-Schaal1]; [@pbio-0020268-Fairbrother1]). ESEs often mediate their effects on splicing through the action of proteins of the SR protein family, which bind to ESEs and recruit components of the core splicing machinery to nearby splice sites ([@pbio-0020268-Graveley1]). Previously, we reported a computational method called relative enhancer and silencer classification by unanimous enrichment (RESCUE)-ESE which identifies ESEs in human genomic sequences using statistical properties of the oligonucleotide composition and splice site strengths of large datasets of exons and introns ([@pbio-0020268-Fairbrother1]). This method identified a set of 238 hexamers (of the 4,096 possible hexamers) which were predicted to possess ESE activity on the basis that (1) they are significantly enriched in human exons relative to introns and (2) they are significantly more frequent in exons with weak (nonconsensus) splice sites than in exons with strong (consensus) splice sites. Tests of splicing enhancer activity using an in vivo splicing reporter system confirmed ESE activity for a representative sequence from each of ten clusters of RESCUE-ESE hexamers ([@pbio-0020268-Fairbrother1]). The function of this set of hexamers was further confirmed by the observation that ESE activity was reduced significantly in nine out of ten point mutants chosen to eliminate RESCUE-ESE hexamers and the observation that the set of RESCUE-ESE hexamers was also predictive in analyzing a list of published mutations that cause exon skipping in the human hypoxanthine phosphoribosyl transferase gene ([@pbio-0020268-Fairbrother1]). A variety of other selection-based methods have been used to identify sets of sequences that are capable of functioning as ESEs. These SELEX methods isolate ESEs from a complex pool of random sequence by iteratively selecting and amplifying the fraction of molecules that can function as ESEs in a reporter assay ([@pbio-0020268-Tian1]; [@pbio-0020268-Coulter1]; [@pbio-0020268-Liu1]; [@pbio-0020268-Schaal1]). These methods have yielded a variety of sequence motifs, and ESE activity of representative sequences has been demonstrated in reporter systems. Often, these motifs have not been refined to a degree where it is possible to reliably design single point mutations that disrupt ESE function ([@pbio-0020268-Tian1]; [@pbio-0020268-Coulter1]; [@pbio-0020268-Liu1]; [@pbio-0020268-Schaal1]). Despite this, a few previous studies have identified several disease alleles where the disruption of a conserved splicing enhancer corresponds to observed splicing defects, a noteworthy example being splicing mutations in the breast cancer gene BRCA1 ([@pbio-0020268-Liu2]; [@pbio-0020268-Orban1]). To date, this type of analysis has been limited to only a few genes. While mutational studies on model splicing substrates have proven an effective means of characterizing individual ESEs, the ability to draw general conclusions about ESE function has been complicated by additional features that vary between substrates. Features such as transcript secondary structure, adjacent negative elements, and the possible contribution of splicing factors associated with the transcription machinery could all modulate ESE activity on a given substrate and prevent any single substrate from serving as a paradigm for all aspects of exon recognition problems. To dilute the contribution of sequence context and to develop general rules for splicing, we have used genomic data to survey the strength of selection on RESCUE-ESE hexamers 2 across several thousand exons. Here we test the hypothesis that, in addition to protein-coding requirements, human exon sequences are also significantly constrained by the requirement to encode ESEs. We have developed a population genetic approach (variant elimination reinforces functionality \[VERIFY\]) which exploits the simple principle that, because they are selected against, deleterious alleles will tend to be underrepresented in the pool of sequence variants that are common in a population ([@pbio-0020268-Graur1]). Taking advantage of the huge repository of genetic information represented by the human single nucleotide polymorphism (SNP) database, we determined the ancestral allele for exonic SNPs by comparison to the chimpanzee genome. This information makes it possible to distinguish between SNPs derived from mutations that disrupt predicted ESEs and those derived from mutations that create predicted ESEs, allowing an assessment of the degree of selection to conserve ESEs in human exons. While we have tested a small subset of RESCUE predictions in a functional assay, many sequences proposed to possess ESE activity have not been validated. This work provides additional evidence that RESCUE-ESE hexamers are physiologically important when they occur in human exons and suggests a way to use SNP data to validate oligonucleotide motifs proposed to have biological activity. Results/Discussion {#s2} ================== To assess the relationship between the locations of common genetic variation in human genes and ESEs, we screened the public SNP database (dbSNP; build 112) for SNPs that mapped to human exons. A set of biallelic reference SNPs was used, excluding entries that (1) mapped to multiple regions in the human genome, (2) mapped to repetitive elements, or (3) were derived from transcript sequence data, e.g., through comparison of expressed sequence tags (ESTs) (see [Materials and Methods](#s3){ref-type="sec"} for details). The remaining SNPs were searched against a large database of human genes containing approximately 121,000 internal exons annotated by aligning available human cDNAs to the assembled genome using the GENOA genome annotation system (see [Materials and Methods](#s3){ref-type="sec"}). This search identified 9,862 SNPs that were localized to an internal exon (aligned perfectly to the genomic sequence over a 33-base segment centered on the polymorphic position). ESE Density Is Highest and SNP Density Lowest near Splice Sites {#s2a} --------------------------------------------------------------- Recording the position of each SNP within the corresponding exon revealed that SNP density is not uniform along exons ([Figure 1](#pbio-0020268-g001){ref-type="fig"}). Consistent with previous observations ([@pbio-0020268-Majewski1]), SNP density was approximately 20--30% lower near both the 3′ splice site (3′ss) and the 5′ splice site (5′ss) of human exons than in the interior of exons, and reached a plateau at about 25--30 bases from the splice sites. The distribution of RESCUE-predicted ESE hexamers along exons had roughly an inverse relationship to the SNP density, with the highest density of ESEs observed near the 5′ and 3′ splice junctions and a lower density in the interior of exons ([Figure 1](#pbio-0020268-g001){ref-type="fig"}). Previously, ESE activity has been observed to vary as a function of the distance between the ESE and the adjacent splice sites, with the highest activity in vitro and in vivo for ESEs positioned closest to splice sites ([@pbio-0020268-Nelson1]; [@pbio-0020268-Lavigueur1]; [@pbio-0020268-Graveley2]). Thus, selective pressure is likely to be higher on ESEs located near splice junctions relative to ESEs in the interior of exons, which could explain the trend in ESE density shown in [Figure 1](#pbio-0020268-g001){ref-type="fig"}. As a consequence of the increased density of ESEs near splice sites, mutations that occur in exons near splice sites should have a higher likelihood of disrupting ESEs and therefore be more likely to be eliminated by purifying selection. Thus, selection on ESEs could potentially explain the trend in SNP density seen in [Figure 1](#pbio-0020268-g001){ref-type="fig"}. In order to more directly test the hypothesis that ESE disruption mutations are subject to negative selection, we conducted a large-scale analysis of sequence variation in human exons. ::: {#pbio-0020268-g001 .fig} Figure 1 ::: {.caption} ###### Density of Predicted ESEs and SNPs along Human Exons RESCUE-ESE hexamers were searched against a database of 121,000 internal human exons. ESE density (blue curve) was determined as the fraction of hexamers beginning at the given exon position in this dataset that were contained in the RESCUE-ESE set. SNP density (red curve) was determined analogously using SNPs from dbSNP mapped to the exon database. Both curves were smoothed by averaging the densities over a leading (3′ss) or lagging (5′ss) window of ten nucleotides. :::  ::: Estimating the Frequency of ESE Disruption in Human SNPs {#s2b} -------------------------------------------------------- The frequency of ESE disruption for simulated (randomly generated, unselected) mutations was compared to the frequency of ESE disruption observed in SNPs, which represent mutations that have survived selection to become reasonably frequent in the human population. SNPs are shaped by the interplay between the mutation process and the process of natural selection. Comparing simulated mutations to natural variations is an effective way to decouple these processes, and this approach has been used by others to study selective pressure in protein-coding genes ([@pbio-0020268-Gojobori1]; [@pbio-0020268-Nei1]; [@pbio-0020268-Kowalczuk1]). Here, we describe the analysis of simulated and natural mutations using a variation of the McDonald--Kreitman test, a widely used statistical test for detecting selection in genes, that we have adapted to measure the strength of selection acting on ESEs using data from several thousand exons ([@pbio-0020268-McDonald1]; [@pbio-0020268-Jenkins1]). When considering SNPs, in order to distinguish mutations that disrupt predicted ESEs from those that create predicted ESEs, the identity of the ancestral allele must be established. Since the mutations that created most human polymorphisms occurred less than 1 million years ago ([@pbio-0020268-Slatkin1]; [@pbio-0020268-Miller1]), long after the human--chimpanzee divergence of 5 million years ago ([@pbio-0020268-Stauffer1]), the orthologous chimpanzee exon will almost always represent the sequence of the ancestral allele. Each of the 9,862 mapped human SNPs described previously was aligned (using a 33-nucleotide sequence window centered around the polymorphic position) to unassembled reads from the genome of the chimpanzee *Pan troglodytes,* accessed through the NCBI trace archives (<ftp://ftp.ncbi.nih.gov/pub/TraceDB/pan_troglodytes/>). As the trace archives represented several-fold coverage of the chimp genomic sequence, most SNPs matched to several sequence reads. Whenever one allele of a human SNP consistently matched the chimpanzee sequence in all high-quality alignments, that allele was designated as the ancestral allele, and the other allele at that position was designated as a variant allele. The 8,408 SNPs that satisfied this criterion were then annotated for predicted ESEs in both the ancestral and variant sequence, simply by comparing the six overlapping hexanucleotides that differed between the two alleles to the set of RESCUE-ESE hexamers and recording the number of matching hexamers. It is well known that current SNP databases contain a certain rate of error. The SNP consortium estimated that about five percent of their submissions were false positives attributed to base calling errors ([@pbio-0020268-Altshuler1]). Incorrect mapping can also result in the misclassification of nearly identical paralogous regions as SNPs ([@pbio-0020268-Bailey1]; [@pbio-0020268-Cheung1]). The rate of false positives in dbSNP can be conservatively estimated by resequencing DNA that has been collected from many individuals. SNPs that cannot be validated in such a manner are either rare SNPs that were not present in the sample or are false positives of the SNP discovery method. Recent resequencing studies validate 60--86% of the entries in dbSNP ([@pbio-0020268-Carlson1]; [@pbio-0020268-Reich1]). We anticipated a lower rate of false positives in the 8,408 SNPs that were used in this analysis because we removed error-prone categories of SNPs (such as those derived from ESTs or duplicated regions) from our data set. Despite this expected improvement, our initial analysis focused on the subset of 2,561 SNPs that had been validated by resequencing and were thus assumed to be free of errors (see [Materials and Methods](#s3){ref-type="sec"}). This precaution was taken because the measurement of ESE disruption is particularly sensitive to artifacts (unpublished data). The annotation of RESCUE-ESE hexamers in a biallelic SNP results in one of four possible outcomes: no ESE hexamers in either allele (ESE neutrality, − −), one or more ESE hexamers only in the ancestral allele (ESE disruption, + −), one or more ESE hexamers only in the variant allele (ESE creation, − +), or one or more ESE hexamers in both alleles (ESE alteration, + +). The latter category is referred to as ESE alteration because the sets of RESCUE-ESE hexamers in the ancestral and variant alleles are, of course, different, and therefore may not necessarily be recognized with the same affinity by the same *trans*-factor(s). The relative frequencies of these four outcomes are listed in [Figure 2](#pbio-0020268-g002){ref-type="fig"}A in the row labeled "Selected (SNP)." Since the 238 RESCUE-ESE hexamers represent only a small fraction (approximately 6%) of the 4,096 possible hexanucleotides, it was not surprising that a large majority of SNPs fell into the ESE-neutral category. To determine whether the rate of ESE disruption in SNPs was higher or lower than what would be expected from unselected mutations, we performed a Monte Carlo (random) simulation of point mutations in human exons. ::: {#pbio-0020268-g002 .fig} Figure 2 ::: {.caption} ###### Analysis of the Effects of SNPs and Unselected Mutations on Predicted ESEs \(A) The percentages of the four prediction outcomes. ESE disruption (+ −), ESE alteration (+ +), ESE neutrality (− −), and ESE creation (− +) changes are listed for the set of 2,561 validated SNPs (selected) and for the set of 100,000 simulated (unselected) mutations. \(B) Synonymous and nonsynonymous mutations were analyzed separately and then compared using the MH test for homogeneity. All outcomes passed the MH test for homogeneity (H~0~:Outcome~synon~ ≈ Outcome~nonsynon~; *p* \< 0.05) and could, therefore, be combined into a summary OR (weighted combination of the ORs measured in the synonymous and nonsynonymous sets). The height of each bar can be interpreted as the odds that the listed outcome will occur in the evolutionarily selected set of mutations (SNPs) relative to the odds that the same outcome will occur in the unselected (simulated mutation) set. Error bars extend one standard deviation on either side of the calculated value. :::  ::: Reduced Frequency of ESE Disruption in SNPs versus Unselected Changes {#s2c} --------------------------------------------------------------------- Mutations were simulated at random in human exons using nucleotide substitution frequencies that reflect the mutational biases observed in unselected regions of the genome with similar nucleotide composition. In vertebrates, the greatest biases in nucleotide-to-nucleotide substitution frequencies are related to the hypermutability of C residues ([@pbio-0020268-Duncan1]) and the higher rate of transitions (C ↔ T, A ↔ G) relative to transversions (purine ↔ pyrimidine). Rates for different base changes have been estimated from analysis of aligned sequences which are assumed to be under no selective pressure. For example, transitions make up 67--70% of all substitutions in human pseudogenes and 65.5% of all substitutions in genomic repeat sequences ([@pbio-0020268-Graur1]; [@pbio-0020268-Hardison1]; [@pbio-0020268-Zhang1]). Using the substitution frequencies derived from a recent large-scale study of nucleotide substitution patterns in processed pseudogenes ([@pbio-0020268-Zhang1]), mutations were simulated in the set of approximately 121,000 GENOA-annotated internal human exons and the effects on predicted ESEs were analyzed as described above for SNPs. The simulation captured the influence of nearest-neighbor bases on the pattern and rate of nucleotide substitution. These nearest-neighbor effects are particularly pronounced for CpG dinucleotides, where an elevated C-to-T mutation rate is observed as an indirect consequence of cytosine hypermethylation. Comparing the results of this simulation to those observed for SNPs (listed as "Selected (SNPs)" in [Figure 2](#pbio-0020268-g002){ref-type="fig"}A), the most striking difference was observed for the category of ESE disruption: 13.6% of simulated (unselected) mutations caused ESE disruption compared to only 10.9% of SNP mutations. This difference implies a significant selective disadvantage for mutations that disrupt RESCUE-ESE hexamers. To assess the degree of selection on ESEs, one standard measure is the relative risk (RR), defined in this instance as the ratio of the frequency of ESE disruption in SNPs to the frequency of ESE disruption for unselected mutations (e.g., RR = 10.9%/13.6% = 0.80, using the pooled data from [Figure 2](#pbio-0020268-g002){ref-type="fig"}A). In this instance, we preferred to use the slightly more complex odds ratio (OR) measure to quantify this effect (defined in [Materials and Methods](#s3){ref-type="sec"}) because of its better statistical properties ([@pbio-0020268-Pagano1]). As expected, the SNP dataset was greatly enriched for synonymous variation relative to the simulated mutation dataset. There is a 1.3:1 ratio of synonymous:nonsynonymous changes in the SNP dataset compared to a 0.5:1 ratio of synonymous:nonsynonymous changes in the simulated dataset. This difference, which has been observed many times, suggests that more than 60% of mutations that change the amino acid sequence are eliminated by natural selection ([@pbio-0020268-Graur1]). In order to account for the potentially confounding effect of selection occurring at the protein level, SNPs and simulated mutations were divided into synonymous and nonsynonymous groups and analyzed separately. After controlling for the higher frequency of synonymous mutations in dbSNP, the selective pressure to avoid disrupting ESEs was approximately equal for the synonymous and nonsynonymous classes of mutations (the Mantel--Haenzel \[MH\] test of homogeneity indicated no significant differences in the magnitude of the effect across all comparisons; χ^2^ \< 0.5). This observation confirms our previous result and alleviates the concern that the analysis might have been confounded by the effects of selection acting at the protein level. The summary OR (the weighted combination of the separate synonymous and nonsynonymous analysis) for ESE disruption was 0.82 ± 0.05 ([Figure 2](#pbio-0020268-g002){ref-type="fig"}B), which implies that natural selection has eliminated approximately 18% of arising point mutations that disrupt RESCUE-ESE hexamers (*p* \< 0.001). Base changes that alter one or more predicted ESE hexamers but result in creation of other RESCUE-ESE hexamers (ESE alteration) are also selected against, but to a somewhat lesser degree ([Figure 2](#pbio-0020268-g002){ref-type="fig"}B, "+ +" category). This observation is not surprising given that these changes may alter the specific combination of SR proteins which interact with the exon and consequently alter ESE activity. In our previous study, we found that ESE alteration mutations would often cause an increase or decrease in enhancer activity, as determined in vivo using a splicing reporter construct ([@pbio-0020268-Fairbrother1]). SR proteins generally have distinct, though sometimes overlapping, RNA binding specificities and vary in their ability to activate splicing ([@pbio-0020268-Graveley2]; [@pbio-0020268-Liu1]). Therefore, some ESE alteration mutations that result in one SR protein replacing another SR protein may weaken the ESE and disrupt splicing. In addition, there are situations where the simultaneous binding of multiple activator proteins on a substrate is critical for correct processing of that pre-mRNA ([@pbio-0020268-Tian2]). In such a case, it is unlikely that one ESE could be exchanged for another without deleterious consequences. At first glance, the overrepresentation of some categories in this analysis may seem surprising. However, this is simply a consequence of the underrepresentation of disruption and alteration mutations in the SNP pool causing the remaining two categories of variation, ESE neutrality and ESE creation, to appear slightly overrepresented in SNPs relative to unselected mutations ([Figure 2](#pbio-0020268-g002){ref-type="fig"}B). As mutations that result in a selective disadvantage are rapidly eliminated from the population, an increasing fraction of the mutations that persist as SNPs will be selectively neutral ([@pbio-0020268-Graur1]). In other words a neutral variant will, on average, be eliminated less rapidly than a disadvantageous variant and so the set of neutral variations will come to represent an increasing fraction of the total (diminishing) pool of variants. The analysis presented here divides SNPs into four categories based on ESE annotation. As variations from two of these categories (ESE disruption and ESE alteration) are shown to be preferentially eliminated by natural selection, the remaining two categories (ESE neutrality and ESE creation) will represent a larger fraction of a diminished total pool and, therefore, appear enriched ([Figure 2](#pbio-0020268-g002){ref-type="fig"}B). Selective Pressure Is Strongest for ESEs Located near Splice Sites {#s2d} ------------------------------------------------------------------ Experiments measuring the splicing activity of substrates with variations in the distance between a well-characterized ESE and the 3′ss have demonstrated a strong proximity effect, with ESE activity decreasing as the distance from the splice site increases ([@pbio-0020268-Lavigueur1]; [@pbio-0020268-Graveley2]). Although the closest ESEs tested in the distance studies were 70 nucleotides away from the splice sites, other studies have demonstrated that ESEs can function at much closer distances to splice sites ([@pbio-0020268-Nelson1]; [@pbio-0020268-Coulter1]). In order to test the generality of this distance effect, we quantitated the selective pressure on ESEs in distal and proximal windows at both the 3′ss and the 5′ss ([Figure 3](#pbio-0020268-g003){ref-type="fig"}). Validated SNPs that fell within a particular exon region (e.g., the first 20 nucleotides of the 3′ss proximal window is defined as region A in [Figure 3](#pbio-0020268-g003){ref-type="fig"}) were compared to unselected (simulated) mutations that fell in the same region, and summary ORs for ESE disruption were calculated for the four exon regions shown. Consistent with higher ESE activity for ESEs located near splice sites, we observed a pronounced increase in the conservation of RESCUE-ESE hexamers located within 20 bases of either the 5′ss or 3′ss (*p* \< 0.05) relative to predicted ESEs located further from splice sites ([Figure 3](#pbio-0020268-g003){ref-type="fig"}). ::: {#pbio-0020268-g003 .fig} Figure 3 ::: {.caption} ###### Selection against Disruption of Predicted ESEs in Different Exon Regions Summary ORs were calculated for mutations that disrupt RESCUE-ESEs as in [Figure 2](#pbio-0020268-g002){ref-type="fig"}, for each of four regions spanning the length of a typical human internal exon. The heights of the blue bars represent the odds that an ESE will be disrupted by a mutation in the set of 2,561 validated SNPs (selected mutations) relative to the odds of disruption in the set of 100,000 simulated (unselected) mutations. Error bars extend one standard deviation on either side of the calculated value. :::  ::: SNP Analysis Identifies Conserved ESEs {#s2e} -------------------------------------- Our observations that SNPs tend to avoid disrupting RESCUE-ESE hexamers (see [Figure 2](#pbio-0020268-g002){ref-type="fig"}) and that the magnitude of this selection increases near splice sites ([Figure 3](#pbio-0020268-g003){ref-type="fig"}) indicate that this set of hexamers represents a physiologically important collection of sequences across many human exons and genes. Although all RESCUE-ESE hexamers tested to date have ESE activity in cell culture assays, we have tested only a small fraction of the 238 individual hexamers, and presumably some members of this set may be false positives of the RESCUE-ESE method. In order to better define functional hexamers in the RESCUE-ESE set, we repeated the selected versus unselected comparisons for each hexamer individually using the larger set of 8,408 exonic SNPs described previously. For each RESCUE-ESE hexamer, we counted cases where a hexamer was interrupted by a mutation that has survived selection (SNP mutations) and compared this frequency to the value we would expect in the absence of selection (simulated mutations). For this analysis we used a simple RR measure, defined as the frequency with which a hexamer overlaps with SNPs relative to the frequency with which the same hexamer overlaps with simulated mutations. This ratio (frequency~SNP~/frequency~simulated~) was calculated for each hexamer and provided a means of assessing the selective pressure on each hexamer. The RR for a hexamer that was under no additional selective pressure would therefore be equal to 1.0, and a hexamer under increased selection would have an RR of less than 1. Consistent with our previous analysis (see [Figure 2](#pbio-0020268-g002){ref-type="fig"}), the majority of RESCUE-ESE hexamers (162 hexamers) had an RR of less than 1 ([Figure 4](#pbio-0020268-g004){ref-type="fig"}A), suggesting that many RESCUE-ESE hexamers are subject to purifying selection. ::: {#pbio-0020268-g004 .fig} Figure 4 ::: {.caption} ###### Measuring Selective Pressure on Each RESCUE-ESE Hexamer Any point mutation alters six overlapping hexamers, and so a database of 8,408 SNP mutations alters a total of approximately 50,000 hexamers in the wild-type (ancestral) allele. In considering all 238 RESCUE-ESE hexamers, the frequency of each ESE hexamer in the total set of ancestral alleles was recorded for the database of SNPs and simulated mutations (8,408 SNP mutations and 100,000 simulated mutations). The ESE frequency in the SNP set was divided by the ESE frequency in the simulated set to calculate the RR for each of the 238 hexamers. \(A) The distribution of RR for all 238 ESE hexamers is plotted on a logarithmic scale. A resampling strategy was used to identify 57 ESE hexamers that were significantly conserved (pink bars have an RR less than 1; *p* \< 0.05) and also six ESE hexamers that were not conserved (blue bars have an RR greater than 1; *p* \< 0.05). \(B) The output of RESCUE-ESE was compared for several vertebrate genomes (human, mouse, pufferfish, and zebrafish). The set of 238 human RESCUE-ESE hexamers was divided into nonoverlapping subsets based on their conservation in the RESCUE-ESE output generated from other vertebrates. The proportion of ESEs that were significantly conserved in the SNP analysis (as described above in \[A\]) were recorded for each subset of RESCUE-ESE hexamers and are represented as pink sectors in the pie chart. :::  ::: While the somewhat limited size of the currently available SNP databases limits our power to detect selection acting on individual hexamers, it was possible to detect a subset of hexamers that displayed a statistically significant level of conservation. A bootstrap sampling strategy identified a total of 57 hexamers with RRs of significantly less than 1 ([Figure 4](#pbio-0020268-g004){ref-type="fig"}A, pink bars) compared to only six hexamers with RRs significantly greater than 1 ([Figure 4](#pbio-0020268-g004){ref-type="fig"}A, blue bars). By comparison, testing a set of 238 arbitrary hexamers would be expected to yield approximately 12 significant hexamers in each of these categories at a *p* value cutoff of 0.05. Included within the set of 57 conserved ESEs are hexamers corresponding to the well-characterized "GAR" and AC-rich ESE classes ([@pbio-0020268-Coulter1]) and several other types of ESEs (see [Supporting Information](#s4){ref-type="sec"}). A comparison of RESCUE-ESE predictions for four different vertebrates (G. Yeo, S. Hoon, B. Venketesh, and C. B. B., unpublished data) revealed a strong correspondence between within-species conservation (SNP analysis) and cross-species conservation ([Figure 4](#pbio-0020268-g004){ref-type="fig"}B). In other words, the ESE hexamers that appeared to be under the greatest selective pressure within the time frame of the SNP analysis (the last hundreds of thousands of years) were more likely to retain the characteristics used by the RESCUE approach to identify ESEs over the time frame of vertebrate speciation (the last tens to hundreds of millions of years). While only a minor fraction (11%) of the hexamers that appear exclusively in the human lineage were significantly conserved in the SNP analysis (RR \< 1; *p* \< 0.05), about half of the hexamers predicted to be ESEs in all four vertebrates examined (human, mouse, zebrafish, and pufferfish) were significantly conserved in the SNP analysis. In the future, with a larger SNP database and more genomes available, it should be possible to use these methods to analyze more individual hexamers for evidence of selective pressure. As mentioned previously, the public SNP database used in this study contains a significant fraction of entries that could not be validated by resequencing. If these unvalidated entries in dbSNP were errors, they would not be expected to specifically avoid ESEs, or particular exon positions. Therefore, SNP artifacts are likely to reduce, rather than increase, the apparent significance of the biases in the nature and distribution of SNPs that we have observed. Here we have shown, using polymorphism data, that RESCUE-ESE hexamers have been preferentially conserved in the recent evolutionary history of the human lineage and that the strength of this conservation increases with increasing proximity to the splice sites. These results imply that splicing imposes important constraints on the evolution of human exons. As the size and quality of the SNP database is rapidly increasing, measures of selection deriving from SNP data are likely to become increasingly useful for evaluating the function of short, degenerate sequence elements like ESEs. For example, it should be possible to use the VERIFY method to analyze selection on motifs that are postulated to control transcription, polyadenylation, messenger RNA (mRNA) stability, or translation. The current build of dbSNP contains 360,000 SNPs that are located in a "locus region" within 2 kb upstream of the transcript start site or 500 bp downstream of the polyadenylation site ([@pbio-0020268-Sherry1]). There are also more than 500,000 SNPs localized to the untranslated regions of mRNAs where elements that modulate translation, polyadenylation, and mRNA stability are thought to be located. The analysis presented here used 8,400 SNPs to study selection on ESE hexamers. Making the simplifying assumptions that ESE hexamer frequencies are both uniform and slightly elevated in exons and are mutated without bias, we can roughly estimate the statistical power of the VERIFY method to detect selection (see [Materials and Methods](#s3){ref-type="sec"}). For example, the set of ESE hexamers, as a whole, is under selective pressure (OR = 0.82), but this selective pressure will vary according to hexamer, with some hexamers being highly conserved and others less so ([Figure 4](#pbio-0020268-g004){ref-type="fig"}). Our ability to detect conservation at the level of the individual hexamers depends upon (1) the degree of hexamer conservation (large differences are easier to detect than small differences) and (2) the number of times a SNP interrupts an ESE hexamer (larger data size increases our confidence in individual measurements and, therefore, increases our ability to measure differences). The VERIFY method could detect about 70% of strongly conserved hexamers (OR \> 0.5) in the set of 238 ESE hexamers using 8,400 SNPs. A similar type of analysis performed with 500,000 SNPs would not only detect almost all the strongly conserved motifs but would be able to detect motifs under a weaker degree of selection (OR \> 0.77) with a similar power (approximately 70%). As an alternative to using polymorphisms within a species, VERIFY could also utilize variations between closely related species to study selection on gene control elements. In addition, this work identifies two new categories of mutations that are under selective pressure: mutations that disrupt or alter RESCUE-ESE hexamers. This increased selective pressure presumably reflects the increased likelihood that such variations will alter splicing and, therefore, gene activity. Several cases of polymorphisms which result in allele-specific differences in splicing have been reported ([@pbio-0020268-Betticher1]; [@pbio-0020268-Stallings-Mann1]; [@pbio-0020268-Stanton1]). Here, we identify features such as proximity to splice sites and ESE disruption and alteration that should prove useful in discovering additional polymorphisms that affect splicing. As splicing mutations constitute at least 14% of disease-causing mutations ([@pbio-0020268-Stenson1]), polymorphisms that affect splicing would be good candidates for association studies intended to identify genetic contributors to quantitative traits or diseases. An additional benefit of using variations that are predicted to result in altered splicing relates to the feasibility of validating an RNA phenotype. While variations predicted to alter protein activity may require gene-specific activity assays or antibodies, RNA phenotypes such as exon skipping can be readily screened by RT-PCR, thus enabling large scale phenotyping of genotyped cell lines. Materials and Methods {#s3} ===================== {#s3a} ### SNP data {#s3a1} Build 112 of dbSNP was downloaded from the NCBI ftp server ([ftp.ncbi.nih.gov](ftp.ncbi.nih.gov)) and parsed from the XML format, which integrates data (e.g., annotation, sequence, transcript) from several sources ([@pbio-0020268-Sherry1]). A current description of the database structure is available online (<http://ncbi.nih.gov/>). To limit the contribution of erroneous SNPs, several filters were applied to produce the set of SNPs used in this analysis. First, SNPs located within a coding region in the assembly annotation were required to align to an internal exon in the GENOA exon dataset. GENOA is a genome annotation script that annotates exons by spliced alignment of mRNA/cDNA sequence to an assembled genome, applying a number of checks on the quality of the resulting alignments (see below). SNPs that mapped to multiple genomic regions, to known repetitive elements, to regions where the reading frame differed in the genomic/transcript alignment, or to an entry that had been contributed in the context of spliced transcripts (e.g., through comparison of ESTs) were excluded. We also removed entries where both versions of a SNP contained a match to either the human exon database or the chimpanzee *(Pan troglodytes)* genome (<ftp://ftp.ncbi.nih.gov/pub/TraceDB/pan_troglodytes/>). From the resulting pool of reference SNPs, entries associated with genotype data (population frequency) were used to build a validated SNP set. Unless otherwise indicated, SNPs were aligned to other sequences only when a perfect (33/33) base match was obtained using BLAST ([@pbio-0020268-Altschul1]) with one or the other of the polymorphic alleles. ### Exon data {#s3a2} Datasets of internal human exons were generated by spliced alignment of cDNA sequences from GenBank (release no. 134) to the assembled, masked human genome sequence (GoldenPath assembly HG13, <http://genome.ucsc.edu>) using the genome annotation script GENOA (D. H., Lee P. Lim, R.-F. Yeh, U. Ohler, and CBB, unpublished data). The exon/intron structures were inferred from at least two cDNA/genomic alignments, and the reading frame was determined from the CDS annotation of the GenBank cDNAs. As RESCUE-ESE hexamers were identified in constitutively spliced exons and alternative splicing sometimes results in ambiguous splice site positions, we confined our analysis to exons which showed no (cDNA) evidence of alternative splicing. Therefore, exons present in some, but not all, cDNA alignments overlapping a genomic region were excluded. In addition, it was required that each exon (1) be flanked by introns whose ends matched the consensus terminal dinucleotides of U2-type or U12-type introns (GT-AG, GC-AG, or AT-AC) and (2) be in an open reading frame that spanned at least three exons (the exon being considered and at least one flanking exon on either side). ### Simulation of exon mutations {#s3a3} A Monte Carlo simulation was used to estimate background frequencies, in humans, of ESE-disrupting, ESE-creating, ESE-neutral, and ESE-altering mutations in internal exons. A PERL script simulated point mutations in sequences during multiple passes through the GENOA exon dataset. For each exon considered, the script utilized a set of randomly generated numbers to dictate the following sequence of decisions: (1) whether a mutation would occur in an exon, (2) the position where a mutation would occur, and (3) the identity of the variant allele. Mutations were simulated in present-day internal coding exons with a probability of occurrence proportional to the length of the exon and no a priori strand or position bias. The output mutations were stored in reference SNP format and annotated with predicted ESEs as described in the text. The mutation probabilities and the pattern of substitution are functions of the input sequence (i.e., the base to be mutated and the 3′ neighboring base). The values of all 12 nucleotide-to-nucleotide substitution probabilities used here were considered for all four possible 3′ neighboring-base contexts. The distribution of dinucleotide substitution probabilities (48 possible) used in this work was derived from a large study of mutation in ribosomal protein pseudogenes (all substitution probabilities were extracted from [Figure 2](#pbio-0020268-g002){ref-type="fig"} except those for CpG, which were obtained from [Figure 1](#pbio-0020268-g001){ref-type="fig"} using the assumption that both transversion values, i.e., CpG to ApG and CpG to GpG, were of equal probability \[[@pbio-0020268-Zhang1]\]). These probabilities reflect the context effect of the nucleotide 3′ of the substituted base averaged over all four possible 5′ contexts. If neighboring-nucleotide effects are not independent and the relevant context encompasses more sequence than is being considered in the simulation, this may result in a slight effect on the substitution pattern generated by the simulation. The dominant well-known contextual influences, such as those seen in the CpG dinucleotide, are captured in the simulation. Total genomic sequence was used to derive the substitution rates and simulate mutations. Features that could potentially influence the mutation process, such as the extent of CpG methylation, regional GC content, or recombination rates, could not be explicitly incorporated into the simulation. However, the analysis and simulation excluded first exons (reducing the contribution of unmethylated CpG in CpG islands), and these other variables affected the result of this analysis only to the degree that they altered the substitution pattern, as distinct from the overall mutation rate. ### Analysis {#s3a4} Mutations were annotated in terms of their effect on overlapping RESCUE-ESE hexamers as ESE disruption (+ −), ESE creation (− +), ESE alteration (+ +), or ESE neutrality (− −). In the annotations, the first position refers to the wild-type or ancestral allele, where "+" indicates one or more ESE hexamers and "−" indicates no ESE hexamers. The ESE alteration category can include mutations in which a net gain or a net loss of ESE hexamers is annotated. All annotations used the set of 238 RESCUE-ESE hexamers described previously ([@pbio-0020268-Fairbrother1]). ORs (rather than RRs) were used to quantify the difference between the set of experimentally validated SNPs and simulated mutations in [Figures 2](#pbio-0020268-g002){ref-type="fig"} and [3](#pbio-0020268-g003){ref-type="fig"}. Briefly, the relative OR is defined here as where P(outcome\|selection) represents the probability that SNP (a mutation under selection) will have a particular ESE annotation outcome (+ +, + −, − −, or − +). The extent of selection was measured separately for synonymous and nonsynonymous mutations and standard methods were used to calculate 95% confidence intervals ([@pbio-0020268-Pagano1]). The MH test was used to determine whether the degree of selection was independent of synonymy. Experimentally validated SNPs that were located in the first or last 70 nucleotides of an exon were evaluated for ESE disruption as a function of position. As no restrictions were placed on exon size, it was possible for SNPs in small exons to be placed in multiple categories (as was the case for approximately 20% of all SNPs). ORs were used to measure the extent of selection, and significance was assessed at an α level of 0.05 (one-tailed) for ESE disruption mutations in the four different regions of an exon. Bootstrap sampling was used to determine hexamers that were significantly avoided by SNPs. RR was used to compare the selection on individual hexamers over the course of 5,000 trials. For each trial, 8,408 SNPs were sampled with replacement from the set of 8,408 SNPs described in the text. The frequency of cases where an ESE coincided with an SNP was calculated for each of the 238 RESCUE-ESE hexamers and divided by the expected frequency for that hexamer (determined through simulation). This ratio of frequencies was used to estimate RR for each hexamer. Dividing the instances where RR was greater than 1 by the number of trials (5,000) provided a bootstrap *p* value for each of the 238 hexamers. The interspecies comparisons considered the output of RESCUE-ESE on four vertebrate genomes: human, mouse, zebrafish, and pufferfish. The set of human hexamers was divided into subsets according to the degree of conservation across vertebrates in the following manner: Hexamers that were only present in humans defined set 1; hexamers present in human and mouse, set 2; hexamers present in human, mouse, and one of the fish species, set 3; hexamers present in human, mouse, zebrafish, and pufferfish, set 4. The degree of overlap between these hexamer sets and the RESCUE-ESE hexamers with an RR of significantly less than 1 was recorded and displayed in [Figure 4](#pbio-0020268-g004){ref-type="fig"}B. Power calculations for single hexamer analysis were performed as described previously ([@pbio-0020268-Pagano1]; method, pages 243--246; estimation of standard error as a function of sample size, page 355). For the purposes of the power calculation the mutations were assumed to be unbiased and the ESE hexamers were assumed to have a slightly elevated frequency (1.5/4,096 = 3.7 × 10^−4^ rather than 2.4 × 10^−4^), and so a particular ESE would be expected to be interrupted by a SNP at any one of six positions with a frequency of 6 × 3.7 × 10^−4^ = 2.2 × 10^−3^. This probability was used to estimate the number of variations that would interrupt (disrupt or alter) an ESE hexamer with and without selection for the different SNP database sizes and degrees of selection (ORs) described in the text. Supporting Information {#s4} ====================== URLs {#s4a} ---- <http://genes.mit.edu/burgelab/rescue-ese/>. An online tool to annotate RESCUE-ESE hexamers in exons. <http://genes.mit.edu/burgelab/Supplementary/fairbrother04/>. Contains exon, SNP, RR, and ancestral allele databases used and/or generated in this study. We thank the Pan troglodytes Genome-NIH Intramural Sequencing Center and the SNP consortium for the data provided on their web sites. We are grateful to Immaculata De Vivo, David Altshuler, and the reviewers for insightful comments on the manuscript. This work was supported by a Pharmaceutical Research and Manufacturers of America postdoctoral fellowship (to WGF) and by a grant from the National Institutes of Health (NIH) (to CBB), NIH grant R37-GM34277 (to PAS), National Science Foundation grant 0218506 (to PAS), National Cancer Institute grant P30-CA14051 to the Cancer Center Support (to PAS), and a Functional Genomics Innovation award from the Burrough Wellcome fund (to CBB and PAS). **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** WGF conceived and designed the experiments. WGF performed the experiments. WGF analyzed the data. WGF and DH contributed reagents/materials/analysis tools. CBB and PAS were the principal investigators. WGF wrote the paper. Academic Editor: Sean Eddy, Howard Hughes Medical Institute and Washington University Citation: Fairbrother WG, Holste D, Burge CB, Sharp PA (2004) Single nucleotide polymorphism--based validation of exonic splicing enhancers. PLoS Biol 2(9): e268. 3′ss : 3′ splice site 5′ss : 5′ splice site dbSNP : public SNP database ESE : exonic splicing enhancer EST : expressed sequence tag MH test : Mantel--Haenzel test mRNA : messenger RNA OR : odds ratio RESCUE : relative enhancer and silencer classification by unanimous enrichment RR : relative risk PERL : Practical Extraction and Report Language SNP : single nucleotide polymorphism VERIFY : variant elimination reinforces functionality

PubMed Central

2024-06-05T03:55:47.254655

2004-8-31

PMC514885

Introduction {#s1} ============ The development of higher eukaryotes depends on the establishment of complex spatiotemporal patterns of gene expression. Thus, an important key to understanding development is to decode the transcriptional control of patterned gene expression. The segmentation gene network of *Drosophila* has long been one of the prime paradigms for studying the role of transcription control in pattern formation ([@pbio-0020271-Carroll1]; [@pbio-0020271-Rivera-Pomar1]). The regulation within the network is almost entirely transcriptional, and many of the *cis*- and *trans*-acting components are well characterized. The network comprises maternal and zygotic factors that act in a hierarchical fashion to generate increasingly refined and complex expression patterns along the anterior-posterior (ap) axis in the blastoderm embryo ([@pbio-0020271-St2]; [@pbio-0020271-Driever1]; [@pbio-0020271-Pankratz1]; [@pbio-0020271-Sprenger1]; [@pbio-0020271-St1]; [@pbio-0020271-Furriols1]): The maternal factors form gradients stretching along the entire ap axis; the zygotic gap factors are expressed in one or more broad, slightly overlapping domains; the pair-rule genes are expressed in seven stripes and segment-polarity genes in fourteen stripes, prefiguring the segmental organization of the larva; finally, the homeotic genes specify segment identity (for schematic see [Figure 6](#pbio-0020271-g006){ref-type="fig"}A). ::: {#pbio-0020271-g006 .fig} Figure 6 ::: {.caption} ###### Module Predictions within the Segmentation Gene Network \(A) Schematic depiction of the regulatory relationships within the segmentation gene network. \(B) Ahab-predicted modules in the control regions of segmentation genes were classified based on their composition into pair-rule driven (pr, red), maternal/gap driven (mg, green), and mixed but predominantly pair-rule (pr(mg), light red) or predominantly maternal/gap driven (mg(pr), light green); see text for details. For each gene, the number and type of modules in the control region is shown; grouping of genes is indicated by brackets and follows the hierarchy as depicted in (A). The type of regulatory input a gene receives is indicative of its position within the gene network. :::  ::: Many of the segmentation genes are transcription factors themselves; their principal targets are segmentation genes acting at the same level or below. From a large body of genetic and molecular studies (for review see [@pbio-0020271-Akam1]; [@pbio-0020271-Cohen1]; [@pbio-0020271-McGinnis1]; [@pbio-0020271-St2]; [@pbio-0020271-Martinez1]; [@pbio-0020271-Pankratz1]), the following broad rules for regulation within the network have been gleaned (cf. schematic in [Figure 6](#pbio-0020271-g006){ref-type="fig"}A): Gap genes receive input from the maternal factors; the gap genes of the trunk heavily cross-regulate, while the gap genes of the head do not. The pair-rule genes are divided into a primary and a secondary tier: The primary pair-rule genes generate their seven-stripe pattern mainly through maternal and gap input, while the secondary pair-rule genes depend on (primary) pair-rule gene input; but the debate about which pair-rule genes belong to the primary tier is not resolved ([@pbio-0020271-Carroll1]; [@pbio-0020271-Klingler1]; [@pbio-0020271-Klingler2]). Segment-polarity genes receive pair-rule gene input, and the homeotic genes receive both gap and pair-rule input. Like other factors controlling the transcription of protein-encoding genes, the segmentation gene transcription factors bind to *cis*-regulatory elements, also called modules, and positively or negatively regulate the recruitment of the basal transcription machinery to the core promoter (for review see [@pbio-0020271-Gray1]; [@pbio-0020271-Arnone1]; [@pbio-0020271-Zhou1]; [@pbio-0020271-Blackwood1]; [@pbio-0020271-Roeder1]; [@pbio-0020271-Naar1]; [@pbio-0020271-Roth1]; [@pbio-0020271-Arnosti1]). Specifically, the maternal factors were found to act as activators, while the gap factors act mostly as repressors; however, there is a body of data suggesting that gap factors can act as activators or repressors in a context-dependent fashion (see below). The expression patterns of the segmentation genes are typically complex, and in many cases different aspects of the pattern are controlled by separate modules. An individual module typically receives input from multiple transcription factors and contains multiple binding sites for each of the factors; in many cases the relevant binding sites are clustered within a small interval of 0.5--1 kb. The combinatorial and redundant nature of the input and its clustering are features that are readily exploited for the computational detection of transcriptional control elements. We have recently developed an algorithm, Ahab, which uses a thermodynamic model to detect *cis*-regulatory modules ([@pbio-0020271-Rajewsky1]). Ahab uses binding site information for multiple transcription factors participating in a common process and seeks an optimal binding of the factors to a given sequence window. Binding site information for the factors is provided in the form of position weight matrices ([@pbio-0020271-Stormo1]), which Ahab uses to infer binding energies. Ahab then optimizes the total free energy of binding the factors to the sequence. The factors compete for binding with a local background model computed from the base composition within the sequence window; the competition between factors is treated as in standard thermodynamics. The result is then the best partitioning of the sequence window into binding sites and background. The total free energy under this partitioning is taken as the score, and can be used to rank modules. Thus, in contrast to other methods for module detection ([@pbio-0020271-Berman1]; [@pbio-0020271-Halfon1]; [@pbio-0020271-Markstein1]; [@pbio-0020271-Papatsenko1]; [@pbio-0020271-Rebeiz1]), Ahab requires no predefined factor-dependent cutoffs, which means that clusters of weak sites can be detected. We used Ahab for a genome-wide prediction of segmentation gene modules with maternal and gap input and found that it recovers known modules with excellent success ([@pbio-0020271-Rajewsky1]). Here, we use Ahab to identify novel modules within the segmentation gene network. We test 16 significant novel predictions and find that 13 faithfully produce pattern elements of the endogenous gene, while the remaining three produce more or less aberrant blastoderm patterns. Our combined computational and experimental analysis increases the number of characterized segmentation modules by 50% and provides effective de novo control region dissections for ten of the 29 genes with gap and pair-rule patterns. Furthermore, we systematically analyze Ahab\'s prediction of binding site composition for all experimentally validated modules. By correlating the expression patterns of modules with their binding site composition, we are able to show that the composition of modules is generally well fitted to the distribution of input factors, and we are able to determine the mode of action for six of the nine maternal/gap input factors. Finally, we explore Ahab\'s predictive ability when binding site information is less well defined, as is the case with the pair-rule factors. Despite the handicap, Ahab traces the global architecture of the segmentation gene network and pinpoints the unexpected behavior of *odd skipped* as a primary pair-rule gene. Results {#s2} ======= Prediction and Validation of Segmentation Modules {#s2a} ------------------------------------------------- As the principal arena for our investigation, we selected the top two tiers of the segmentation gene network, namely the gap and pair-rule genes (for references see [Dataset S1](#sd001){ref-type="supplementary-material"}). Using Ahab, we searched the genomic regions surrounding these genes for *cis-*regulatory modules containing clusters of binding sites for maternal and gap factors. As input for Ahab, we provided binding site information (in the form of position weight matrices derived from the literature; [Dataset S2](#sd002){ref-type="supplementary-material"}) for nine transcription factors: the maternal factors Bicoid (Bcd), Hunchback (Hb), Caudal (Cad), the Torso-response element (TorRE), and Stat92E (D-Stat), and the gap factors Kruppel (Kr), Knirps (Kni), Giant (Gt), and Tailless (Tll). Note that the weight matrices for Kni and Tll are quite unspecific, which leads to an increased number of binding site predictions. Conversely, the available binding site information for D-Stat and Gt is rather limited and thus appears artificially specific, resulting in fewer predictions. Ahab was run over the genomic regions of 29 genes with gap and pair-rule patterns consisting of 0.75 Mb of total genomic sequence (see [Materials and Methods](#s4){ref-type="sec"}). We experimented with two adjustable parameters of Ahab, free energy cutoff and the order of the background model, i.e., whether pairs or triples of bases are used as background sequence. We favored the lower order background, which is less stringent and increases the number of factor binding sites, and set the free energy cutoff at 15, which is approximately four standard deviations above the mean of genome-wide window scores ([Figure 1](#pbio-0020271-g001){ref-type="fig"}A). The window size was set at 500 bp, which we had previously found to deliver the most efficient recovery of known modules ([@pbio-0020271-Rajewsky1]). ::: {#pbio-0020271-g001 .fig} Figure 1 ::: {.caption} ###### Ahab Predictions and Recovery of Known Modules \(A) Histogram of genome-wide window scores for the Ahab mg run (maternal/gap input, window size 500 bp, window shift 50 bp, background model 2). As free energy cutoff we chose 15, which is approximately four standard deviations above the genome-wide mean (indicated by light blue line). \(B) Pie chart summarizing results of Ahab predictions for gap and pair-rule genes, including recovery of known modules and testing of novel predictions. (C--F) For the genomic regions of selected gap and pair-rule genes, the free energy profiles of two Ahab runs (mg and mgpr) are shown. The free energy cutoffs are marked by dotted lines; statistically significant predictions for the mg run are marked by black arrow heads (cf. [Figure 4](#pbio-0020271-g004){ref-type="fig"}). In the header above, the blastoderm expression pattern of the locus is depicted schematically, anterior to the left, posterior to the right. The position of experimentally validated modules within the control region is delineated by colored bars; the aspect of the endogenous pattern they drive is highlighted in matching color. Overall, the control regions of the gap genes *hb* and *Kr* and of the primary pair-rule genes *eve* and *h* are computationally well delineated with maternal/gap input. References: (1) [@pbio-0020271-Schroder1], (2) [@pbio-0020271-Margolis1], (3) [@pbio-0020271-Hoch1], (4) [@pbio-0020271-Goto1], (5) [@pbio-0020271-Fujioka2], (6) [@pbio-0020271-Riddihough1], (7) [@pbio-0020271-Howard1], and (8) [@pbio-0020271-Langeland1]. :::  ::: Under these conditions, Ahab predicts 52 modules within the genomic region of the 29 genes of interest, an average of about two modules per gene. This hit rate represents a 5-fold enrichment compared to the genome-wide rate. Of the 52 predicted modules, 43 are located in intergenic regions, nine in introns, and none in coding regions, indicating a bias of the predictions toward transcriptional control regions. Of the 31 known modules, we recover 22 as significant predictions (score \>15; because of overlaps, 20 Ahab predictions cover the 22 known modules), and three overlap with free energy peaks just below the cutoff ([Figure 1](#pbio-0020271-g001){ref-type="fig"}; cf. [Figure 4](#pbio-0020271-g004){ref-type="fig"}). In the six cases where Ahab misses known modules completely, the reasons are most likely missing input factors (e.g., *hkb\_ventral\_element* module; [@pbio-0020271-Hader1]) or a low number of binding sites (e.g., *ems\_head* module; [@pbio-0020271-Hartmann1]). The likelihood of recovering 22 modules at random is negligible (*p* \< 10^−8^). We also predict 32 novel modules, and we expect at least some predictions with scores below 15 to be functional as well. ::: {#pbio-0020271-g004 .fig} Figure 4 ::: {.caption} ###### Correlation of Expression Patterns with Module Composition Based on the expression pattern they give rise to, known and newly validated modules are sorted into anterior, posterior, and terminal (if expression bridges the 50% EL line, the module is labeled ant/post), and their binding site composition is evaluated using Ahab output from the mg run. The expression pattern of a module is depicted schematically (anterior = 100% EL, left; posterior = 0% EL, right), followed by name of gene, name of module, recovery as significant prediction (marked by X) or as subthreshold peak (marked by (X)) in D. melanogaster and *D. pseudoobscura,* distance to the gene\'s transcription start site (negative values denote upstream location), and binding site composition. For references see [Dataset S1](#sd001){ref-type="supplementary-material"}. Expression patterns of previously known modules are in black, those of newly validated modules are in dark pink, and modules with unfaithful/unstable patterns are in light pink. Binding site composition is given in the form of integrated profile values for individual input factors (see [Materials and Methods](#s4){ref-type="sec"}); higher color intensity emphasizes higher values. Diagnostic features are emphasized by black trim: In anterior modules Bcd sites are overrepresented and Cad sites are underrepresented, while in posterior modules Cad sites are overrepresented and Bcd sites underrepresented. Terminal modules are enriched in TorRE sites. :::  ::: For experimental validation, we selected 16 module predictions with scores greater than 15 and five with scores below 15 ([Figures 2](#pbio-0020271-g002){ref-type="fig"} and [3](#pbio-0020271-g003){ref-type="fig"}), located near genes with gap and pair-rule patterns whose control regions had not or only partially been dissected: *cad*, *cap 'n\' collar (cnc), Dichaete (D), fork head (fkh), gt, kni, knirps-like (knrl), nubbin (nub), ocelliless (oc), POU domain protein 2 (pdm2), odd skipped (odd),* and *sloppy paired 2 (slp2)*. We used the free energy profiles to delineate the module and then tested its ability to drive blastoderm expression using a *lacZ* reporter construct (see [Materials and Methods](#s4){ref-type="sec"}). All of the predicted modules we tested drive expression in the blastoderm. However, the faithfulness of the patterns produced by the modules varies. Of the 16 modules with scores greater than 15, 13 produce faithful patterns that reproduce one or more aspects of the endogenous pattern, two produce unfaithful patterns, and one has an unstable, insertion-dependent pattern. Of the five modules with scores below 15, two produce faithful and three produce unstable blastoderm patterns. This indicates that Ahab has excellent success in predicting modules driving blastoderm expression and that the free energy cutoff is well chosen, with few false positives and negatives. The fact that unfaithful or unstable patterns are produced by some of the modules is likely a reflection of the fact that Ahab makes predictions simply on the basis of the total free energy without any explicit rules as to the number and type of factors that have to contribute to the binding. By comparing the composition of modules of different degrees of faithfulness or stability, one can attempt to formulate such rules (see below). ::: {#pbio-0020271-g002 .fig} Figure 2 ::: {.caption} ###### Expression Patterns Driven by Ahab-Predicted Modules I Ahab-predicted modules in the control region of gap and pair-rule genes were tested by fusing putative modules to a basal promoter driving *lacZ* (module-basal promoter-lacZ; [@pbio-0020271-Thummel1]). The genomic regions, with free energy profiles, for two Ahab runs (mg and mgpr) are shown on the right. The free energy cutoffs are marked by dotted lines; mg run predictions with scores greater than 15 are marked by black arrowheads, tested subthreshold peaks with scores below 15 by open arrowheads. The transcribed region of the locus is marked in blue, the experimentally tested genomic regions are marked by pink bars and named according to distance from transcription start site to middle of the enhancer, and previously known modules are marked by orange bars. The endogenous gene expression is shown on the left (blue frame), the expression pattern driven by the module(s) in the center (pink frame). Embryos are oriented anterior to left, dorsal up. In a few cases, the patterns driven by Ahab-predicted modules are unfaithful to the endogenous gene expression; we distinguish "unfaithful" and insertion-dependent "unstable" patterns. For further description see text. (A) *gt,* (B) *cnc,* (C) *oc,* (D) *D,* (E) *cad,* (F) *fkh,* and (G) *slp2.* References: (1) [@pbio-0020271-Berman1], (2) [@pbio-0020271-Gao1], (3) [@pbio-0020271-Lee1], and (4) [@pbio-0020271-Pankratz2] and [@pbio-0020271-Rivera-Pomar2]. :::  ::: ::: {#pbio-0020271-g003 .fig} Figure 3 ::: {.caption} ###### Expression Patterns Driven by Ahab-Predicted Modules II See legend for [Figure 2](#pbio-0020271-g002){ref-type="fig"}. (A) *kni,* (B) *knrl,* (C) *pdm2,* (D) *nub,* and (E) *odd.* :::  ::: Using Ahab for the Dissection of Segmentation Gene Control Regions {#s2b} ------------------------------------------------------------------ The gap gene *gt* is initially expressed in two domains in the blastoderm, one anterior and one posterior; as cellularization progresses, the anterior domain splits into two stripes, and, finally, a third expression domain develops at the anterior terminus. We predict three modules*, gt\_(−1), (−3),* and *(−6),* all of which we tested; in addition, we tested one subthreshold peak further upstream, *gt\_(−10)* (see [Figure 2](#pbio-0020271-g002){ref-type="fig"}A). We can account for all *gt* pattern elements: the subthreshold *gt\_(−10)* faithfully produces the anterior expression, *gt\_(−6)* produces the anterior tip expression, and the *gt\_(−3)* module produces the posterior expression (cf. [@pbio-0020271-Berman1]). Interestingly, *gt\_(−1)* is bifunctional and produces both the anterior and the posterior expression domain. The gap gene *kni* is expressed in two domains in the blastoderm, one at the anterior tip and one in the posterior, but only the module driving the posterior expression had previously been identified ([@pbio-0020271-Pankratz2]). In addition to the known module *kni\_kd,* we predict two additional modules, one further upstream, *kni\_(−5),* and one in the first intron, *kni\_(+1)*. The *kni\_(−5)* module faithfully produces the expression at the anterior tip, while the *kni\_(+1)* module drives an imprecise *kni* pattern with an aberrant anterior and an abnormally widened posterior expression domain (see [Figure 3](#pbio-0020271-g003){ref-type="fig"}A). The sister gene *knrl* is expressed in the same pattern as *kni*. We find two significant predictions in the control region; we tested one, *knrl\_(+8),* which produces an unfaithful pair-rule-like pattern (see [Figure 3](#pbio-0020271-g003){ref-type="fig"}B). The less well known gap genes *nub* and *pdm2* are both expressed in a broad posterior domain; *pdm2*, but not *nub,* develops a segmental pattern during gastrulation. The control regions of the two genes have not been dissected ([@pbio-0020271-Kambadur1]). We find one significant prediction for *nub, nub\_(−2),* and two for *pdm2, pdm2\_(+1)* and*(+3). nub\_(−2*) faithfully reproduces the posterior expression of the gene (see [Figure 3](#pbio-0020271-g003){ref-type="fig"}D). For *pdm2, pdm2\_(+1)* faithfully reproduces the posterior domain as well as the segmental expression of the gene, while *pdm2\_(+3)* produces line-dependent variable patterns of blastoderm expression (see [Figure 3](#pbio-0020271-g003){ref-type="fig"}C). The *cad* gene is expressed both maternally and zygotically. Its zygotic expression in the blastoderm consists of a single posterior stripe. We make a single significant prediction, *cad\_(+14),* which faithfully reproduces the pattern (see [Figure 2](#pbio-0020271-g002){ref-type="fig"}E). *fkh* is initially expressed in a single domain at the posterior end, to which a second domain at the anterior end is added later in the blastoderm. We make a single significant prediction, *fkh\_(−2),* which faithfully produces the early domain at the posterior end (see [Figure 2](#pbio-0020271-g002){ref-type="fig"}F). The head gap gene *cnc* is expressed in two domains, an anterior cap and a collar. Our single significant prediction, *cnc\_(+5),* faithfully produces the pattern (see [Figure 2](#pbio-0020271-g002){ref-type="fig"}B). Similarly, the single significant prediction for *oc, oc\_(+7),* faithfully produces the single head gap domain of the endogenous gene (see [Figure 2](#pbio-0020271-g002){ref-type="fig"}C). *D* is initially expressed in a broad domain encompassing the entire segmented portion of the blastoderm embryo, and an anterior patch is added at the end of the blastoderm. The control region of *D* has not been dissected ([@pbio-0020271-Sanchez-Soriano1]). Our single significant prediction, *D\_(+4),* faithfully produces the early blastoderm pattern (see [Figure 2](#pbio-0020271-g002){ref-type="fig"}D). Finally, the pair-rule genes: *slp1* and *slp2* are first expressed in a gap-like pattern in the head, followed by expression in seven and then fourteen stripes. The dissection of the upstream region of *slp1* had identified the stripe element but not the gap-like expression in the head ([@pbio-0020271-Lee1]). We find a subthreshold peak upstream of *slp2* that nicely reproduces the missing head gap pattern (see [Figure 2](#pbio-0020271-g002){ref-type="fig"}G). *odd* is first expressed in a pair-rule and then in a segmental pattern and has traditionally been placed among the secondary pair-rule genes, which are thought to generate their pattern through pair-rule input rather than direct maternal/gap input. Surprisingly, we find two significant predictions in the upstream region of the gene, *odd\_(−3)* and*(−5).* Both these modules drive expression in two stripes: *odd\_(−3)* drives expression in stripes 3 and 6, while *odd\_(−5)* drives expression in stripe 1 and a broader region encompassing stripes 5 and 6 of the endogenous pattern (see [Figure 3](#pbio-0020271-g003){ref-type="fig"}E). This behavior is reminiscent of the two-stripe modules of *eve (eve\_stripe3\_7* and *eve\_ stripe4\_6).* Thus, at least four of the seven *odd* stripes are formed as individual stripes by maternal/gap input rather than as a complete seven-stripe pattern, indicating that *odd* has primary pair-rule character. Overall, our experimental validation demonstrates that Ahab is highly successful in predicting modules that drive patterned expression in the blastoderm. The algorithm finds missing modules that complement existing ones to collectively produce the expression pattern of a gene and identifies, with surprising accuracy, relevant modules in previously undissected control regions. Most of the modules faithfully produce pattern elements of the endogenous gene, suggesting that our delineation of modules, which is based on the free energy profile of the prediction, is generally quite accurate. Module Composition and Pattern of Expression {#s2c} -------------------------------------------- Ahab\'s success in finding modules encouraged us to examine in greater detail its prediction of the binding site content of modules. We sought to examine whether the expression patterns of the previously known and our newly tested modules correlate with their composition. In its optimization procedure, Ahab fits all input factors simultaneously to the genomic region of interest, while experimental sites for transcription factors are typically determined in the absence of any competition. Ahab reports binding site composition in the form of integrated profile values, which tally the fractional occupancy of sites for a given factor, and are thus a measure of the strength of binding by this factor (see [Materials and Methods](#s4){ref-type="sec"}). In order to gauge the accuracy of Ahab predictions of module composition, we examined how well Ahab performs in recovering known binding sites (for detailed description see [Materials and Methods](#s4){ref-type="sec"}). Overall, the recovery of known sites ranges from 50% to 100%, with the most specific factors/position weight matrices showing the best recovery. The missed sites are typically weak and are not misattributed to other factors but rather to background. Thus, Ahab should provide a reliable profile of module composition. In order to correlate the binding site composition with the ap expression pattern of the modules, we charted the previously known modules and all the newly validated modules with faithful expression and sorted them according to their expression along the ap axis (see [Figure 4](#pbio-0020271-g004){ref-type="fig"}). We ask which, if any, features are diagnostic. The Maternal Factors {#s2d} -------------------- In anterior modules (driving expression at 50%--100% egg length ), Bcd sites are overrepresented and Cad sites underrepresented (see [Figure 4](#pbio-0020271-g004){ref-type="fig"}), including seven known and six newly tested modules. In posterior modules (driving expression at 0%--50% EL), Bcd sites are underrepresented and Cad sites overrepresented, including five known and five newly tested modules. Finally, in terminal modules (driving expression at 0%--20% and 80%--100% EL), TorRE sites are strongly overrepresented, including four known and one newly tested modules. In addition to the TorRE-terminal signature, terminal modules expressed at the anterior terminus often contain Bcd sites, and those expressed at the posterior end, Cad sites. Thus, there is a strong positive correlation between the expression pattern of the module and the maternal input they receive, supporting the general interpretation that the maternal factors act as transcriptional activators in their realm of expression. To take a closer look at this relationship, we computed for each input factor and for every position along the ap axis the average number of binding sites found in the modules driving expression at that position. We plotted this number as a function of ap position and compared the resulting curve with the input factor distribution as determined by Reinitz and coworkers ([@pbio-0020271-Myasnikova1]) ([Figure 5](#pbio-0020271-g005){ref-type="fig"}A). For TorRE, the distribution of binding sites beautifully follows the expression profile of the input factor (as inferred from expression of its negative regulator, Capicua), indicating that binding sites are present almost exclusively where the cognate factor is active. The distributions of Bcd and Cad binding sites broadly conform with the anterior and posterior gradients of their respective input factors. The rise in the curves at the posterior terminus for Bcd and at the anterior terminus for Cad is caused by terminal modules expressed at both ends of the embryo. Overall, for the maternal activators, the binding site composition of modules is well fitted to the input factor distribution. ::: {#pbio-0020271-g005 .fig} Figure 5 ::: {.caption} ###### Ap Distribution of Binding Sites and Cognate Input Factors \(A) Plots depict distribution of input factors (black) along the ap axis (anterior tip = 100, posterior tip = 0) (based on [@pbio-0020271-Myasnikova1]) and the average number of binding sites (as measured by integrated profile values; [Figure 4](#pbio-0020271-g004){ref-type="fig"}) found in all modules driving expression at a given percent EL (red) (see [Materials and Methods](#s4){ref-type="sec"}). For TorRE, Bcd, and Cad, the distributions of binding sites and input factors are positively correlated. For Hb, Gt, and Kr, the distributions are negatively correlated; note that the number of binding sites is particularly high in modules expressed adjacent to the expression domain of these factors. In the case of Hb, modules with more Hb sites than Bcd sites (blue) show negative correlation with input factor distribution, and modules with fewer Hb sites than Bcd sites (green) show positive correlation, indicating bimodal function of Hb. For Kni and Tll, no clear correlations are found, possibly because of the unspecificity of their weight matrices. \(B) Information scores of the Kr, Kni, and Tll weight matrices. :::  ::: The Gap Factors {#s2e} --------------- The situation regarding the gap factors is more complex. When examining the distributions of Hb, Gt, and Kr binding sites and comparing them with the input factor distributions, we clearly find an anticorrelative relationship: The number of sites is lower in regions where the cognate factor is present, and higher in regions where the factor is absent ([Figure 5](#pbio-0020271-g005){ref-type="fig"}A). Remarkably, the number of sites is particularly high in regions immediately adjacent to the expression domain of the factor. These findings are consistent with the experimental evidence that gap factors act as repressors. Thus, modules which have many sites efficiently suppress expression within the domain of the input factor, and permit expression only outside the domain. The great majority of modules conform to this anticorrelative relationship; we can therefore conclude that, overall, repression is the prevalent mode of action for these gap factors. However, we do find some modules that appear to be coextensively expressed with the presumptive repressors. One possible explanation is that the input factor has a different mode of action in these modules, that is, instead of repression it may mediate activation. Hb appears to be an example for such a switch in the mode of action. We find many modules with a small number of Hb sites that are coextensively expressed with Hb in the anterior, and it has been shown experimentally that Hb function is context dependent: Repressor function has been demonstrated for several posterior modules (e.g., *kni\_kd, eve\_stripe3\_7,* and *eve\_stripe4\_6*) ([@pbio-0020271-Pankratz2]; [@pbio-0020271-Fujioka2]), while activator function has been demonstrated for several anterior modules (e.g., *hb\_anterior, Kr\_CD1,* and *eve\_stripe2*) ([@pbio-0020271-Treisman1]; [@pbio-0020271-Hoch1]; [@pbio-0020271-Small1]; [@pbio-0020271-Stanojevic1]). It is thought that Hb is converted from a repressor to an activator by the concurrent presence of homeobox factors such as Bcd ([@pbio-0020271-Zuo1]; [@pbio-0020271-Simpson-Brose1]). We examined the composition of these two sets of known modules and found that in the posterior modules, in which Hb acts as a repressor, the profile values of Hb exceed those of Bcd, while in anterior modules, in which Hb acts as an activator, the profile values of Hb are lower than those of Bcd. When we apply the simple rule suggested by this observation to all modules containing Hb sites, we find that it significantly improves the picture: the Hb\>Bcd (Hb as repressor) set is strongly negatively correlated with Hb factor expression, while the Hb\<Bcd (Hb as activator) set is positively correlated with Hb factor expression (the only exception is the *D\_(+4)* module, which drives expression in a broad domain straddling the 50% EL line). Thus, the global distribution profile of Hb sites can largely be explained by introducing a simple contextual rule. By contrast, for Gt and Kr, the number of modules expressed coextensively with the input factor is comparatively small. In the case of Gt, all experimental evidence points to its acting as a repressor. Increasing the spatiotemporal resolution of the plot to reflect the modulation of Gt expression over time may be sufficient to account for the presence of Gt sites in at least some of the potentially "noncompliant" modules (*cnc\_(+5), oc\_(+7), oc\_otd\_early,* and *hb\_ant*). In the case of Kr, context-dependent function has been suggested, but mostly based on tissue culture experiments ([@pbio-0020271-Sauer1]; [@pbio-0020271-La1]; [@pbio-0020271-La2]). The four potentially noncompliant modules (*Kr\_CD2, run\_stripe3, nub\_(+5), D\_(+4))* are clearly expressed coextensively or overlapping with the Kr input factor. Since the average number of binding sites is low in these modules, it is possible that Kr acts as a repressor but that this manifests itself only in a reduced expression level. In fact, the *Kr\_CD2* module has been noted to be more weakly expressed than its sister module *Kr\_CD1,* which lacks Kr sites ([@pbio-0020271-Hoch1]), but there are too many other differences in their binding site composition to draw any firm conclusions. These noncompliant modules provide a solid experimental platform for resolving the issue of whether or not Kr truly switches its mode of action in vivo. Finally, for Kni and Tll, most experimental evidence points to repression, but context-dependent activation has been suggested in a few cases ([@pbio-0020271-Langeland2]; [@pbio-0020271-Margolis1]; [@pbio-0020271-Kuhnlein1]; [@pbio-0020271-Hartmann1]). As noted at the beginning, the weight matrices for both factors are fairly unspecific ([Figure 5](#pbio-0020271-g005){ref-type="fig"}B), resulting in a lower level of confidence in the predictions, which typically show a large number of binding sites. When plotting binding site and input factor distributions, no clear positive or negative correlations are visible ([Figure 5](#pbio-0020271-g005){ref-type="fig"}A), suggesting either strong context-dependent function---which is not really supported by the extant literature---true indiscriminate binding, or simply poor binding site information. Unfaithful Modules {#s2f} ------------------ In our experimental tests, we found a few novel modules that drive unfaithful patterns. Can we understand their behavior based on the composition profile of the module? We observed two flavors of unfaithful expression: strong invariant and weak variable. The *kni\_(+1)* module is an example of the former: It drives expression in a posterior domain that is wider than the endogenous pattern (see [Figure 3](#pbio-0020271-g003){ref-type="fig"}A). When compared to the faithful *kni\_kd* module, *kni\_(+1)* contains the same types of binding sites, but with different profile values: The profile values for the activator Cad are higher and the ones for the repressors Hb, Kr, and Tll are lower (see [Figure 4](#pbio-0020271-g004){ref-type="fig"}). This suggests that an increase in activator binding together with a decrease in binding by adjacently expressed repressors may be responsible for the widening of the posterior domain. The *pdm2\_(+3)* module is an example of weak and unstable expression (see [Figure 3](#pbio-0020271-g003){ref-type="fig"}C), which we find more often when analyzing subthreshold peaks. Such modules typically suffer from a reduced number of activator sites and an increase in sites for coextensively expressed repressors (see [Figure 4](#pbio-0020271-g004){ref-type="fig"}). Thus, the two flavors of unfaithful patterns, strong invariant and weak variable, appear to correlate with the ratio of activator to repressor sites in the module and the degree to which the distributions of the relevant input factors are compatible. Further experimental and computational work will be required to determine precise module composition rules, but both faithful and unfaithful modules can contribute to defining them. Evolutionary Conservation {#s2g} ------------------------- The availability of the Drosophila pseudoobscura genome makes it possible to ask how well segmentation modules are conserved. In a previous study, [@pbio-0020271-Emberly1] showed that the degree of sequence conservation between D. melanogaster and D. pseudoobscura is not significantly higher in known segmentation modules than in surrounding noncoding regions, suggesting that sequence conservation per se is not sufficient to identify such modules. We obtain the same result for our Ahab predictions (data not shown). However, when we run Ahab over the aligned segmentation gene control regions in *D. pseudoobscura,* using D. melanogaster weight matrices as input, we recover as significant predictions about the same number of known modules as in *D. melanogaster,* indicating that there is substantial *functional* conservation (see [Materials and Methods](#s4){ref-type="sec"}). However, only 24 of the 35 known and newly validated modules that are recovered in D. melanogaster also score as significant predictions in *D. pseudoobscura,* with an additional seven as subthreshold peaks (see [Figure 4](#pbio-0020271-g004){ref-type="fig"}). Conversely, four subthreshold D. melanogaster modules are recovered as significant predictions in *D. pseudoobscura,* and three known modules are recovered only in D. pseudoobscura. Thus, modules with maternal/gap input appear to be in some evolutionary flux, which needs to be taken into consideration if evolutionary conservation is employed as a tool in module discovery. Regulatory Input within the Segmentation Gene Hierarchy {#s2h} ------------------------------------------------------- Given Ahab\'s success in predicting modules with maternal and gap input, we decided to expand the analysis to the entire segmentation gene network and explore the algorithm\'s performance when less well defined binding site information is available. To this end, we included the control regions of a total of 48 genes: To the genes with gap-like and pair-rule patterns, we added segment-polarity and homeotic genes (for references see [Dataset S1](#sd001){ref-type="supplementary-material"}). Concurrently, we expanded the set of binding site inputs. The maternal and gap factors were used as before (mg run). In addition, we collected binding site information from the literature for the pair-rule factors Hairy (H), Even skipped (Eve), Runt (Run), Fushi tarazu (Ftz), Ftz transcription factor 1 (Ftz-f1), Paired (Prd), and Tramtrack (Ttk) ([Dataset S2](#sd002){ref-type="supplementary-material"}). For all these factors the available binding site information is generally less extensive and relies less on in vivo and more on in vitro experiments such as Selex. This again has the consequence that the weight matrices are artificially more specific, resulting in the prediction of fewer sites but higher scores for a match. The pair-rule factors were run by themselves (pr run) and in combination with the maternal and gap factors (mgpr run), with window size 500 and background model 2. In the combined control regions of the entire set of 48 segmentation genes, which total 1.7 Mb in length, we find 82 significant peaks for the mg run (score \>15), 56 for the pr run (score \>15), and 69 for the mgpr run (score \>22, cutoff set to equal genome-wide mean plus four standard deviations), in total 145 distinct putative modules, an average of approximately three per gene. Interestingly, the mg run and pr run peaks are completely nonoverlapping. We determined the relative contribution of maternal/gap and pair-rule input to each predicted module by evaluating its binding site composition as revealed by the mgpr run, i.e., using all input factors. Modules were classified into four types: maternal/gap driven, pair-rule driven, or driven by both but with bias towards maternal/gap input or pair rule input (see [Materials and Methods](#s4){ref-type="sec"}). The number and types of modules found within the control region of each target gene are shown in [Figure 6](#pbio-0020271-g006){ref-type="fig"}B. For the genes with gap-like expression, maternal and gap input strongly predominates; for pair-rule and segment-polarity genes, pair-rule input predominates. The homeotic genes receive both types of input. This global result reflects very well the overall regulatory structure of the segmentation gene network. However, we find interesting exceptions to the global rules. Among the pair-rule genes, *odd* stands out as receiving unexpectedly strong maternal/gap input. *odd* is expressed in a pair-rule and then segment-polarity pattern ([@pbio-0020271-Coulter1]) and has traditionally been placed among the secondary pair-rule genes ([@pbio-0020271-Klingler1]; [@pbio-0020271-Pankratz1]; [@pbio-0020271-Pick1]). But as our dissection reveals (see [Figure 3](#pbio-0020271-g003){ref-type="fig"}E), *odd* receives strong maternal/gap input and generates at least four of its seven stripes via two-stripe modules, suggesting that it in fact belongs to the primary pair-rule tier. In addition, as noted above, the control region of the secondary pair-rule gene *slp2* contains a subthreshold peak with maternal/gap input that drives its early gap-like expression in the head region (see [Figure 2](#pbio-0020271-g002){ref-type="fig"}G). Finally, we also examined the position of known and predicted modules relative to the transcription start site of the gene ([Figure 7](#pbio-0020271-g007){ref-type="fig"}) We found that maternal/gap-driven (mg run) modules are strongly biased toward the proximal upstream region (−6 to 0 kb), the first 2 kb of intronic space, and the proximal downstream region (+2 to +4 kb). This clustering is found for the gap, pair-rule, and segment-polarity genes, whose genomic organization is typically simple, but not for the homeotic genes, which typically have much larger control regions and multiple large introns, with wide scattering of predicted modules. For pair-rule-driven (pr run) modules, a similar though less pronounced clustering is observed (data not shown). ::: {#pbio-0020271-g007 .fig} Figure 7 ::: {.caption} ###### Genomic Position of Modules Position of modules predicted by the Ahab mg run relative to the transcription start site of the cognate loci; predictions for the homeotic genes are excluded. The number of modules found at a given position is shown in blue. The black line indicates the probability of a module occurring at a given position (calculated by dividing the number of modules at a given position by the number of control regions extending to that position). The stippled black line shows that probability if modules were randomly distributed. Modules with maternal/gap input are clustered within the first 6 kb upstream, in the first 2 kb of intronic space, and around 2 kb downstream (measured from the end of the gene). :::  ::: Discussion {#s3} ========== In this study we have demonstrated that the Ahab algorithm can be used successfully for two purposes: the prediction of novel segmentation modules within genomic sequence and the prediction of module binding site composition. The computational analysis of control regions with Ahab dramatically improves the efficiency of the experimental dissection, allowing us to significantly increase the number of validated *cis-*regulatory elements from 31 to 46 and to provide effective de novo dissections for ten segmentation genes. Two principal factors contribute to this success. First, the existing experimental data for the segmentation gene network provide a rich substrate for the computational effort. Second, the biochemistry underlying the regulation of transcription, that is, the binding of transcription factors to DNA, is well described by equilibrium thermodynamics ([@pbio-0020271-von1]; [@pbio-0020271-Berg1], [@pbio-0020271-Berg2], [@pbio-0020271-Berg3]; [@pbio-0020271-Ptashne1]), and thus Ahab\'s use of equilibrium conditions to predict the number, type, and occupancy of binding sites within a window of genomic sequence mimics the intrinsic process. The global analysis of the segmentation gene hierarchy shows that the prevalence of maternal/gap input strongly correlates with gap-like expression, while the prevalence of pair-rule input strongly correlates with segmental expression. Integrating the inputs over all modules within the control region of a gene provides a reliable indication of its type of expression pattern and position within the hierarchy. In fact, the integrated predictions are so accurate as to pinpoint abnormalities in the gene classification, such as the known head gap function of *slp2,* and also the hitherto unknown primary pair-rule character of *odd*. Since our knowledge of input factor sites is incomplete (particularly regarding the pair-rule factors), these positive results are likely to reflect the redundant and combinatorial nature of the input. Ahab performs well not only in identifying modules, but also in predicting their composition, thus permitting an analysis of binding site composition under uniform criteria for the entire set of known and newly validated maternal/gap-driven modules. Gene expression studies in mutant embryos have revealed the global regulatory interactions within the segmentation gene network ([@pbio-0020271-St2]; [@pbio-0020271-Pankratz1]; [@pbio-0020271-Rivera-Pomar1]; [@pbio-0020271-Furriols1]), but are not suited to uncover redundancies within the network or to separate direct from indirect effects. This becomes possible by examining the inputs into the *cis-*regulatory modules. We find that the vast majority of the modules expressed in the early blastoderm contain maternal factor sites, which strongly suggests that the maternal gradient systems of Cad, Hb, Bcd, and Torso (through its transcriptional effectors) have most, if not all, of the early zygotic patterning along the ap axis under their direct control. Together with the strong interdependence of the maternal gradient systems, this massively parallel output would explain the coordinated and long-range effects on segmentation gene expression patterns that are observed when maternal factors are titrated up or down through genetic manipulation. Further, by correlating the binding site content of modules driving expression at a given position with input factor distributions, we are able to infer the mode of action for six of the nine factors and to show that modules are generally well fitted to the distributions of their positive and negative input factors. The maternal factors act as activators within their domain of expression, while the gap factors act largely as repressors. This overall result confirms previously existing data and demonstrates that the rules gleaned earlier from rather small datasets generalize very well over the entire set. Interestingly, our data also provide support for the idea that Hb functions in a bimodal fashion and suggest a simple rule for its context-dependent switch from repression to activation. Modules with few Hb and many Bcd sites drive expression in the anterior half of the embryo, while modules with more Hb than Bcd sites do not. Depending on module composition and Bcd availability, Hb can thus activate transcription; this Bcd/Hb synergy could serve to bolster transcriptional activation in regions where Bcd levels taper off. For Kni and Tll, the mode of action cannot be assessed on the basis of the extant binding site information. The comparison of modules with faithful and unfaithful or unstable patterns provides some interesting additional clues for composition rules, such as the ratio and compatibility of activator and repressor sites. However, to address the question of how the precise domain boundaries are established within a given region of the embryo, a more detailed examination of composition rules and of the internal organization of modules will be needed, specifically of rules governing the number, affinity, spacing, and arrangement of binding sites. This analysis will require different types of experimentation as well as additional computational analysis. The performance of Ahab is influenced by a number of parameters, but the most important is the quality of the input factor weight matrices. To further improve weight matrices, more sites for undersampled factors will have to be collected (D-Stat and Gt), and existing sites for the unspecific factors will have to be scrutinized (Kni and Tll). More importantly, the relative affinity of binding sites for their factor will have to be measured in a more comprehensive fashion. The ideal experiment would measure, under identical conditions, the relative binding affinity of the consensus sequence to all possible single base mutations in the consensus binding site ([@pbio-0020271-Benos1]). For some segmentation modules, the currently predicted binding site composition is clearly insufficient to explain their expression, indicating that some of the relevant input has not been characterized. We have experimented with motif-finding algorithms and found that novel, biologically functional binding motifs can be identified by searching for locally overrepresented motifs within known modules and filtering out the known input factor binding sites (see [@pbio-0020271-Rajewsky1]; J. F., M. P., M. D. S., and U. G., unpublished data), which suggests that computational methods can also assist the identification of novel input factors. With an Ahab run that recovers 70% of the known modules with predominant maternal/gap input, we predict another 32 putative modules in the control regions of gap and pair-rule genes. Most of these look plausible in terms of genomic location and composition, and as our validation shows, many drive blastoderm expression that faithfully reproduces the endogenous pattern of the gene. However, we also found modules whose expression does not match the endogenous pattern (unfaithful/unstable) or whose composition does not suggest any coherent expression pattern (e.g., no activator sites); among the latter are some predictions dominated by Kni and Tll sites, which are potentially problematic because of the unspecificity of their weight matrices. The apparently anomalous modules could drive expression at later stages of development or could simply be artifacts of improper delineation or missing relevant input. A more intriguing possibility is that some of these modules are in evolutionary transit---nascent or dying. Such modules might be held in check by relatively few point mutations ("pseudo" modules), by nearby insulator elements, or by restricted access to the basal promoter when competing with the functional modules. The effort to discover the true nature and function of these anomalous modules will be aided by the computational and experimental comparison of corresponding modules in D. melanogaster and D. pseudoobscura. Materials and Methods {#s4} ===================== {#s4a} ### Position weight matrices and Ahab runs {#s4a1} When possible, previously compiled position weight matrices were used: for Bcd, Hb, Cad, TorRE, Kr, Kni, and Tll ([@pbio-0020271-Rajewsky1]), and for Ftz, Prd\_HD, and Ttk ([@pbio-0020271-Papatsenko1]). For H, Run/CBF, and D-Stat, we directly used in vitro selection data ([@pbio-0020271-Melnikova1]; [@pbio-0020271-Van1]; [@pbio-0020271-Yan1]). For Eve\_HD, the alignment was taken from the literature ([@pbio-0020271-Hoey1]), for Gt, Eve\_t2, and Ftz-f1, footprinted sites from the literature were aligned ([@pbio-0020271-Hoey1]; [@pbio-0020271-Biggin1]; [@pbio-0020271-Ueda1]; [@pbio-0020271-Jiang1]; [@pbio-0020271-Capovilla1]; [@pbio-0020271-Fujioka1]; [@pbio-0020271-Florence1]; [@pbio-0020271-Yu1]; [@pbio-0020271-Shimell1]). The binding sites, alignments, and weight matrices used plus references are listed in [Dataset S2](#sd002){ref-type="supplementary-material"}. For description and mathematical details of the algorithm, see [@pbio-0020271-Rajewsky1]. All runs were carried out on *Drosophila* genome sequence Release 2 after masking tandem repeats in the genomic sequence as described in [@pbio-0020271-Rajewsky1]. Control regions were defined as the sequence surrounding a gene and limited by the two flanking genes, up to a maximum of 20 kb upstream and 10 kb downstream, and with a buffer for the flanking genes of 2 kb upstream and 1 kb downstream. For the homeotic genes, no maximum for the upstream or downstream extension of the control region was imposed. ### Mapping of known modules {#s4a2} The genomic position of known modules was derived from literature ([@pbio-0020271-Papatsenko1]; [@pbio-0020271-Rajewsky1]) or mapped to genomic sequence from the literature using restriction sites, PCR primers, or distances relative to transcription start site. For a complete list and description see [Dataset S3](#sd003){ref-type="supplementary-material"}. ### Significance of Ahab predictions {#s4a3} To assess the significance of Ahab module predictions, we calculated the overlap between predictions and known modules in basepairs, and compared it with the overlap achieved when predictions are randomly placed within the delineated control regions (minus masked and coding sequence). We failed to match the actual overlap through 10^8^ randomizations, resulting in an estimate of *p* \< 10^−8^ for the significance of the recovery of known modules by Ahab. When we remove from the calculation the 13 modules that were used for the construction of weight matrices *(Kr\_CD1, Kr\_CD2\_AD1, eve\_stripe3\_7, eve\_stripe2, h\_stripe5, h\_stripe6, hb\_anterior\_actv, hb\_central\_&\_posterior\_actv, kni\_kd, oc\_head, tll\_K2, tll\_P2,* and *tll\_P3*), along with the *kni, hb,* and *tll* control regions, which contain no additional annotated modules, we find *p* = 4.9 × 10^−6^. ### Ahab recovery of known binding sites {#s4a4} The experimental binding sites that define our weight matrices are derived from a variety of in vitro experiments that typically neglect competition between transcription factors, whereas Ahab, in its prediction of binding sites, fits all factors simultaneously. To gauge whether Ahab can be used as a predictor of module composition, we examined what fraction of known binding sites the algorithm recovers. The only free parameter in the comparison is the profile value (between 0 and 1), which measures the fractional occupancy of a site by its factor; a profile value of 1 means that a site is always occupied by its factor. A site was scored as found if the prediction exceeded a certain profile value cutoff and overlapped the experimental footprint by more than 50%. [Table 1](#pbio-0020271-t001){ref-type="table"} correlates the recovery of sites with the specificity of the weight matrices for two profile value cutoffs. Overall, the recovery ranges from 50% to 100%, with the most specific factors/matrices showing the best recovery. ::: {#pbio-0020271-t001 .table-wrap} Table 1 ::: {.caption} ###### Recovery of Known Binding Sites :::  The table shows the fraction of known maternal and gap factor binding sites recovered by Ahab, with profile value cutoffs of 0.25 and 0.5, respectively. The specificity of the weight matrices ("WM Specificity") is characterized in terms of the distribution of profile values reported by Ahab when run over the sequence of all modules containing known binding sites. The numbers indicate the portion of profile values that exceed the cutoff compared to all profile values; for example, column five for Kr means that 60% of the predicted sites have a profile value greater than 0.50 ::: We further examined whether Ahab misses known sites by misclassification. We found that Ahab generally does not misattribute the missing sites to another factor. A cogent example is provided by Cad and Hb, which have very similar binding sites containing an oligoT stretch. Surprisingly, none of the 21 Hb sites that were missed at a profile value cutoff of 0.5 were misclassified as Cad; conversely, only one of the 11 missed Cad sites was classified as Hb. This discrimination is far better than that achieved by a simple weight matrix scan over the same modules: for this scan, we counted information scores greater than five, which is the score of the weakest experimental binding sites, and overlaps between matrix and binding site of 50% or more. The matrix scans correctly classified 29/43 Hb sites and 16/21 Cad sites; but misclassified five Hb sites as Cad and two Cad sites as Hb. Taken together, Ahab finds the majority of known binding sites and rarely misclassifies; it is thus a reliable indicator of module composition. A complete listing of the integrated profile values reported by Ahab for known, newly validated, and predicted modules is available in [Dataset S6](#sd006){ref-type="supplementary-material"} (mg run) and [Dataset S7](#sd007){ref-type="supplementary-material"} (mgpr run). ### Recovery of modules in *Drosophila pseudoobscura* {#s4a5} To assess the conservation of known and Ahab-predicted modules, we aligned D. melanogaster and D. pseudoobscura genomic sequence as described in [@pbio-0020271-Emberly1] and ran Ahab over the aligned D. pseudoobscura control regions, with D. melanogaster weight matrices as input and with cutoffs for significant predictions (15 in D. melanogaster) and subthreshold peaks (12 in D. melanogaster, equal to genome-wide mean plus three standard deviations) set to obtain equivalent numbers of predictions in *D. pseudoobscura. D. pseudoobscura* predictions were then mapped to D. melanogaster coordinates and examined for overlap with the known and predicted D. melanogaster modules. ### Processing of Ahab output and module classification {#s4a6} To associate predictions from different Ahab runs, each run was processed and the highest point on the free energy plot within an interval of the window size was marked as a "peak." Peaks are thus spaced by at least the window size. Peaks in two different runs correspond if they are closer than half the window size; their correspondence is unique and order independent. For the three-way comparison, the mg and pr runs were separately compared to the mgpr run. In no case did mg run and pr run peaks correspond without at least one of them matching a mgpr run peak. For the purposes of broadly classifying predicted modules as to type of input, we defined four classes: mostly maternal/gap input, mostly pair-rule, and mixed input but with a bias towards maternal/gap or pair-rule. Two classification methods were used. The first relied on a single Ahab run with all factors (mgpr run) and then compared the sum of the maternal/gap factor profile values for a given module with the sum of the pair-rule profile values, after normalization to make the mean and standard deviation of maternal/gap profile values over all peaks equal to the mean and standard deviation of all pair-rule profile values. An alternative scheme used the free energy plots for the three runs (mg, pr, and mgpr), identified corresponding peaks, and then compared their rank in the different runs. The two methods yielded very similar results. Since Ahab does not adapt its window size to the data, modules that are wider than the window size needed to be delineated to be captured accurately. To this end, we defined the start of the module as the first local maximum in the free energy plot that is above the cutoff. The end point was initialized as the other end of the corresponding Ahab window. The plot was then scanned from left to right, and when another local maximum or rise in window score above the cutoff was encountered, the end point of the module was reset as the end of the corresponding window. The sequence of all delineated predicted modules is available in [Dataset S5](#sd005){ref-type="supplementary-material"}. ### Molecular biology and RNA in situ hybridization {#s4a7} Module predictions were tested as follows. The module was delineated within the genomic sequence as described above and further expanded to include good primer sites for touch-down PCR. Primers were designed following manufacturer\'s guidelines (Clontech, Palo Alto, California, United States), restriction sites (Xba, Asp718) were added for subsequent cloning. Genomic PCR products were cloned into TOPO (Invitrogen, Carlsbad, California, United States), sequenced to confirm identity, and subcloned into Casper hs43ßGAL ([@pbio-0020271-Thummel1]). A Fasta file with the primers and cloned regions is available in [Dataset S4](#sd004){ref-type="supplementary-material"}. Transgenic fly strains were generated using standard methods. For each construct three independent insertions were analyzed for expression patterns by RNA in situ hybridization with a *lacZ* probe. RNA in situ hybridizations were carried out as described by Noordermeer and Kopczynski (<http://www.fruitfly.org/about/methods/RNAinsitu.html>). ### Delineation of protein and transcript patterns {#s4a8} The protein expression profiles of the maternal and gap input factors were obtained from <http://flyex.ams.sunysb.edu> ([@pbio-0020271-Myasnikova1]) (temporal class 4, 10% strip, normalized and registered by FRDWT, averaged over 5% EL). In cases where these data were not available, input factor expression profiles were inferred from literature (D-Stat) or our own data (TorRE, measured by expression of the negative regulator Capicua). The output transcript patterns of segmentation gene modules were determined using images of our own RNA in situ hybridizations of blastoderm embryos, and complemented by data from the literature. Embryos were viewed in the sagittal plane, and the intersection of the domain boundaries with the longitudinal axis was determined and calculated as percent EL. Measurements were performed using the Zeiss (Oberkochen, Germany) Axiovision 3.1 measurement tool and averaged over 2--5 embryos. A complete listing of the references for the expression patterns of segmentation genes is found in [Dataset S1](#sd001){ref-type="supplementary-material"}. To generate the plots in [Figure 5](#pbio-0020271-g005){ref-type="fig"}A, we calculated, for every input factor and for every position along the ap axis, the average of the integrated profile values reported by Ahab for the modules driving expression at that position. Values were calculated in 1% EL increments, then averaged over 5% EL. Supporting Information {#s5} ====================== The Gbrowse display of free energy profiles for genome-wide Ahab runs (mg, pr, mgpr) can be viewed at <http://edsc.rockefeller.edu/cgi-bin/gbrowse_ms/cgi-bin/gbrowse?src=fly>. Dataset S1 ::: {.caption} ###### Segmentation Genes Referred to in This Study The dataset gives name, symbol, flybase identifier, and references for expression pattern, control region dissection, and binding site information. (178 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S2 ::: {.caption} ###### Compilation of Position Weight Matrices and Binding Sites Used in This Study (8 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S3 ::: {.caption} ###### Sequence Information for Known Segmentation Modules in Fasta Format (80 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S4 ::: {.caption} ###### Sequence Information for Transgenic Constructs Used in This Study in Fasta Format (37 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S5 ::: {.caption} ###### Sequence Information for Ahab-Predicted Modules in the Control Regions of 48 Segmentation Genes Data based on mg run, Fasta format. (53 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S6 ::: {.caption} ###### Profile Value Output for Ahab Mg Run Input: Bcd, Hb, Kr, Gt, Kni, Tll, Cad, TorRE, and Dstat. Performed over defined sequences of known modules ([Dataset S3](#sd003){ref-type="supplementary-material"}), tested constructs ([Dataset S4](#sd004){ref-type="supplementary-material"}), and Ahab-predicted modules ([Dataset S5](#sd005){ref-type="supplementary-material"}). (62 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S7 ::: {.caption} ###### Profile Value Output for Ahab Mgpr Run Input: Bcd, Hb, Kr, Gt, Kni, Tll, Cad, TorRE, Dstat, H, Eve\_HD, Eve\_t2, Run, Ftz, Ftz-f1, Ttk, and Prd\_HD. Performed over defined sequences of known modules ([Dataset S3](#sd003){ref-type="supplementary-material"}), tested constructs ([Dataset S4](#sd004){ref-type="supplementary-material"}), and Ahab-predicted modules ([Dataset S5](#sd005){ref-type="supplementary-material"}). (77 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a8} ----------------- The FlyBase (<http://flybase.bio.indiana.edu>) accession numbers for the genes and gene products discussed in this paper are Bcd (FBgn0000166), Cad (FBgn0000251), Capicua (FBgn0028386), *cnc* (FBgn0000338)*, D* (FBgn0000411), D-Stat (FBgn0016917), Eve (FBgn0000606), *fkh* (FBgn0000659)*,* Ftz (FBgn0001077), Ftz-f1 (FBgn0001078), Gt (FBgn0001150), H (FBgn0001168), Hb (FBgn0001180), Kni (FBgn0001320), *knrl* (FBgn0001323)*,* Kr (FBgn0001325), *nub* (FBgn0002970), *oc* (FBgn0004102), *odd* (FBgn0002985)*, pdm2* (FBgn0004394), Prd (FBgn0003145), Run (FBgn0003300), *slp2* (FBgn0004567), Tll (FBgn0003720), TorRE (cf. FBgn0003733), and Ttk (FBgn0003870). We thank Saurabh Sinha for customizing GBROWSE so that it can display multiple free energy plots superimposed on the genome annotations. We are very grateful to D. Arnosti, C. Desplan, S. Small, and J. Reinitz for sharing materials, information, and, above all, strong opinions and lively discussions. This work was in part supported by National Science Foundation DMR 0129848 (EDS) and National Institutes of Health grant GM066434 (EDS and UG). **Conflict of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** MDS and UG conceived and designed the experiments. MP, JF, and HQF performed the experiments. MDS, UU, EDS, and UG analyzed the data. UU and MP generated the figures. MDS, EE, and NR performed Ahab runs. UG wrote the paper. Academic Editor: Mark A. Krasnow, Stanford University School of Medicine ¤ Current address: Department of Biology, New York University, New York, New York, United States of America Citation: Schroeder MD, Pearce M, Fak J, Fan HQ, Unnerstall U, et al. (2004) Transcriptional control in the segmentation gene network of *Drosophila.* PLoS Biol 2(9): e271. ap : anterior-posterior Bcd : Bicoid Cad : Caudal *cnc* : *cap \'n\' collar* *D* : *Dichaete* D-Stat : Stat92E EL : egg length Eve : Even skipped *fkh* : *fork head* Ftz : Fushi tarazu Ftz-f1 : Fushi tarazu transcription factor 1 Gt : Giant H : Hairy Hb : Hunchback Kni : Knirps *knrl* : *knirps-like* Kr : Kruppel mg run : maternal/gap run mgpr run : combined maternal/gap and pair-rule run *nub* : *nubbin* *oc* : *ocelliless* *odd* : *odd skipped* *pdm2* : *POU domain protein 2* pr run : pair-rule run Prd : Paired Run : Runt *slp2* : *sloppy paired 2* Tll : Tailless TorRE : Torso-response element Ttk : Tramtrack

PubMed Central

2024-06-05T03:55:47.258595

2004-8-31

PMC514886

Health food stores have long hawked fish oil capsules as a cure-all for everything from migraines to heart disease. And though such claims are often weak on scientific evidence, fish oil, it turns out, is no snake oil. A recent review of scientific studies concludes that omega-3 fatty acids can indeed protect against heart disease, and the American Heart Association now recommends fish oil capsules for patients with coronary heart disease. Fatty acids come in hundreds of varieties, distinguished primarily by their structure, which in turn determines their physiological role. Unlike proteins or genes---which are polymers made up of amino acids and nucleotides, respectively---fatty acids are a large group of compounds containing long chains of carbon and hydrogen atoms with a carboxylate group (acid) attached at the end. It is this asymmetrical chemical configuration that gives fatty acids their unique properties. Fatty acid diversity comes from variations in the length of the carbon chain and in the number of double bonds between carbons. Fatty acids with one or more double bonds are called unsaturated fatty acids. Fatty acids play an essential role in metabolism, providing the cell with a concentrated source of energy, and form the structural foundation of the cell membrane, where they are most conspicuous and perhaps best understood. Long-chain (unbranched) fatty acids, which run ten to 22 carbons long, are the most common fatty acids in animal cells and the most studied. One much less understood class of fatty acids---the monomethyl branched-chain fatty acids (mmBCFAs)---has been found in organisms from bacteria to humans, but its role remains obscure. In this issue of *PLoS Biology*, Marina Kniazeva et al. explore the origin and function of mmBCFAs in the worm Caenorhabditis elegans and find that these relatively obscure fatty acids play a crucial role in growth and development. mmBCFAs are abundant in diverse genera of bacteria, which use a supply of branched-chain amino acids and enzymes to assemble the fatty acid chains. mmBCFA biosynthesis has been characterized in bacteria, but not in eukaryotes. (Worms, and humans, are eukaryotes; our cells have nuclei.) Here, Kniazeva et al. identified worm genes that are homologous to the gene that codes for an enzyme called elongase in another eukaryote, yeast. Elongases are enzymes that extend the length of fatty acid chains by two carbons. To see what kind of fatty acid molecules the homologous worm genes were synthesizing, the authors used a technique called RNA interference (RNAi) to "silence" the genes\' expression in the worms. Surprisingly, two of the eight inhibited genes had a specific effect on branched-chain fatty acid levels: *elo-5* and *elo-6*. Inhibiting *elo-5* function had deleterious effects on the growth and development of the worms. The progeny of worms treated as embryos with RNAi for *elo-5* stopped growing at the first larval stage, while the progeny of worms treated at later stages developed to adulthood but got progressively sicker and showed reproductive problems. These defects were corrected when the researchers fed the mmBCFAs directly to the worms, indicating that these mmBCFAs are essential for normal larval growth and development. Given the widespread distribution of mmBCFAs in organisms as diverse as bacteria and humans, it\'s perhaps not too surprising that they regulate essential physiological functions during animal development. It\'s still not clear, however, what all the components of the fatty acid manufacturing machinery are or how an organism monitors production levels. And though it\'s still an open question as to how these ubiquitous molecules function in mammals, the fact that they have been conserved throughout evolution underscores their importance---and suggests they may play a similar role.

PubMed Central

2024-06-05T03:55:47.264223

2004-8-31

PMC514887

A fundamental principle guiding the publication of scientific results is that the data supporting any scholarly work must be made fully available to the research community, in a form that allows the basic conclusions to be evaluated independently. In the context of molecular biology, this has typically meant that authors of a paper describing a newly sequenced genome, gene, or protein must deposit the primary data in a permanent, public data repository, such as the sequence databases maintained by the DNA Data Bank of Japan (DDBJ), European Bioinformatics Institute (EBI), and National Center for Biotechnology Information (NCBI). Similarly, we, members of the Microarray Gene Expression Data Society (MGED; <http://www.mged.org>), believe that all scholarly scientific journals should now require the submission of microarray data to public repositories as part of the process of publication. While some journals have already made this a condition of acceptance, we feel that submission requirements should be applied consistently and that journals should recognize ArrayExpress ([@pbio-0020317-Brazma2]), Gene Expression Omnibus (GEO) ([@pbio-0020317-Edgar1]), and the Center for Information Biology Gene Expression Database (CIBEX) ([@pbio-0020317-Ikeo1]) as acceptable public repositories. To this end, the members of MGED propose the following as a new paradigm for the publication of microarray-based studies. (1) Authors should continue to take primary responsibility for ensuring that all data collected and analyzed in their experiments adhere to the "Minimum Information about a Microarray Experiment" (MIAME) guidelines and should continue to use the MIAME checklist ([www.mged.org/Workgroups/MIAME/miame\_checklist.html](www.mged.org/Workgroups/MIAME/miame_checklist.html)) as a means of achieving this goal. (2) Scientific journals should require that all primary microarray data are submitted to one of the public repositories---ArrayExpress, GEO, or CIBEX---in a format that complies with the MIAME guidelines. (3) Public databases should work with authors and scientific journals to establish data submission and release protocols to assure compliance with MIAME guidelines. (4) To assist with the review process, the databases should continue to work in collaboration with publishers to provide qualified referees with secure means of accessing prepublication data. Authors should be strongly encouraged to submit data to the databases during review. Naturally, data should be protected from general release prior to either publication or authorization from the data submitters, whichever comes first. At a minimum, journals should require valid accession numbers for microarray data as a requirement for publication, and these accession numbers should be included in the text of the manuscript to allow members of the community to find and access the underlying data. Since its inception in 1999, MGED has been working with the broader scientific community to establish standards for the exchange and annotation of microarray data. In December 2001, we proposed the MIAME guidelines ([@pbio-0020317-Brazma1]) and requested that interested parties provide feedback on its relevance and utility. The feedback from both researchers and scientific journals was overwhelmingly positive, yet almost everyone who responded also asked for help in implementing these guidelines. Subsequently, in the summer of 2002, we submitted an open letter to various journals (e.g., [@pbio-0020317-Ball1], [@pbio-0020317-Ball2]) urging the community to adopt the MIAME requirements for microarray data publication. We provided a checklist so that authors could ensure that sufficient information to allow their data to be re-analyzed by others would be available. Again, the response from the community was extremely positive, and most of the major scientific journals now require publications describing microarray experiments to comply with the MIAME standards. While the adoption of these standards has greatly improved the accessibility of microarray data, much of it remains on individual authors\' websites in a variety of formats; consequently, obtaining and comparing datasets remains a significant challenge. Clearly we need additional requirements for publication that include submission of expression data to public data repositories. Though one might ask why this requirement was not part of the original MIAME recommendation, the answer is quite simple---MIAME was ahead of its time. While NCBI and the EBI had developed nascent microarray data repositories, and work was underway to create a similar database at the DDBJ, submitting data to these databases was a considerable burden for authors. However, since that time, improvements in the data-entry utilities available for GEO ([www.ncbi.nlm.nih.gov/geo](www.ncbi.nlm.nih.gov/geo)), ArrayExpress ([www.ebi.ac.uk/arrayexpress](www.ebi.ac.uk/arrayexpress)), and CIBEX ([cibex.nig.ac.jp](cibex.nig.ac.jp)), as well as a growing number of commercial and academic software packages capable of writing MAGE-ML documents ([@pbio-0020317-Spellman1]) that can be directly submitted to these public databases, have lowered the barriers for data submission to the point where we as a community **must** now reconsider that submission to one of these databases be a requirement. Requiring authors to submit microarray data to a public database will provide a number of distinct advantages to the entire research community. (1) These established repositories have a commitment to continued community service and to providing some level of assurance that published gene expression datasets will continue to be available into the future. (2) Having the data available in these public repositories in a standardized format will not only make them more accessible, but it will allow expression data to be integrated with other relevant data, including the available genome sequences, single nucleotide polymorphism and haplotype mapping information, the literature, and other resources that can aid in further interpretation of expression patterns. Although many authors now provide some or all of this information, the established databases are much more likely to assure that these links are maintained and current. (3) Curation of data submitted to public data repositories will assist authors, reviewers, and publishers in assuring that the data comply with the MIAME requirements, further enhancing their utility. (4) The standardization of microarray data formats will enable the development of additional data analysis and integration tools and makes it easier for scientists to access, query, and share data. (5) Finally, submission prior to publication will make it easier for referees to access the data confidentially, facilitating the review and publication process. In the same way that availability of sequence data had a profound impact on a wide range of disciplines, we believe that requiring that microarray data be deposited in public repositories as a necessity for publication will accelerate the rate of scientific discovery. What this proposal requires is a change in the way in which we approach the publication of microarray-based studies. Both authors and journals have a responsibility to assure that the requisite data are available, and because submitting MIAME-compliant data can take considerable time and effort, this process should be factored into review and publication timelines. However, while this process may be time consuming and painful at first, we believe that the benefits of building an open repository of microarray data will far outweigh any initial disadvantages. As always, it is our sincere hope that these suggestions stimulate discussion within the community and that together we can arrive at a consensus that ensures that microarray data are widely and easily accessible. Finally we would like to urge the DDBJ, EBI, and NCBI to work together towards exchanging all MIAME-compliant microarray data. Catherine A. Ball is in the Department of Biochemistry, Stanford University School of Medicine, Stanford, California, United States of America; Alvis Brazma is at the European Bioinformatics Institute, European Molecular Biology Laboratory, Hinxton, United Kingdom; Helen Causton is at the Clinical Sciences Centre/Imperial College Microarray Centre, Imperial College, London, United Kingdom; Steve Chervitz is Bioinformatics Engineer at CIS Enterprise Data Group, Affymetrix, Emeryville, California, United States of America; Ron Edgar is at the National Center for Biotechnology Information, National Library of Medicine, Bethesda, Maryland, United States of America; Pascal Hingamp is at the Laboratoire Technologies Avancées pour le Génome et la Clinique, Faculte des Sciences de Luminy, Centre d\'Immunologie de Marseille Luminy, Universite Aix-Marseille-II, Marseille Cedex, France; John C. Matese is at the Lewis-Sigler Institute for Integrative Genomics, Carl Icahn Laboratory, Princeton University, Princeton, New Jersey, United States of America; Helen Parkinson is at the European Bioinformatics Institute, European Molecular Biology Laboratory, Hinxton, United Kingdom; John Quackenbush is at The Institute for Genomic Research, Rockville, Maryland, United States of America; Martin Ringwald is at the Jackson Laboratory, Bar Harbor, Maine, United States of America; Susanna-Assunta Sansone is at the European Bioinformatics Institute, European Molecular Biology Laboratory, Hinxton, United Kingdom; Gavin Sherlock is in the Department of Genetics, Stanford University School of Medicine, Stanford, California, United States of America; Paul Spellman is Computational Scientist at Lawrence Berkeley National Laboratory, Berkeley, California, United States of America; Chris Stoeckert is Research Associate Professor in the Department of Genetics, Center for Bioinformatics, University of Pennsylvania, Philadelphia, Pennsylvania, United States of America; Yoshio Tateno is Professor at the Center for Information Biology and DNA Data Bank of Japan, National Institute of Genetics, Research Organization of Information and Systems, Mishima, Japan; Ronald Taylor is Senior Research Scientist II at Computational BioSciences Group, Biological Sciences Division, Pacific Northwest National Laboratory, Richland, Washington, United States of America; Joseph White is at The Institute for Genomic Research, Rockville, Maryland, United States of America; and Neil Winegarden is at the University Health Network, Microarray Centre, Toronto, Ontario, Canada. Citation: Ball CA, Brazma A, Causton H, Chervitz S, Edgar R, et al. (2004) Submission of microarray data to public repositories. PLOS Biol 2(9): e317. CIBEX : Center for Information Biology Gene Expression Database DDBJ : DNA Data Bank of Japan EBI : European Bioinformatics Institute GEO : Gene Expression Omnibus MIAME : Minimum Information about a Microarray Experiment MGED : Microarray Gene Expression Data Society NCBI : National Center for Biotechnology Information

PubMed Central

2024-06-05T03:55:47.264949

2004-8-31

PMC514888

To build the complex body plan of higher organisms, thousands of genes must act in a coordinated fashion, becoming active at the right time and in the right place to define structures like head, thorax, and abdomen, or cell types like skin, muscle, and bone. One of the central questions for developmental biologists is how such specific spatiotemporal expression of genes is achieved. The general mechanism of the control of gene expression is well understood: Special proteins, called transcription factors, bind to short stretches of DNA near a gene. By docking to such binding sites, they activate or repress the transcription of the gene into mRNA (which is then translated into protein). Transcription factors often act in a combinatorial fashion---that is, several different factors have to bind in close proximity to each other to achieve a particular transcriptional outcome. As a consequence, their binding sites form clusters, called regulatory elements or modules. In many contexts, the genes that are activated or repressed encode transcription factors themselves, forming a cascade of transcriptional control events. One such transcriptional control hierarchy is the segmentation gene network in the fruitfly Drosophila. Organized in four tiers and acting in combinatorial fashion, the segmentation genes lay out the anterior-posterior axis of the embryo. In a stepwise refinement of expression patterns, they translate broad, overlapping gradients formed by maternally provided transcription factors into a periodic pattern of 14 discrete stripes that prefigure the 14 segments of the larva. The segmentation gene network has long been one of the prime paradigms for studying transcriptional control, and many researchers have worked over the years to experimentally dissect the regulatory interactions within the hierarchy. For some of the most important genes, the regulatory elements driving their expression and the favored binding sites have been identified. Nevertheless, the picture of transcriptional regulation within the segmentation gene network has remained incomplete. This is where the research reported by Mark Schroeder et al. comes in: With the sequence of entire genomes available, it\'s possible to use existing binding site information to computationally search the neighborhood of genes for regulatory elements. The difficulty here is that in higher organisms such as Drosophila, the binding sites are typically short and variable, and the search space is large; on the other hand, the fact that sites cluster---where transcription factors work in concert---aids the task. To identify regulatory elements, the researchers developed an algorithm, named Ahab, that models the behavior of multiple transcription factors competing for binding sites and fine-tunes the search by detecting clusters of weak sites. Using this approach, Schroeder et al. identified 52 regulatory elements within the segmentation gene network, 32 of them novel. The authors tested a large number of the newly identified modules experimentally by placing them in front of reporter genes that reveal where the modules drive expression within the developing fly. They showed that almost all modules faithfully reproduce the expression pattern of the endogenous gene. To better understand the way segmentation gene modules function, the researchers then systematically analyzed their predicted binding site composition. They correlated the composition of modules with the expression they produce and with the distribution of the transcription factors that bind to them. They were thus able to glean basic composition rules and to derive the mode of action for most of the factors, that is, whether they act as activators or as repressors.[](#pbio-0020319-g001){ref-type="fig"} ::: {#pbio-0020319-g001 .fig} ::: {.caption} ###### Segmentation in the early Drosophila embryo :::  ::: Overall, Schroeder et al. show that a computational search can greatly reduce the experimental effort necessary for finding regulatory elements within the genomic sequence. Their study provides an example of how experimental and quantitative methods can be combined to achieve a more global analysis of the regulatory interactions within a transcriptional network.

PubMed Central

2024-06-05T03:55:47.266146

2004-8-31

PMC514889

Identifying the causative mutation for a disease can be the first step to a potential cure. This task is not always trivial. Often the initial strategy is to look for the variations within a mutated gene that alter its protein coding sequence, as these mutations often alter the gene\'s function. However, in a growing number of cases, the causative mutation is a "synonymous" mutation---a change in the coding sequence of a gene that doesn\'t change the sequence of the protein coded by the gene. This type of mutation may be responsible for Seckel syndrome, a human disease characterized by dwarfism. In Seckel syndrome, the mutation doesn\'t alter the protein sequence itself but instead results in the skipping over of a portion of the protein coding sequence (an exon), a process called altered splicing. The disease-causing potential of this type of splicing mutation has only recently gathered attention. Splicing assembles the exons of a transcribed gene (the RNA copy) into the right order while removing the non-coding sequences of the RNA (introns). This highly regulated process is coordinated by a number of sequences within a gene, including splice sites that precede and follow the exon, as well as by exonic splicing enhancers (ESEs), which help recruit the factors (proteins) necessary to insure proper splicing. Although splice sites have optimal (consensus) sequences, there is some variability amongst individual splice site sequences that allows splicing to take place to a greater or lesser extent. ESEs facilitate splicing, especially when a gene\'s splice sites vary from the consensus sequence. Candidate ESEs have previously been identified based on their more frequent occurrence in exons that are adjacent to non-consensus splice sites. In this issue of *PLoS Biology*, William Fairbrother et al. investigated the functionality of these putative ESEs. If they are functional, the authors reasoned, then mutations that disrupt them would be selected against---that is, these mutations would tend to be discarded---in the human genome. To this end Fairbrother et al. developed a computational method, which they call VERIFY (for "variant elimination reinforces functionality"), to evaluate the selective pressure on ESEs. They took advantage of a public database of all single nucleotide polymorphisms (DNA changes at a single point) within the human genome and compared them to the chimpanzee genome; this allowed the authors to infer the identity of the ancestral gene (or allele). By determining which allele is ancestral and which is the variant, the researchers could then distinguish the mutations that created ESEs from mutations that disrupted ESEs. Mutations that altered or disrupted ESEs were under-represented, leading the authors to conclude that predicted ESE sequences evolve under a more stringent level of selection than exonic sequences with no predicted ESEs. This selective pressure was greater for predicted ESEs located near the splice signals than for ESEs that were located within the exon. This result was consistent with experimental findings that ESE strength diminishes with distance from the splice site. As more vertebrate genomes are sequenced and the public database of single nucleotide polymorphisms continues to grow, this type of computational method will become increasingly valuable. It can help confirm the functionality or role of candidate regulatory elements thought to control various aspects of gene expression, and in so doing, offer insights into the complex machinations required to maintain the healthy operation of the human genome.

PubMed Central

2024-06-05T03:55:47.266835

2004-8-31

PMC514891

Background ========== In an era of cost containment, the need for rigorous examination of the cost-effectiveness of drugs, as well as their clinical effectiveness, is widely recognized not only by governments but also by the pharmaceutical industry \[[@B1]-[@B4]\]. Messages framed differently, but with the same basic content, have reached the community of prescribing physicians, who have come to understand that, although an effective drug may be prescribed for patients who would benefit from it, unless the drug is cost-effective, the resources that are expended might produce greater benefits for other patients. Such messages and updated recommendations to prescribing doctors, in addition to results derived from recent large, randomized trials, continue to have only minimal, if any, impact on the prescribing habits of doctors. The latest such example \[[@B5]\] concerns the outcomes of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT), indicating that diuretics could be more effective than angiotensin-converting enzyme (ACE) inhibitors or calcium channel blockers in the treatment of hypertension, and at a much lower cost \[[@B6]\]. Despite this convincing evidence, a study presented at the annual conference of the American Heart Association in March 2004 showed that spending on antihypertensive drugs essentially doubled (from approximately \$6 billion to approximately \$12 billion) between 1990 and 2002. The explanation most commonly offered is that \"doctors selected the more costly antihypertensive agents.\" Since cost-effectiveness is conventionally required for evaluating the efficacy of alternative healthcare interventions, the perspective commonly taken is that of the health services \[[@B7]\]. Therefore, establishing the superiority or equivalence of a new intervention relative to the standard one has been extended not only to new drug entities, but also to generic versions of innovator drugs \[[@B8]\], surgical techniques \[[@B9]\], medical devices \[[@B10]\], and such diverse factors as medical protocols \[[@B11]\]. Cost-effectiveness and clinical effectiveness should be pursued simultaneously to ensure that health care is efficient, ethical, and beneficial to patients. This paper deals with only one aspect of clinical effectiveness: drug treatment benefit and how it may be ascertained from claims of therapeutic equivalence. Ethical Issues ============== The basis for the scientific and ethical underpinnings for the design and conduct of randomized clinical trials is the *uncertainty principle*, which states that a patient should be enrolled in a randomized controlled trial only when substantial uncertainty exists as to which of the trial treatments would benefit the patient more \[[@B12]\]. From this principle derives the fundamental ethical challenge of equivalence trials, reflected in the researcher\'s explicitly expressed belief that \"the new drug might be not different from the old drug,\" a fact that should be acknowledged in the consent process whereby trial subjects are informed that \"it is not known which drug is better or whether they are the same.\" Nevertheless, demonstrating equivalence of the drugs being compared implies starting from the assumption that the new drug is better. In other words, the hypothesis to be tested in equivalence trials (and the hypothesis that is refuted if equivalence is shown) is that one treatment is superior to the other. Altruistic patients are more likely to agree to participate in such a trial, whereas other, less altruistic patients are more likely to decline participation, as their interest lies in treatments with proven efficacy. Obviously, this situation is more patient-favorable than are placebo-controlled trials, in which the individual patient\'s well-being may be subordinated to the good of others \[[@B13],[@B14]\]. Placebo-controlled trials are still used extensively to demonstrate the effectiveness of new drugs; however, a paradigm shift appears to be steadily emerging in this area \[[@B15],[@B16]\]. Speaking metaphorically, Urquhart stated that placebos are predestined to be \"roadkill on the highway of medical progress\" \[[@B17]\]. For circumstances in which no increased risk for patients is foreseen, use of placebo-controlled trials seems appropriate and ethical, provided the patients are fully informed and that they give their written, informed consent. However, if these patients and their doctors were to find the placebo-controlled studies inappropriate, and if they were to exercise their option in large numbers, these studies would become unfeasible, regardless of the ethical justifications, scientific considerations, views of the trial sponsor, or, ultimately, the expectations of regulatory authorities. Apart from the extreme opinions that challenge the placebo-controlled trials as unethical \[[@B18]-[@B24]\] and those that advocate proactive use of the active-controlled equivalence trials \[[@B25],[@B26],[@B31]\] or question their scientific merits \[[@B16],[@B26]-[@B32]\], a balanced approach is needed (i.e., one that recognizes the use of placebos in instances wherein efficacy cannot otherwise be demonstrated, and the use of active-controlled trials as the design of first choice when scientifically sound circumstances require it). Simultaneous use of both alternatives might be necessary in selected cases. The issue of *assay sensitivity* ================================ Defined as \"the ability of a study to distinguish between active and inactive treatment,\" assay sensitivity is the *sine qua non*for the validity of equivalence claims derived from any active-controlled equivalence/noninferiority study. Methodological flaws affecting one or several of the specific elements inherent in assay sensitivity itself seem to have been more a rule than an exception in many trials carried out during the past decade. Illustrating this point is a systematic review of trials published between 1992 and 1996 that claim equivalence \[[@B33]\]. In the review, the authors showed that: • 88 papers were evaluated for five equivalence-specific methodological attributes. • Only 45 (51%) of the 88 reports specifically identified demonstration of equivalence as their aim; the others attempted to show superiority or did not state any research aim. • An equivalence boundary was set and confirmed with an appropriate statistical test in 23% of the reports; in 67% of reports, equivalence was declared after a failed test for comparative superiority; in 10%, the claim of equivalence was not evaluated statistically. • Sample sizes were calculated in advance in 33% of reports. • In 25% of reports, sample size was 20 patients per group or fewer. The main concern with such \"equivalence claims\" is certainly the risk of harm to patients, as poor sensitivity has the potential to cause a type II error (false conclusion of no efficacy) and thereby to thwart satisfaction of public health needs for effective medicines. Just as important as paying careful attention to all aspects of assay sensitivity is acknowledging from the outset that a large number of pharmaceutical products present sensitivity problems. That is, agents otherwise known to be clinically effective are often indistinguishable from placebo in well-designed and well-conducted trials. For this reason, such drugs are useless as comparators in active-controlled trials. A typical example is ondansetron, an antiemetic that, despite its known clinical effectiveness, showed no effect in many placebo-controlled trials \[[@B34]\]. Claims of equivalence of a new antiemetic agent with ondansetron would therefore be unreliable, given the lack of assay sensitivity of ondansetron (despite many trials in which it had proven to be superior in comparison with placebo). Similar examples include agents belonging to the class of antidepresssants, analgesics \[[@B35]\], beta-blockers used in postinfarct patients \[[@B36]\], antihypertensives, ACE inhibitors used in patients with heart failure, antianginal agents, and antihistamines. The explanation for this serious problem lies in the great variability of the random placebo effect, which at times may profoundly confound the direction and the magnitude of treatment effects, especially in studies based on small sample sizes \[[@B37]\]. Regarding the example of ondansetron, the incidence of nausea and vomiting ranged from 10% to 96% in the placebo-controlled trials. Furthermore, regarding situations in which multiple trials have demonstrated the efficacy of the active control when compared to placebo, the potential exists for referral bias due to eventual nonreporting of negative results. The risk in such instances is that the smallest clinically relevant effect of the control drug may not be valid \[[@B38]\]. Rationale for choice of active control ====================================== In contrast to the scenarios described above, equivalence and noninferiority trials should be undertaken only when a well-proven standard therapy exists (i.e., when the intended control drug is accepted as the standard of care for the particular indication). Investigators should be confident that the efficacy of the control drug was proven to be superior in a previous placebo-controlled trial and that this efficacy will be preserved under the conditions of the current trial (i.e., the control drug has an established, predictable and quantifiable effect). Doubts about the validity of these assumptions mean uncertainty as to whether the two drugs in the current trial, which are allegedly equivalent, really are effective to a similar degree, or are equally ineffective, or cannot be evaluated definitively because the trial design was inadequate to demonstrate the real differences between the two agents. The goal of showing equivalence =============================== A recent editorial by Alderson \[[@B39]\] concluded as follows: \"We need to create a culture that is comfortable with estimating and discussing uncertainty.\" This observation applies especially to the field of equivalence/noninferiority trials. Increasing the degree of certainty in these trials is a matter of paying careful attention to the elements of study design, conduct, and analysis -- all supposed to mirror as closely as possible the design, conduct, and analysis performed in previous evaluations of the current active control against placebo. Such trials should be reported in a transparent and explicit fashion, to acknowledge that they are not really equivalent to superiority trials \[[@B40]\]. The primary objective of equivalence/noninferiority trials is to demonstrate that the efficacy of the new treatment matches that of the control treatment. However, \"equivalence\" should not be interpreted to mean 100% (absolute equivalence can never be demonstrated), but that despite some degree of difference, the two agents are clinically indistinguishable. Closer scrutiny should be afforded the secondary objectives of the study, as they might demonstrate some sort of superiority over the control, such as a more favorable safety profile, easier administration, or reduced cost. Alternatively, results might indicate that the new agent would be a reliable second-line treatment. All too often in the past, when trials that were designed to demonstrate the superiority of an agent over its comparator failed to reject the null hypothesis (i.e., a statistically significant difference was not demonstrated), results were interpreted as proof of the equivalence of the two drugs. A dangerous mismatch of the goals of the superiority and equivalence trials arises when the general reasoning employed in planning and evaluating superiority trials is simply extrapolated to active-controlled trials. The aim of the superiority trial is to rule out the equality of the two agents being compared by rejecting the null hypothesis that the two agents are the same. Failure to reject the null hypothesis does not mean that equivalence can be assumed. Lack of superiority might be consistent with equivalence but does not prove it. In other words, \"absence of evidence of a difference is not evidence of absence of a difference\" \[[@B41]\]. In equivalence trials, the goal is to rule out all differences of clinical importance between the two agents being compared. This goal is accomplished by rejecting the null hypothesis that the smallest difference of clinical importance exists in favor of the standard-of-care regimen (i.e., in favor of the active control in the current trial). Therefore, establishing equivalence is contingent upon determining what specifically and precisely constitutes a clinically important difference. This process translates into the need to prove that the two interventions do not differ by more than a certain amount, defined as the \"equivalence margin\" (i.e., the tested agent is not inferior to the active control by more than the predefined margin). Methodological requirements =========================== Patient compliance with therapy ------------------------------- To assure the adequacy of the compliance component of assay sensitivity, prescreening of subjects selected to participate in active-controlled trials is necessary, as is reliable assessment of patient compliance with the trial requirements by means of appropriate methodologies \[[@B42],[@B43]\]. Commonly, compliance is defined as the degree of correspondence between the patient\'s current dosing history and the prescribed drug regimen \[[@B44]\]. This seemingly simple definition covers the wide variability in patient compliance in the use of prescribed drugs. The degree of drug exposure has an impact on important clinical outcome variables and cost-effectiveness parameters \[[@B45],[@B46]\]. Knowledge of the drug\'s kinetics and dynamics may allow pharmacokinetic/pharmacodynamic modeling, to address the consequences of temporal dosing patterns that result from variations in patient compliance with the recommended treatment regimen \[[@B17]\]. The most compelling example of treatment noncompliance occurs with antihypertensive medications. Noncompliance seems to be the main reason that blood pressure is adequately controlled in fewer than one fourth of patients treated for hypertension, both in the US and in European countries \[[@B47],[@B48]\]. The classical \"pill count\" method of assessment grossly overestimates patient compliance, as self-reporting of medication use is highly skewed toward reports of excellent compliance \[[@B49]\] Over-reliance on inaccurate self-reports of compliance in research studies can result in misleading conclusions about both the efficacy of treatment and the dose-response relationships \[[@B48]\]. Electronic pill boxes that register the date and time of each access have become the \"gold standard\" and could be a valuable complement to conventional self-reporting of compliance \[[@B50],[@B51]\]. Furthermore, compliance with the protocol-specified regimen can be improved by prescreening patients who are eligible for recruitment to active-controlled studies, with the aim of assessing the ability of individuals to comply with study-specific requirements. Concomitant medication ---------------------- Use of co-medication during active-controlled studies, whether the result of self-medication or prescription medication, can distort the study\'s final results. Co-medication can interfere with response to the tested drug or the control drug, or it can influence the trial endpoints and lead to false-positive conclusions of equivalence. Use of non-trial medication is quite common in clinical trials in general and should be assessed and minimized, particularly in active-controlled studies. Patients\' baseline characteristics and outcome features -------------------------------------------------------- A basic assumption is that the active agent in an equivalence/noninferiority study should have retained its known (historical) effect, demonstrated in a previous placebo-controlled comparison. Patients participating in the current study should be as similar as possible to the patients in the placebo-controlled trial with respect to all baseline values and treatment variables that might influence outcome. These variables include symptoms, signs, risk factors, morbidity, compliance with therapy, responsiveness to drug effects, nonuse of prohibited concomitant medication, consistent diagnostic criteria, inclusion and exclusion criteria, unbiased assessment of endpoints, and reasons for dropping out. Failure to achieve this similarity from the outset, failure to ensure high-quality study conduct, or both, can introduce bias into the study and compromise assay sensitivity. The classical method to minimize systematic differences between study groups is *randomization*, (i.e., random allocation of patients to test or control groups). Further, *double blinding*is intended to minimize potential biases resulting from differences in management, treatment, or assessment of patients, or differences in interpretation of results that could arise as a result of the subject\'s or investigator\'s knowledge of the assigned treatment. The type and frequency of outcome events in the current study are expected to be similar to those in the placebo-controlled comparison. Substantial differences, resulting most often from an imbalance in one or more of the variables mentioned above would render interpretation of differences between the new therapy and the active control very difficult. For example, because of lower baseline blood pressure values and fewer associated risk factors, patients in a hypertension study may display fewer outcome events. Choice and importance of outcome variables ------------------------------------------ Equivalence trials are commonly designed to demonstrate that the test treatment is similar in efficacy to the active control, the assumption being that the control treatment is effective under the conditions of the current trial. In reality, however, most equivalence trials are actually noninferiority trials, attempting to show that the new drug is not less effective than the control by more than the defined amount (margin). Presence of assay sensitivity is essential for interpretation of such a study. In cases with doubtful assay sensitivity, a three-arm study design (test drug, active control, and placebo) might be optimal. Apart from being more complex and requiring a larger sample size, such a trial offers the advantage of measuring the effect of the test drug versus placebo while allowing comparison of the test drug and active control in a setting in which assay sensitivity is established by the active control-versus-placebo comparison. By making the active groups in such trials larger than the placebo groups, it is possible to increase the precision of the active drug comparison while minimizing the chance that patients will be randomly assigned to placebo groups. Furthermore, this design allows distinction between adverse events due to the drug and those due to underlying disease (\"background noise\") \[[@B52]\]. As mentioned earlier, equivalence/noninferiority trials should not only focus on efficacy, but also should prospectively define an analytical plan for safety assessment as a secondary objective. Accordingly, appropriate statistical power to detect adverse effects is a necessity, as is collection of data on the comparative safety of each treatment. Failure to meet these requirements not only undermines the chance to exploit a potentially favorable safety profile of the test drug versus the control, but also presents the risk of missing dangerous signals with regard to safety. A recent example relates to mibefradil, a calcium inhibitor that appeared to have an excellent safety profile until postmarketing surveillance revealed cases of sudden death in patients at high risk for polymorphic ventricular tachycardia or patients in whom concomitantly administered drugs either inhibited mibefradil metabolism or otherwise amplified its cardiac risk \[[@B53]\]. A more recent experience with COX-2 inhibitors illustrates another significant problem in current evaluations of new drug safety. In the VIGOR study \[[@B54]\], the extensively marketed product rofecoxib appeared to be inferior to naproxen with regard to the frequency of cardiovascular thrombotic events, raising the question of rofecoxib\'s inferiority versus naproxen\'s superiority to an imputed placebo \[[@B55]\]. Confidence interval and sample size =================================== The margin (Δ) itself clearly communicates a judgment as to what is and is not important. The margin defines the largest difference that is clinically acceptable. Setting the margin is critical to the design of both equivalence and noninferiority trials, and it is commonly established for the purpose of excluding a clinically important difference between treatments. However, what constitutes such a difference may vary widely for each patient and clinician and might fall below the margins set by the designer of the trial. For that reason, careful clinical judgment and statistical reasoning should be exercised in selecting a meaningful difference to be ruled out; furthermore, this difference should be specified and justified a priori in all equivalence/noninferiority trials. At the very least, the equivalence margin should be smaller than the lower 95% confidence limit for the absolute risk difference observed between standard therapy and placebo in the relevant superiority trial \[[@B31]\]. That is, this lower boundary of the 95% confidence interval is the smallest expected effect of the control over the placebo, and it should exceed the established margin \[[@B29],[@B38]\]. Confidence interval is the method of choice to interpret equivalence and noninferiority trials. It defines a range for the possible true differences between the test drug and the active control. If every point within this range reflects a difference that is clinically nonrelevant, then the two agents may be considered equivalent. In other words, for an equivalence trial, the two-sided 95% confidence interval -- defining the range of possible differences between the test and the control agent -- should lie entirely within the interval (-Δ to +Δ) (Fig. [1](#F1){ref-type="fig"}, lines c, d, and e). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Confidence interval approach to analysis of equivalence and non-inferiority trials. :::  ::: For a noninferiority trial, the effect of the new drug may be shown to be similar to or greater than that of the control. The possible difference of interest occurs only in the - Δ direction, and the 95% confidence interval should lie entirely to the right of the - Δ value (Fig. [1](#F1){ref-type="fig"}, line a). A p value associated with the null hypothesis of noninferiority can be calculated. A trial that is intended to demonstrate noninferiority may actually allow a claim of superiority for the test drug. In such a case, the one-sided 95% confidence interval lies to the right of not only the - Δ but also the zero line (Fig. [1](#F1){ref-type="fig"}, line g). A p value can be calculated to verify whether the superiority test is sufficiently small to reject the hypothesis of no difference at the 5% α level (p \< 0.05). A claim of superiority, however, would imply a careful assessment of the test drug\'s safety profile, which should be similar to or better than that of the control to increase the strength of the evidence in favor of superiority. A less favorable safety profile raises the question of whether the claimed superiority outweighs the eventual adverse effects and therefore requires a careful quantification of the overall risk-benefit in clinical terms. This latter emphasis is meant as a reminder that claimed superiority of a test agent in the context of a noninferiority trial is, in fact, superiority to \"no treatment,\" based on the proven superiority of the control agent against placebo in a previous trial. Another possible scenario is that of a superiority trial that fails to detect a significant difference between the two agents being compared. An investigator who anticipates this outcome at the outset of the trial may want to downgrade the goal from superiority to noninferiority. That change is legitimate, provided that a noninferiority margin has been prospectively defined and the 95% confidence interval shown to lie to the right of the - Δ (Fig. [1](#F1){ref-type="fig"}, lines a and f). A post hoc definition of the margin is not acceptable. For calculating sample size, values should be specified for the range of equivalence (Δ), and the α (type I error) and β (type II error) values should be selected on the basis of the same principles as for comparative trials. The distinction between one-sided and two-sided tests of statistical significance carries over into the confidence interval approach recommended by the Committee for Proprietary Medicine Products (CPMP) guidelines \[[@B56],[@B57]\], which provide key information for making decisions about equivalence/noninferiority. Unlike superiority trials, in which intention-to-treat analysis is the rule, in equivalence/noninferiority trials both intention-to-treat and per-protocol analysis should be run. Both types of analysis would be expected to lead to similar conclusions for a robust interpretation. In a recent article, Gomberg-Maitland et al. \[[@B58]\] suggested the use of standard guidelines for reporting equivalence/noninferiority trials to facilitate qualitative assessment of the methodology applied (Table [1](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Guidelines for reporting equivalence/noninferiority trials. (*From Gomberg-Maitland et al*). ::: --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 1\. A table comparing the inclusion and exclusion criteria with those of previous trials on which the standard therapy was based. 2\. A flow diagram delineating the number of eligible patients screened, the number randomised, the number of patients assigned to each group, the number of withdrawals and crossovers, and the number of patients in each group who successfully completed the trial on assigned treatment. 3\. A statement on the projected and actual total treatment exposure (patient-years), the minimum per-patient exposure, and the respective impact of withdrawals and crossovers on exposure to initially assigned treatment. 4\. The rationale of setting the margin of acceptable difference with specific reference to the minimum clinically important treatment effect and with the established efficacy advantage for the control over placebo. Where event rate ratios or floating margins are utilized, the rationale for their use, their prespecified criteria for adjustment, and the margin or ratios used to determine sample size should be provided. 5\. The minimum requisite number of primary events should be established at the outset. 6\. A comparison of event rates during treatment with the active control in the trial and in the historical trials that established its efficacy compared with placebo. --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ::: Conclusions =========== Each year, major advances in drug discovery generate a seemingly endless supply of new drugs in virtually all therapeutic areas; however, in most cases these new drugs provide only incremental improvement in efficacy, safety, or the overall risk-benefit ratio. In addition, ethics-based restrictions on the use of placebos as comparators have enhanced the viability of equivalence and noninferiority trials as viable alternatives for registration purposes, for risk management once the drug is on the market, and for a marked increase in confidence in the new drug on the part of doctors and patients. Given the complexity of these study designs, careful attention should be paid to the proper use of specific epidemiological features so as to avoid methodological deficiencies that may harm patients if clinically inferior treatments are erroneously deemed equivalent to the standard of care, or if potentially superior therapies are discarded as merely \"equivalent.\"

PubMed Central

2024-06-05T03:55:47.267386

2004-8-17

PMC514892

Background ========== The superfamily of proteins containing the C-type (Ca-dependent) lectin-like domain (CTLD) is a large group of extracellular proteins characterized by evolutionary flexibility and functional versatility \[[@B1],[@B2]\]. Its members have been extensively studied because of their involvement in diverse physiological processes, and their ability to bind selectively a wide variety of ligands. As the superfamily name suggests, carbohydrates (in various contexts) are primary ligands for CTLDs and this binding is Ca-dependent \[[@B3]\]. However, the fold has been shown to specifically bind proteins \[[@B4]\], lipids \[[@B5]\] and inorganic compounds including CaCO~3~and ice \[[@B6]-[@B9]\]. In several cases, the domain is multivalent and may bind both protein and sugar \[[@B10]-[@B12]\]. Three studies using the whole-genome approach have been published analyzing the distribution of the superfamily in *C. elegans*\[[@B13]\], *D. melanogaster*\[[@B14]\] and human \[[@B15]\]. An early study \[[@B2]\] attempted to generalize findings on vertebrate CTLD-containing proteins (CTLDcps), and to classify them into groups. This classification included 7 groups and, although not sufficient to describe later known CTLDcps even in mammals and other vertebrates, has been widely used by CTLD researchers. The recent work of Drickamer and Fadden \[[@B15]\] provided an updated classification of human and mouse CTLDcps, based on a comprehensive analysis of CTLDcps encoded by the human genome; this comprises 14 groups. These whole-genome studies and genome annotation projects demonstrated the relative abundance of CTLDcps and importance of the domain. Known fish C-type lectins ------------------------- A number of fish CTLDcp sequences have been reported separately in the literature and public sequence databases. The best-studied and most distinct set are serum antifreeze proteins (AFPs) from several cold-water-living species \[[@B7],[@B16],[@B17]\]. These sequences consist mostly of just a CTLD, and were classified as group VII members based on domain architecture. A three-dimensional structure of the sea raven antifreeze protein has been determined experimentally \[[@B18]\]. Apart from AFPs, several other soluble bony-fish CTLDcps have been described: 5 isoforms of *Salmo salar*serum lectin (SSL) \[[@B19]\], three collectins from different Cyprinidae carp family species \[[@B20]\], skin mucus protein AJL-2 \[[@B21]\] and two C-type lectins (eCL-1 and eCL-2) from gills of Japanese eel \[[@B22]\], two lectins from rainbow trout liver \[[@B23]\], a carp lectin \[[@B24]\], goldfish lectin OL-1 (GI: 26000685, unpublished), and a liver lectin from *Gillichthys mirabilis*(long-jawed mudsucker), annotated as \"mannose receptor C\" \[[@B25]\]. Known membrane-bound CTLDcps from bony fishes include a polycystic kidney disease protein 1 (PKD1) orthologue from *Fugu*\[[@B26]\], a rainbow trout Kupffer cell receptor homologue \[[@B27]\], and a set of putative killer cell receptors (KLR) identified recently \[[@B28]\]. Although predicted coding sequences for CTLDcps from winter flounder (GI:28394504, unpublished) and medaka fish \[[@B29]\] do not contain a recognizable transmembrane (TM) domain, based on CTLD sequence and, in the case of the medaka CTLDcp, domain structure, they should be assigned to group II, as the absence of TM regions may be a result of incomplete prediction. The only known CTLDcp sequence from cartilaginous fishes is a tetranectin homologue from reef shark cartilage \[[@B30]\]. Fugu genome sequence -------------------- The *Fugu rubripes*genome, available since 2002 \[[@B31]\], is the second vertebrate genome sequenced. It is 8 times smaller than the human genome and is proving to be an effective instrument in analyzing the human genome because of its compactness, low content of repetitive elements and the relatively large evolutionary distance between fish and mammals, which is estimated to be about 430 Myr \[[@B32]\]. Currently three versions of the *Fugu rubripes*genome assembly are publicly available. The second version of the assembly (v.2), constructed from 4.1 million sequencing reads (5.4 X sequence coverage), was reported in the original publication announcing the completion of the *Fugu rubripes*genome sequencing \[[@B31]\]. The third version (v.3) was released in August 2002, has slightly better coverage (5.7X) and improved scaffold contiguity. Sequence data for all three assembly versions can be downloaded from the Joint Genome Institute web site \[[@B33]\]. The JGI site and the EnsEMBL web site \[[@B34]\] are the two main portals to the *Fugu rubripes*genome annotation. Although EnsEMBL and JGI annotations and genome browsers are different, they share the same gene and transcript structure predictions created by the EnsEMBL pipeline. Several analyses of the draft *Fugu*genome sequence targeting different protein families have been published recently \[[@B35]-[@B39]\], which showed its usefulness for evolutionary and functional studies as well as gene discovery. Here we present an analysis of the presence of the CTLD superfamily in the draft assembly of the *Fugu rubripes*genome. Results ======= Comparison of assembly versions 2 and 3 --------------------------------------- At the time this study was started, annotation of the v.3 assembly was not yet published; hence, most of our analysis was done with v.2 of the assembly and later mapped to the v.3 assembly. From our experience, there is no substantial difference between v.2 and v.3 assemblies in the amount of sequence information and its quality, although the v.3 assembly contains longer scaffolds due to more extensive linkage. Despite very high similarity at the sequence level, the v.3 assembly annotation contains no history information that would provide links between contigs, genes, and transcripts in the second and the third versions of the assembly. None of the stable identifiers for genes, transcripts or peptides from v.2 are present in v.3. This information cannot be generated by usual procedures used in EnsEMBL (e.g. ID Mapping Application, which is a part of EnsEMBL Java APIs) and has to be obtained by sequence comparisons. This lack of correspondence creates difficulties for the sequence analyzer and end point reader. To facilitate analysis and allow comparison, references to feature identifiers for both of the assemblies are given in Table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### CTLD-encoding genes identified in the Fugu rubripes genome.^a^ ::: ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- **Name** **Description** **v.2 gene ID** **v.3 gene ID** ----------------------------------------------------------------------- ------------------------------------------------------------------------------------------- ---------------------------------- ------------------------------------ **I Hyalectans** AGGRECAN AGGRECAN [ANUFRUG00000000095]{.underline} [ANUFR2G00000000089]{.underline} AGGRECAN-F1 Fugu aggrecan paralogue [ANUFRUG00000000081]{.underline} [ANUFR2G00000000077]{.underline} BREVICAN BREVICAN SINFRUG00000078610 SINFRUG00000151617 BREVICAN-F1 Fugu brevican paralogue SINFRUG00000074933 SINFRUG00000128229,\ SINFRUG00000128230,\ SINFRUG00000128231 NEUROCAN NEUROCAN SINFRUG00000054833 SINFRUG00000150572,\ SINFRUG00000150573,\ SINFRUG00000150574,\ SINFRUG00000150576 NEUROCAN-F1 Fugu Neurocan paralogue [ANUFRUG00000000142]{.underline} [ANUFR2G00000000154]{.underline} VERSICAN VERSICAN [ANUFRUG00000000144]{.underline} [ANUFR2G00000000164]{.underline} VERSICAN-F1 Fugu versican paralogue (fragment containing EGF, CTLD and CCP domains) [ANUFRUG00000000061]{.underline} [ANUFR2G00000000059]{.underline} VERSICAN-F2 Fugu versican paralogue (fragment containing link and Ig domains) [ANUFRUG00000000043]{.underline} [ANUFR2G00000000041]{.underline} **II Dendritic cell receptors, mono-ctld macrophage receptors, ASGR** DC-SIGN-F1 Fugu DC-SIGN paralogue [ANUFRUG00000000029]{.underline} [ANUFR2G00000000027]{.underline} DC-SIGN-F2 Fugu DC-SIGN paralogue [ANUFRUG00000000067]{.underline} [ANUFR2G00000000063]{.underline} DC-SIGN-F3 Fugu DC-SIGN paralogue [ANUFRUG00000000069]{.underline} [ANUFR2G00000000065]{.underline} DC-SIGN-F4 Fugu DC-SIGN paralogue [ANUFRUG00000000071]{.underline} [ANUFR2G00000000067]{.underline} DC-SIGN-F5 Fugu DC-SIGN paralogue [ANUFRUG00000000073]{.underline} [ANUFR2G00000000069]{.underline} DC-SIGN-F6 Fugu DC-SIGN paralogue [ANUFRUG00000000109]{.underline} [ANUFR2G00000000105]{.underline} DC-SIGN-F7 Fugu DC-SIGN paralogue [ANUFRUG00000000085]{.underline} [ANUFR2G00000000123]{.underline} DC-SIGN-F8 Fugu DC-SIGN paralogue [ANUFRUG00000000087]{.underline} [ANUFR2G00000000081]{.underline} DC-SIGNR DCSIGN receptor [ANUFRUG00000000027]{.underline} [ANUFR2G00000000025]{.underline} HML2 Similar to human macrophage lectin SINFRUG00000060881 SINFRUG00000120587 SRCL Scavenger receptor with C-type lectin SINFRUG00000071148 SINFRUG00000134389 SRCL-F1 Putative Fugu paralogue of SRCL SINFRUG00000064389 SINFRUG00000152316 XLCMCL eXtra Large Coiled coil region containing Membrane C-type Lectin [ANUFRUG00000000053]{.underline} [ANUFR2G00000000051]{.underline} **III Collectins** COLEC10 COLEC10 SINFRUG00000077039 SINFRUG00000125405 **MGC3279** **Uncharacterized collectin family member** SINFRUG00000064196 SINFRUG00000147955 **IV Selectins** SELECTIN-E E-Selectin [ANUFRUG00000000001]{.underline} [ANUFR2G00000000001]{.underline} SELECTIN-L L-SELECTIN [ANUFRUG00000000003]{.underline} [ANUFR2G00000000003]{.underline} SELECTIN-P P-SELECTIN [ANUFRUG00000000005]{.underline} [ANUFR2G00000000005]{.underline} **VI Multi-CTLD molecules. Macrophage Mannose Receptor (MMR) family** DEC205 DEC205 [ANUFRUG00000000011]{.underline} [ANUFR2G00000000011]{.underline} Endo180 Endo180 SINFRUG00000058766 SINFRUG00000152106 MManR Macrophage mannose receptor SINFRUG00000071196 SINFRUG00000126868,\ SINFRUG00000134363 MManR-F1 Fugu mannose receptor paralogue (fragment) SINFRUG00000064600 SINFRUG00000152797 MManR-F2 Fugu macrophage mannose receptor paralogue. [ANUFRUG00000000039]{.underline} [ANUFR2G00000000035]{.underline}\ [ANUFR2G00000000037]{.underline} MManR-F3 Fugu paralogue of MMR-family gene SINFRUG00000066378 SINFRUG00000152288 MManR-F4 Fugu paralogue of MMR-family gene (fragment) SINFRUG00000078047 SINFRUG00000152861 MManR-F5 Fugu MMR-family member (fragment) [ANUFRUG00000000091]{.underline} [ANUFR2G00000000085]{.underline} PLA2R Phosopholipase A2 receptor [ANUFRUG00000000009]{.underline} [ANUFR2G00000000009]{.underline} **VIII MT-75, layilin** LAYILIN Layilin [ANUFRUG00000000089]{.underline} [ANUFR2G00000000083]{.underline} LAYILIN-F1 Fugu layilin paralogue [ANUFRUG00000000075]{.underline} [ANUFR2G00000000071]{.underline} MT-75 MT-75 SINFRUG00000084745 SINFRUG00000145404 **IX Tetranectin family** CLECSF1 CLECSF1 SINFRUG00000050048 SINFRUG00000136890 SCGF SCGF [ANUFRUG00000000125]{.underline} [ANUFR2G00000000121]{.underline} TETRANECTIN Tetranectin SINFRUG00000084961 SINFRUG00000144710 TETRANECTIN-F1 Fugu tetranectin paralogue SINFRUG00000083037 SINFRUG00000149544 **X PKD** PKD1 Polycystic kidney disease protein 1 SINFRUG00000033997 PKD1L2 PKD-1 homologue 2 [ANUFRUG00000000121]{.underline} [ANUFR2G00000000117]{.underline} **XI Attractin family** ATTRACTIN Attractin SINFRUG00000071911 SINFRUG00000136030 ATTRACTIN-F1 Fugu paralogue of Attractin SINFRUG00000060472 SINFRUG00000147061 **KIAA0534** **KIAA0534** SINFRUG00000056251 SINFRUG00000121439 **XII Eosinophil major basic protein family** EMBPL Putative Fugu EMBP-like protein [ANUFRUG00000000023]{.underline} [ANUFR2G00000000021]{.underline} **XIII DGCR family** DGCR2 DGCR2 SINFRUG00000082125 SINFRUG00000155593 **XIV Thrombomodulin family** C1qRP C1qRP [ANUFRUG00000000049]{.underline} [ANUFR2G00000000047]{.underline} C1qRP-F1 Putative Fugu C1qRP paralogue (fragment) [ANUFRUG00000000013]{.underline} disappeared **CETM** **Protein containing CTLD, EGF and transmembrane domains** [ANUFRUG00000000057]{.underline} [ANUFR2G00000000055]{.underline} ENDOSIALIN ENDOSIALIN [ANUFRUG00000000117]{.underline} [ANUFR2G00000000113]{.underline} THROMBOMOD Thrombomodulin SINFRUG00000077807 SINFRUG00000153798 **XV Bimlec** **BIMLEC** **Novel C-type lectin from BCG cell wall induced monocyte** [ANUFRUG00000000007]{.underline} [ANUFR2G00000000007]{.underline} **XVI SEEC** **SEEC** **Novel SCP-EGF-EFG-CTLD containing protein.** [ANUFRUG00000000041]{.underline} [ANUFR2G00000000039]{.underline} **XVII CBCP** **CBCP** **Calx-Beta and CTLD containing protein** [ANUFRUG00000000047]{.underline} [ANUFR2G00000000045]{.underline} **AFP Antifreeze protein** AFPL-F1 Antifreeze protein-like [ANUFRUG00000000045]{.underline} [ANUFR2G00000000043]{.underline} AFPL-F2 Antifreeze protein-like [ANUFRUG00000000139]{.underline} disappeared **F1 Fugu dual-CTLD molecules** FDC-F1 Putative Fugu dual-CTLD protein 1 [ANUFRUG00000000025]{.underline} [ANUFR2G00000000023]{.underline} FDC-F2 Putative Fugu dual-CTLD protein 2 [ANUFRUG00000000037]{.underline} [ANUFR2G00000000033]{.underline} FDC-F3 Putative Fugu dual-CTLD protein 3 [ANUFRUG00000000099]{.underline} [ANUFR2G00000000093]{.underline} FDC-F4 Putative Fugu dual-CTLD protein 4 [ANUFRUG00000000103]{.underline} [ANUFR2G00000000097]{.underline},\ [ANUFR2G00000000099]{.underline} FDC-F5 Putative Fugu dual-CTLD protein 5 [ANUFRUG00000000107]{.underline} [ANUFR2G00000000103]{.underline} FDC-F6 Putative Fugu dual-CTLD protein 6 [ANUFRUG00000000123]{.underline} [ANUFR2G00000000119]{.underline} FDC-F7 Putative Fugu dual-CTLD protein 7 [ANUFRUG00000000101]{.underline} [ANUFR2G00000000095]{.underline} FTCP Putative Fugu triple-CTLD protein [ANUFRUG00000000015]{.underline} [ANUFR2G00000000013]{.underline} **L Link domain** BRAL1 Brain link protein-1 SINFRUG00000078615 SINFRUG00000151615 CD44 CD44 [ANUFRUG00000000113]{.underline} [ANUFR2G00000000109]{.underline} CRTL1 Cartilage linking protein 1 SINFRUG00000078961 SINFRUG00000137046 CRTL1-F1 Putative fugu cartilage linking protein paralogue [ANUFRUG00000000059]{.underline} [ANUFR2G00000000057]{.underline} CRTL1-F2 Putative fugu cartilage linking protein paralogue SINFRUG00000074643 SINFRUG00000142167,\ SINFRUG00000142169,\ SINFRUG00000142171 HAPLN3 Hyaluronan and proteoglycan link protein 3 SINFRUG00000052853 SINFRUG00000155413 HAPLN3-F1 Putative Fugu paralogue of HAPLN3 SINFRUG00000079552 SINFRUG00000129575 Lyve-1 Lymphatic vessel endothelial HA receptor-1 [ANUFRUG00000000077]{.underline} [ANUFR2G00000000073]{.underline} STABILIN-1 Stabilin-1 [ANUFRUG00000000079]{.underline} [ANUFR2G00000000075]{.underline} STABILIN-2 Stabilin-2 SINFRUG00000074867 SINFRUG00000146665 TSG-6 TSG-6 SINFRUG00000075173 SINFRUG00000148136 **NLSLH** **NLSLH** **Novel L-SeLectin Homologue** [ANUFRUG00000000055]{.underline} [ANUFR2G00000000053]{.underline} NLSLH-F1 Fugu CTLD containing gene fragment, NLSLH paralogue [ANUFRUG00000000097]{.underline} [ANUFR2G00000000091]{.underline} **U Unclassified** AGGRECOL Putative Fugu CTLD-containing protein equally similar to aggrecan and placenta collectin. [ANUFRUG00000000083]{.underline} [ANUFR2G00000000079]{.underline} ANZG001 Putative Fugu CTLD-containing protein (fragment) [ANUFRUG00000000019]{.underline} [ANUFR2G00000000017]{.underline} ANZG002 Putative Fugu CTLD-containing protein (fragment) [ANUFRUG00000000021]{.underline} [ANUFR2G00000000019]{.underline} ANZG004 Putative Fugu protein with CTLD and FTP domains [ANUFRUG00000000093]{.underline} [ANUFR2G00000000087]{.underline} ANZG005 Putative Fugu CTLD-containing protein (fragment) [ANUFRUG00000000065]{.underline} disappeared ANZG006 Putative Fugu CTLD-containing protein (fragment) [ANUFRUG00000000111]{.underline} [ANUFR2G00000000107]{.underline} ANZG007 Putative Fugu CTLD-containing protein (fragment) [ANUFRUG00000000063]{.underline} [ANUFR2G00000000061]{.underline} ANZG008 Putative Fugu CTLD-containing protein (fragment) [ANUFRUG00000000017]{.underline} [ANUFR2G00000000015]{.underline} ANZG010 Putative Fugu CTLD-containing protein [ANUFRUG00000000051]{.underline} [ANUFR2G00000000049]{.underline} ANZG011 Putative Fugu CTLD-containing protein [ANUFRUG00000000115]{.underline} [ANUFR2G00000000111]{.underline} CFN3 Protein with CTLD and FN3 domains. [ANUFRUG00000000105]{.underline} [ANUFR2G00000000101]{.underline} DEC205-FUSE Large Fugu protein which looks like a DEC205 fused to another CTLD-containing gene [ANUFRUG00000000119]{.underline} [ANUFR2G00000000115]{.underline} FG75645 Fugu CTLD-containing protein fragment SINFRUG00000075645 SINFRUG00000139863 PTP-GMC1 Protein-tyrosine phosphatase expressed by glomerular mesangial cells [ANUFRUG00000000130]{.underline} [ANUFR2G00000000137]{.underline} ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ^a^All *Fugu*CTLDcps identified in this analysis are listed. Columns 3 and 4 contain stable identifiers for gene models in the v.2 and v.3 assembly databases, respectively. Identifiers starting with ANUFRU and ANUFR2 belong to our predictions on the v.2 and v.3 assemblies, respectively, and are underlined. EnsEMBL gene stable identifiers are given if the original predictions were used. Bolded members denote *Fugu*proteins matched with novel human orthologues. ::: Protein database searches ------------------------- Due to almost complete lack of cDNA or EST sequences for *Fugu rubripes*, most of the EnsEMBL gene structure predictions are based on homology with known protein sequences from other organisms, mostly mammals. We expected a significant fraction of CTLDcps to be conserved between fish and human, and, therefore, to be predicted correctly by EnsEMBL in the *Fugu*genome. So our first approach to detecting *Fugu*CTLDcps was to search a sequence database of predicted *Fugu*proteins with a hidden Markov model (HMM) for the CTLD. This search returned 69 significant matches. Some of the identified genes had a description assigned to them, apparently derived from the description of the sequence they were found to be homologous to. These descriptions, however, could not be used as a reliable basis for assigning orthology and paralogy relationships. For example, a sequence, which we later identified as an Endo180 orthologue (SINFRUG00000058766 in v.2 assembly annotation) is described as \"80 KDA SECRETORY PHOSPHOLIPASE A2 RECEPTOR PRECURSOR PLA2\", while another gene, which we designated as an aggrecan orthologue (SINFRUG00000069597 in v.2 annotation) was annotated as \"ADRENOLEUKODYSTROPHY PROTEIN (ALDP)\". Therefore, we reviewed domain architecture and sequence similarity matches for each of the sequences found to verify phylogenetic relationships. Homology detection ------------------ The results of Inparanoid \[[@B40]\] comparison (see Methods) of all human to all *Fugu*CTLDcps were used to initially cluster the set of *Fugu*proteins and detect approximate orthology/paralogy links. Inparanoid has an important advantage over phylogenetic tree reconstruction software, as it does not require a multiple alignment of sequences but creates a distance matrix of the local pairwise alignments. This method assigned putative human orthologues to 25 *Fugu*proteins. Orthology relationships for the other 44 sequences from the set were established by individual analyses. Revision of CTLDcp gene structure predictions --------------------------------------------- While analyzing phylogenetic relationships predicted by Inparanoid, we discovered several systematic and sporadic mistakes in the EnsEMBL gene predictions. The most widespread mistake was a failure to include exons encoding TM domains into gene structure prediction. Consequently, almost all EnsEMBL-predicted *Fugu*CTLDcps were soluble proteins, whereas very few human CTLDcps are. Simple comparison with the GenScan \[[@B41]\] features overlapping the CTLD-encoding genes showed that absence of TM domains is a result of coding sequence (CDS) mis-prediction rather than a fundamental difference in *Fugu*CTLDcps. GenScan predictions, in turn, could not be used as a basis for our analysis because they sometimes contain regions that are absent from human or mouse orthologues, and often merge neighboring genes. Another general problem was observed with proteins that had a previously unknown domain architecture (see below). In such cases individual domains were split into separate gene models. In addition to these systematic problems, there were multiple sporadic ones. For example, our analysis of the *Fugu*genome shows that, similarly to the human and mouse genomes, the selectin cluster is well conserved and contains all three selectin genes in tandem (SELE, SELL, SELP), located on scaffolds 1045 (32046--41921) and 166 (83937--93826) in the v.2 and v.3 *Fugu*genome annotations, respectively. However, the EnsEMBL annotation contains a prediction of two overlapping genes (v.2: SINFRUG00000085188 and SINFRUG00000085187; v.3: SINFRUG00000123102 and SINFRUG00000123101), one of which is located in the intron of the other (Figure [1](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Fugu genome sequence and annotation.***A. Fugu*selectin gene cluster annotation in the EnsEMBL database (v.2 annotation is shown, v.3 annotation is almost identical to v.2). Gene models predicted by us based on comparison with human selectins are shown in the grey box. As shown, the CTLD is encoded by the 5\' exon in fSELP, fSELL and fSELE; the TM segment is encoded by the 3\' exon. EnsEMBL predicted transcripts, GenScan predictions and similarity features are shown on the tracks below. Stable IDs for EnsEMBL transcripts are given. The TMHMM track shows ORFs encoding TransMembrane regions predicted by the TMHMM program (see Methods). *B.*Fragments of group VI genes found on various scaffolds. CTLD numbers indicate sequential number of CTLD in full-length MManR, while numbers for the CTLD in the partial sequences indicate the MManR CTLD sequence they are most closely homologous to. :::  ::: To solve these problems, we had to manually revise the predicted structure for all genes encoding proteins detected by the protein-level searches, and correct them using supporting evidence available in the EnsEMBL database, as well as additional evidence generated by us. The latter included similarity features produced by genome-wide GeneWise and BLAST searches with CTLD profiles and sequences, transmembrane domain predictions, and similarity matches to the complete sequence of supposed human or mouse orthologues. As the final stage of the CTLDcp identification process, we performed a set of DNA-level comparisons to ensure that the CTLD-containing loci that are not covered by EnsEMBL-predicted genes, or for which transcript predictions are wrong and, thus, not detectable by protein database searches, were not omitted from the analysis. This \"quality control\" step led to identification of an additional set of 25 well conserved CTLDcps, which had both new and known domain architectures, as well as additional individual CTLDs, which were merged with neighboring CTLDcp loci if appropriate. Groups of *Fugu*CTLDcps ----------------------- After all these searches, we had identified a set of 94 *Fugu rubripes*loci encoding CTLDcps (Table [1](#T1){ref-type="table"}), which in total contain 173 individual CTLDs, including PTR/Link-type CTLDs \[[@B42]\]. *Fugu*CTLDcps were named according to their human orthologues, established on the basis of domain composition and sequence similarities. Where more than one homologue was present in *Fugu*, a name was produced by adding a suffix of the form \"-FXX\", where XX is a sequential number of the paralogue, to the name of the closest human homologue. Predicted CTLDcps that do not have homologues among the known CTLDcps have identifiers of the form ANZ000. A few of these novel genes were orthologous to loci in other vertebrate genomes supported by expression data, but otherwise are un-characterized, and were assigned descriptive names (CBCP, Bimlec, SEEC, CETM, NLSLH). We have clustered *Fugu*CTLDcps using the classification scheme for human CTLDcps based on domain composition; this comprises 14 groups \[[@B15]\]. Link/PTR-domain-containing CTLDcps, apart from hyalectans, were placed into a separate group. Among the *Fugu*CTLDcps that did not have mammalian homologues we detected a distinct group of soluble dual-CTLD sequences, which we have called F1 (Figure [2](#F2){ref-type="fig"}). The remainder of the *Fugu*-specific CTLDcps were assigned to the U (Unclassified) group. Gene structure prediction for members of the U group is the lowest in quality, due to lack of supporting evidence apart from similarity to CTLD sequence profiles and GenScan predictions. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **CTLDcps with novel domain architectures.***Fugu*CTLD-containing proteins, which do not fit into the existing CTLDcp classification are shown. Domain abbreviations are explained in the text. Roman numbers near names indicate suggested new group names for the new *Fugu*sequences, which also have new predicted human homologues. C-terminal CTLDs of DEC205-FUSE that are not present in the v.3 assembly are shown in light pink. :::  ::: All but two groups of human CTLDcps have detectable representatives in the *Fugu*genome (Figure [3](#F3){ref-type="fig"}, Table [1](#T1){ref-type="table"}). We did not detect any orthologues for groups V (NK cell receptors) and VII (lithostathine/Reg family). The member repertoire for most of the other groups is very well conserved between *Fugu*and human. However, groups II and III, which include some of the best-studied mammalian CTLDcps, have a significantly different member composition in *Fugu*. In summary: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Phylogenetic relationships between fish and human CTLDs.**A phylogenetic tree built on a ClustalW alignment of a 95% non-redundant collection of predicted *Fugu*CTLDs and known human and fish CTLDs. Link domains and group VI CTLDs were excluded from the alignment. The tree was built by the neighbor-joining method with 100 bootstrap trials using the ClustalW program. PhyloDraw was used to draw the radial cladogram shown. Branches containing CTLDs from CTLDcps belonging to the same group are shaded; group numbers are marked. Lower case prefixes in the identifiers indicate taxonomic origin: h -- *Homo sapiens*, f -- *Fugu rubripes*, zbrfs -- *Danio rerio*(zebrafish), g -- *Gillichthys mirabilis*, gldhs -- *Carassius auratus*(goldfish), carp -- *Cyprinus carpio*(common carp), rsmlt -- *Osmerus mordax*(rainbow smelt), slmn -- *Salmo salar*(Atlantic salmon), wfldr -- *Pseudopleuronectes americanus*(winter flounder), ahrng -- *Clupea harengus*(Atlantic herring), servn -- *Hemitripterus americanus*(sea raven), jpeel -- *Anguilla japonica*(Japanese eel), medak -- *Oryzias latipes*(Japanese medaka), c -- *Paralabidochromis chilotes*(cichlid fish). :::  ::: ### Group I All four members of the lectican group that are present in human have orthologues in the *Fugu*genome. Each of the *Fugu*hyalectan genes is duplicated. One of the *Fugu*versican copies is split between two scaffolds in the v.2 assembly. ### Group II We found only one representative of the asialoglycoprotein receptor (ASGR) family in *Fugu*(HML2), while in human this family has 3 members encoded by a gene cluster on Ch 17 (ASGR1, ASGR2, HML2). The *Fugu*sequence was identified as an HML2 orthologue by phylogenetic analysis based on the alignment of CTLD sequences. Another clearly identifiable member of group II is the orthologue of scavenger cell receptor C-type lectin (SRCL), which is duplicated in *Fugu*and is 50% identical to the human SRCL. The rest of the group II *Fugu*CTLDcps (DC-SIGN-F1 -- DC-SIGN-F8, XLCMCL) do not have clearly identifiable orthologues among known human CTLDcps, although phylogenetic analyses based on CTLD sequence alignment indicate that they are homologous to members of the group II subgroup containing DC-SIGN, Mincle and Dectin-2, which also appear as top hits in BLAST searches. However, this subset of group II *Fugu*sequences co-clusters in phylogenetic trees and is not similar enough to any tetrapod sequence to establish orthology. Four of the sequences (DC-SIGN-F2, DC-SIGN-F3, DC-SIGN-F4, DC-SIGN-F5) are located in a cluster on scaffold 75 in the v.3 assembly. Two members of the subgroup (DC-SIGN-F1 and DC-SIGN-F6) have unstable placements in phylogenetic trees, and may appear on a branch containing human/mouse group V sequences, if the latter are included in the alignment. This association is, however, unstable and may be due to mistakes in CDS prediction or phylogeny reconstruction. Alternatively, it is possible that these sequences are homologous to the common predecessor of group V and group II CTLDcps. ### Group III Although *Fugu*has two collectins, there are no orthologues for mannose binding proteins (MBPs) or pulmonary surfactant proteins (PSP), which are the best studied members of the group in human. Both of the *Fugu*collectins (COLEC10, MGC3279) are well conserved compared with their human orthologues and co-cluster with them in phylogenetic trees. No functional information is available for the novel collectin MGC3279, which was discovered in a large-scale cDNA sequencing project and maps to chromosome 2p25.3 in the v.31 NCBI assembly of the human genome, but the exceptionally high level of conservation between human and fish (\~76% identity) strongly suggests that it is functional and important in both organisms. COLEC10 (collectin liver 1, CL-L1) was originally reported as limited to birds and mammals \[[@B43]\] based on the Zoo-blot analysis. ### Group IV As already mentioned, all three selectin genes found in other vertebrates are present in *Fugu*and have the same genome arrangement. ### Group VI We identified *Fugu*orthologues for all four human group VI members: macrophage mannose receptor (MManR), DEC-205 (CD205), phospholipase A2 receptor (PLA2R) and Endo180. In addition, there are 5 sequences (MManR-F1 -- MManR-F5) showing high similarity to members of the group, four of which do not contain the minimal number of CTLDs (8) present in the known group VI sequences (Figure [1](#F1){ref-type="fig"}). The fragments belong to at least 3 group VI CTLDcps. Although the most parsimonious explanation of the presence of these fragments would be that each of the genes encoding an eight-CTLD molecule (MManR, Endo180 and PLA2R) was copied in a chromosome or genome duplication event, phylogenetic analysis indicates that all five sequences are paralogues of the MManR gene, which, thus, appears to have been duplicated several times. There is one more potential group VI member in *Fugu*. A GenScan-predicted DEC-205-FUSE gene, which was assigned to the U group, encodes a large protein (\~2000 residues) with multiple CTLDs clustered in two groups: 5 at the N terminus and 10 (7 in v.3 assembly) at the C terminus, with an LCCL domain \[named after its presence in Limulus factor C, cochlear protein Coch-5b2, and late gestation lung protein Lgl1; \[[@B44]\]\] and a coagulation factor 5/8 C-terminal domain (discoidin domain, FA58C) lying in the middle separating the two groups of CTLDs (see Figure [2](#F2){ref-type="fig"}). EnsEMBL predictions in the DEC-205-FUSE locus in both versions of the assembly contain a large (4 kb) intron in the region encoding LCCL, FA58C and 8 CTLDs at the center of the molecule. LCCL has been observed in a combination with a CTLD in an invertebrate protein \[[@B45]\], while FA58C has been found only in combination with LCCL, but not with a CTLD \[[@B46]\]. Although there is no supporting cDNA or EST evidence for our predicted gene structure, the small intron sizes (e.g. LCCL is separated by 135 bp from the downstream CTLD) and well-conserved CTLDs, suggest that the prediction may be correct if the corresponding region was correctly assembled. There is no orthologue for DEC-205-FUSE in the human genome. ### Groups VIII and IX We have identified *Fugu*orthologues for all known human members of groups VIII and IX. One member in each of these groups is duplicated in *Fugu*(Layilin and Tetranectin). ### Group X In addition to the PDK1 orthologue, which was identified previously \[[@B26]\], there is at least one more putative group X member, orthologous to a recently identified human and mouse PKD1 homologue PKD1L2 \[[@B47]\]. It is interesting to note that the GenScan-predicted *Fugu*PKD1L2 sequence is very similar to the sequences of human and mouse PKD1L2 cDNAs, even though the latter were deposited in GenBank at the beginning of June 2003 -- after GenScan prediction. This example indicates that *ab initio*GenScan predictions on the *Fugu*genome can be very accurate. ### Group XII We found a single sequence resembling mammalian eosinophil major basic proteins (EMBPs) in *Fugu*(EMBPL). Although the similarity between the mammalian and the fish sequences is very low (\~30% identity), several observations suggest that the *Fugu*EMBP-like sequence is an orthologue of one of the two mammalian genes. First, the overall domain architecture of the fish protein is similar to that of the EMBPs. Although the fish CTLD has a neutral pI (7.1), it is preceded by a 30-residue peptide with a predicted pI of 3.62, analogous to the longer acidic neck of the mammalian EMBPs. In the existing classification \[[@B15]\], the presence of the acidic neck is used as the defining feature of group XII distinguishing it from the other group of single-CTLD soluble proteins (VII). Second, in the phylogenetic trees EMBPL usually appears on the same branch as EMBPs (e.g. Figure [3](#F3){ref-type="fig"}), albeit with low bootstrap support. Third, the exon-intron structure of the CTLD region is identical in fish and mammalian genes. Finally, the fish sequence has the same rare substitution in the fourth position of the WIGL motif as the EMBP sequences (discussed in more detail below). ### Group XIV The thrombomodulin family is fully represented in *Fugu*, with one gene duplicated (C1qRP). In addition, a novel member of the family conserved between *Fugu*and mammals was identified, which we named CETM (for CTLD, EGF, TransMembrane domain) (see Figure [2](#F2){ref-type="fig"}). Multiple full-length cDNA and EST sequences from different tissues found in nucleotide databases indicate that mammalian CETM is ubiquitously expressed. The sequence of the CETM CTLD contains a putative carbohydrate-binding motif (EPN), which is normally associated with mannose specificity. ### Antifreeze-protein-like sequences We identified two putative CTLDcp-encoding loci with similarity to antifreeze proteins: AFPL-F1 and AFPL-F2 (antifreeze-protein-like), almost identical to each other and positioned in tandem on scaffold 1930 in the v.2 assembly. In v.3 of the assembly, the AFPL-encoding region was rearranged and one of the AFPL loci disappeared. The intron-exon structure of the CTLD-encoding region is identical to the structure of the sea raven antifreeze protein gene \[[@B48]\] with three intron insertions (upstream of C1, downstream of the WIGL motif, and between C2 and C3 \[[@B49]\]), and very similar to the structure of the *Salmo salar*serum lectins \[[@B19]\], where only the first two splice sites are present. The Fugu AFPL gene expression is confirmed by an EST sequence BU806418, which covers the whole predicted CDS. ### Link domain containing CTLDcps All link domain-containing proteins identified in mammals are represented in *Fugu*and often are highly conserved between fish and human (e.g. TSG-6, 72%; Stabilin-1, 45% identity); we will consider them as a single group despite their different domain architectures. Predicted members of the CD44 family (CD44 and lymphatic vessel endothelium-specific hyaluronan receptor (Lyve-1)), however, are much more divergent from their human homologues, and it is not clear whether the two loci found in *Fugu*are orthologues of the two human genes or paralogues which arose by duplication of an ancestral gene. In a recently published comprehensive study of another family of the Link group, the hyaluronan and proteoglycan binding link proteins (HAPLN), four homologues were identified in vertebrates (mouse, human and partially zebrafish) each linked to one of the four lecticans \[[@B50]\]. As all lecticans (i.e. group I) are duplicated in *Fugu*, we were expecting to also find duplicate copies of all HAPLN members. However, orthologues of only three HAPLNs were found (CRTL1, BRAL1, HAPLN3), two of which are linked to hyalectans in the same way as in mammalian genomes (CRTL1 with Versican, BRAL1 with Brevican). The state of the assemblies does not allow to determine conclusively whether HAPLN3 is linked to Aggrecan or not. Only two of the *Fugu*lectican gene duplications are accompanied by corresponding HAPLN genes: Aggrecan-F1 is linked to HAPLN3-F1 and the CRTL1 paralogue is present downstream to Versican-F1 in two tandem copies (CRTL1-F1 and CRTL1-F2). In neither version of the assembly could the HAPLN4 homologue be identified downstream to Neurocan or Neurocan-F1. Sequence conservation levels within the HAPLN proteins compared with their human orthologues is quite high (e.g. 76% identity for CRTL1). ### Fugu dual-CTLD CTLDcps The members of this group are soluble proteins with two or three CTLDs, which we initially characterized as fragments of putative macrophage mannose receptor paralogues. However, phylogenetic analysis showed that these proteins constitute a separate group, with no mammalian orthologues detectable in sequenced genome and protein databases. The domain structure prediction is confirmed by three zebrafish cDNAs (CAE17649, CAE17650, CAE17651), which have the same domain organization, although conservation between zebrafish and *Fugu*sequences is only moderate (\~30%). Another homologue with the same domain structure and similarity to the F1 group members, which was returned as the top-scoring hit by BLAST searches in the nrdb, is the SCARF2 protein from a planarian *Girardia tigrina*\[[@B51]\]. A hypothetical dual-CTLD protein from *Drosophila*(NP\_609962), which presumably corresponds to the single member of group B in the *Drosophila*CTLDcp classification of Dodd and Drickamer \[[@B14]\], was also detected as a F1 homologue by BLAST. Novel CTLDcps conserved between *Fugu*and mammals ------------------------------------------------- Discovering novel superfamily members in existing database sequences is one of the most important and exciting outcomes of a systematic computer-based study such as this. We predicted putative *Fugu*orthologues for several uncharacterized mammalian CTLDcps (Bimlec, MGC3279, KIAA0534, CETM, SEEC, CBCP, NLSLH) that are well conserved between *Fugu*and mammals. Most of the predictions were supported by mammalian cDNA sequences from public databases, but for two of them (NLSLH and CBCP) no full-length cDNA from any organism was found in DBs. The high level of genomic sequence conservation over evolutionary time from fish to human, as in the case of NLSLH, and the presence of partial cDNA and EST sequences from rodents and human, as in the case of CBCP, were strongly suggestive that the predictions are correct. The novel CTLDcps that could be attributed to one of the 14 known groups have been discussed in the preceding sections for the corresponding groups; those that do not fit into the existing classification are described below. A large (\~2100 aa) proteoglycan (CBCP), containing a set of chondroitin sulphate proteoglycan (CSPG) repeats \[[@B52]\], which are homologous to the NG2 ectodomain \[[@B53]\], a calcium-binding Calx-β domain \[[@B54]\] and a CTLD, is a novel member of a protein family which had not been reported previously to have members containing CTLDs; examples of this family also include the human MCSP/CSPG4 \[[@B55]\] and mouse FRAS1 \[[@B56]\] genes. The prediction was supported by three overlapping but incomplete cDNA sequences from human and mouse, high levels of conservation between human and *Fugu*(\~50% identity), and the compact structure of the predicted *Fugu*gene. CBCP has been placed in a new CTLD group, XVII; its domain structure is shown in Figure [2](#F2){ref-type="fig"}. We have cloned a full-length cDNA of mouse CBCP confirming the domain structure predicted in this study (A.N. Zelensky, in preparation). The CTLD of CBCP lacks Ca-binding residues, and its long loop region is short, resembling that of the group V CTLDs. Another protein with a novel domain organization, whose prediction is strongly supported by available cDNAs, is SEEC ([S]{.underline}CP, [E]{.underline}GF, [E]{.underline}GF, [C]{.underline}TLD-containing protein) (see Figure [2](#F2){ref-type="fig"}), which is well conserved between human and *Fugu*. Although not described in a publication, a full-length human SEEC cDNA (AK074773) was sequenced in the NEDO high-throughput sequencing project. The predicted *Fugu*SEEC is 63% identical to the human sequence. The sperm-coating glycoprotein (SCP) domain, which is present in a broad set of organisms from yeast and plants to mammals, but whose function is unknown \[[@B57]\], is rarely observed in combination with other domains in proteins; in only one other known protein (from sea urchin) is it found together with an EGF domain \[[@B58]\], and SEEC is the first example of a CTLD-SCP combination. The potential Ca/carbohydrate-binding motif (QPD) characteristic of galactose specificity is present in the CTLD. SEEC has been placed in a new CTLD group XVI. A predicted protein named \"novel L-selectin homologue\" (NLSLH) because its CTLD is most similar to selectin CTLDs is duplicated in *Fugu*(NLSLH and NLSLH-F1) but only moderately conserved (32% identity) between *Fugu*and human. The putative human orthologue is located on Ch1q25.1 about 18 Mb further from the centromere than the selectin cluster and is supported only by EST sequences (AA912157, AA889574), but not cDNAs. No conserved domains except for the CTLD could be detected in the human and *Fugu*NLSLH loci so, if the predictions are correct, NLSLH is a soluble single-CTLD-containing protein. Carbohydrate-binding motifs are not present in the NLSLH and NLSLH-F1 CTLDs. Finally, a type I transmembrane protein Bimlec, whose prediction is supported by a full-length human cDNA, was placed in a new group XV. Dating the CTLDcp duplications ------------------------------ We found 12 groups of unlinked *Fugu*-specific CTLDcp paralogues (Table [1](#T1){ref-type="table"}), and attempted to determine the duplication dates using two approaches: (1) based on the estimation of the number of synonymous nucleotide substitutions (Ks) in the coding sequences and (2) based on the molecular clock hypothesis. For all but two pairs of duplicated genes, Ks values estimated with four different methods (see Methods) were between 1.5--2.5, which indicates a complete saturation of the synonymous sites (Figure [4(A)](#F4){ref-type="fig"}). Ks values so high cannot provide an accurate estimation of the duplication age, but we can conclude with confidence that the CTLDcp gene duplications are at least 150 Myr old, which is the time required for complete saturation of silent sites assuming a mutation rate of 2.5 substitutions/silent site/billion years in fish \[[@B59]\]. If, however, Ks values presented in Figure [4(A)](#F4){ref-type="fig"} and the silent mutation rate are close to correct, the corresponding duplication timeframe is predicted to be 300--500 Myr. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **CTLDcp duplication dates.**A. Average number of synonymous substitutions per synonymous site (Ks) for CTLDcp paralogue pairs based on full-sequence (triangle) and CTLD-only (diamond) alignments, measured with four different methods (see Methods). Error bars show one standard deviation in the CTLD-only measurements. All possible pairwise alignments between the MManR fragments and between the three CRTL1 paralogues were analyzed. Only homologous regions were used for MManR fragment alignments. B. A linearized phylogenetic tree built by the neighbor-joining method from Poisson-corrected distances between ClustalW-aligned sequences of CTLDs 3--5 from *Fugu*, mouse and human MManRs. Sequence of the human PLA2R region containing CTLDs 3--5 was used as an outgroup. T~hm~-- time of separation between human and mouse \[96 Myr; 60\], T~fish~-- time of separation between ray-finned and lobe-finned fishes \[430 Myr; 32\]. Time of duplication (T~dupl~) was calculated using average between molecular clock calibrated with T~hm~and with T~fish~. :::  ::: In order to date the duplications based on molecular clock measurements, we aligned duplicated *Fugu*CTLD sequences with their vertebrate orthologues present in GenBank, and built linearized phylogenetic trees based on the alignments. As human and mouse sequences were invariably available, the divergence time between these two species \[96 Myr; 60\] was used to calibrate the clock, together with the divergence time between Actinopterygii and Sarcopterygii \[430 Myr; 32\]. Symmetrical tree topology ((H, M) (F, F1)), expected for a Actinopterygian-specific duplication, was revealed by at least one phylogeny reconstruction method we used for the following six homologue groups (data not shown): brevican, neurocan, MManR, SRCL, tetranectin and HAPLN3, with duplications dated 369, 284, 397, 377, 360 and 312 Myr, respectively. A typical tree with symmetrical topology is shown for MManR in Figure [4(B)](#F4){ref-type="fig"}. The other six alignments (aggrecan, versican, layilin, attractin, C1qRP, CRTL1) produced trees with topologies suggesting a duplication predating the split between Actinopterygian and Sarcopterygian. The portion of symmetrical topologies (50%) in the CTLD set is similar to the ratio reported by Taylor and coworkers in fish: 15 of 27 (55 %) \[[@B61]\], and 25 of 53 (45%) \[[@B62]\] for bigger and more heterogeneous gene collections. Discussion ========== Draft assembly limitations -------------------------- A systematic study based on draft-quality whole-genome data for an organism like *Fugu rubripes*has some limitations, as the genomic sequence is incomplete, fragmented and sometimes misassembled, and the expressed sequence information is scarce. On the other hand, many of the genomes that are currently being sequenced will be released and remain for sometime in the same state as the *Fugu*genome data are now. Indeed, more than a year after the initial release \[[@B31]\] very few improvements to the *Fugu*genomic data \[v.3 assembly and EST sequencing project; \[[@B63]\]\] have been published. Therefore, it is essential to extract useful biological information from draft-quality whole-genome sequences. Our study is such an attempt. We have mentioned four limitations of the draft-state assembly -- incompleteness, fragmentation, misassembly and lack of expression information. While the last might appear the biggest problem, we found that *ab initio*predictions combined with manual curation and interspecies comparison have proven to be very accurate (e.g. see PKD1L example), thanks to the compactness of the *Fugu*genome, smaller ratio between intron and intergenic region sizes compared with mammalian genes, wealth of data for comparative analysis etc. We do not expect that sequencing the remaining 5% of the *Fugu*genome, which is mostly heterochromatic regions, will lead to discovery of many new CTLDcps. From the comparison of the *Fugu*CTLDcp repertoire discovered by us and found in other fish species independently, the only surprising omission in our results is a MBP orthologue. MBP sequences have been found in several other fish species. Their absence in Fugu may represent a bona fide gene loss. As to the fragmentation, only a few of the CTLDcps are split between scaffolds, namely versican and some MManR paralogues (Figure [1](#F1){ref-type="fig"}). All of the fragmented genes are big, and in most cases the fragments can be combined easily to reveal the full sequence. Finally, misassembly signs were observed in several CTLDcp loci while comparing two versions of the assembly. These showed as presence of repeated regions in the v.2 assembly, which disappeared in the v.3 assembly. Two groups identified in higher vertebrates are not detectable in *Fugu* ------------------------------------------------------------------------ We could not detect CTLDcp representatives for groups V (NK cell receptors) and VII (lithostathine) in the *Fugu*genome. CTLDs in the members of these groups have lost their carbohydrate-binding activities, and perform functions that have, apparently, evolved after evolutionary separation of tetrapods from fish, or which are mediated by other proteins in fish. For example, group VII members are secreted into the digestive tract -- a system that is very flexible evolutionally. Group V is probably one the youngest and most rapidly evolving sets of CTLDcps; its component members vary significantly even between rodents and human, a phenomenon connected to the co-evolution with the acquired immune system proteins that group V CTLDcps interact with. Our conclusion on the absence of group V CTLDcps in the *Fugu*genome is at odds with the conclusions of two studies describing group V CTLDcp evolution in chordates. A recent paper describes possible CD94 homologues (cichlid killer cell lectin receptor, cKLR) in bony fishes *Paralabidochromis chilotes*and *Oreochromis niloticus*, which are encoded by a large multi-gene family with at least 10 members \[[@B28]\]. Another recent work described sequencing of a CD94 homologue in a tunicate \[[@B64]\]. The decision by Sato et al. \[[@B28]\] to assign putative fish killer cell receptors to group V rather than to group II was based on several considerations, including gene structure, absence of canonical Ca^2+^/carbohydrate-binding residues, and phylogenetic analysis based on the CTLD alignment. The latter consideration is mentioned as the most important one. However, as the authors themselves note, bootstrap values for placing cKLR on the group V branch, are \"low to moderate\". Indeed, we found that in phylogenetic trees built using different methods (maximum parsimony, distance estimation method with PAM matrix followed by neighbor-joining tree reconstruction, maximum likelihood) from the ClustalW alignments of cKLR sequences with group V and group II CTLD sequences from *Fugu*, mouse and human, cKLR placement is unstable. As shown in Figure [5](#F5){ref-type="fig"}, on a tree built by the neighbor-joining method we found cKLR on the branch containing the *Fugu*-specific subset of group II CTLDcps (DC-SIGN-F1 -- DC-SIGN-F8), most of which do contain residues required for Ca^2+^/carbohydrate binding. On a tree built by the maximum parsimony method, we found cKLR on a separate branch equally related to group II and group V sequences (not shown). Also, a BLAST search with the complete cKLR sequence (GI 31789959) in the non-redundant NCBI protein database returns members of the ASGR subgroup of group II as top matches. Therefore, we judge that sufficient support for assignment of cKLR to group V is lacking and the question of the presence of the NK-cell receptor family in fishes is still open. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Relationships between fish, mouse and human group V and II CTLDs.**Non-redundant set of CTLD sequences from known human and mouse CTLDcps classified as groups II and V, *Fugu*CTLDcps classified as group II, and putative killer cell receptor from *Paralabidochromis chilotes*(cKLR) were aligned with ClustalW. A consensus phylogenetic tree was built from 100 bootstrap trials using the *protdist*(with PAM distance matrix) and *neighbor*programs from the PHYLIP package. Black triangle shows position of cKLR. Bootstrap values higher than 40 are indicated. :::  ::: As to the putative CD94 homologue from tunicates, it is indeed more similar to CD94 than to any other CTLDcp. However, the low level of sequence homology and the lack of evidence for existence of group V CTLDcps in more advanced taxa does not allow a confident statement that the sequence from tunicates is a CD94 orthologue, rather than a result of convergent evolution. Expansion of the innate immunity CTLDcp groups in *Fugu* -------------------------------------------------------- Unlike pairwise unlinked duplications (see below), tandem duplications and other gene family expansions are limited to two groups, namely the DC-SIGN subgroup of group II and MManR. In mammals, members of these subgroups play an important role in innate immune responses. In particular, DC-SIGN is actively studied due to its ability to bind and internalize a broad range of bacterial and viral pathogens, including HIV-1 and *Mycobacterium tuberculosis*(reviewed in \[[@B65]\]), while MManR is also implicated in binding and phagocytosis of a wide range of microorganisms \[[@B66]\]. Expansion of these groups, most notably the DC-SIGN subgroup, in *Fugu*may reflect a larger role for innate immunity in host defense in lower vertebrates. Interestingly, multi-copy clusters comprising at least 10 genes encoding close cKLR homologues were identified in another cichlid fish species *Oreochromis niloticus*\[[@B28]\], which suggests another parallel between the expanded DC-SIGN subgroup in puffer fish and cKLRs of cichlids. There are no extra members, however, in the *Fugu*collectin group -- another CTLD group directly involved in innate immunity in mammals. Moreover, the mannose binding protein (MBP), which is the best-studied mammalian collectin involved in lectin complement activation pathway, was not detected by us. The absence of MBP orthologues in *Fugu*is rather puzzling, as MBP sequences have been found in several other fish species (*Danio rerio*, *Cyprinus carpio*and *Carassus auratus*; \[[@B20]\]), and are well conserved within the *Cyprinidae*carp family. The collectin family is also present and expanded in the Urochordate *Ciona intestinalis*with nine collectin genes identified in the draft genome sequence \[[@B67]\], although it is not clear whether one of these nine genes is an MBP orthologue. Given the role of MBPs in complement activation in mammals, and their presence and level of conservation in the carp family, it is possible that the *Fugu*MBP orthologue does exist but is not covered by the draft genome sequence. Complement-activating C-type lectins from lower organisms have been identified but not completely sequenced \[[@B68]\]; they have multiple CTLDs as in CPL-III from the protochordate *Clavelina picta*\[[@B69]\] or lack the collagen domain and show more similarity to other CTLDcps such as the glucose-binding lectin (GBL) from another tunicate, *Halocynthia roretzi*\[[@B70]\]. *Fugu*dual CTLD molecules -- a missing link between vertebrate and invertebrate CTLDs? -------------------------------------------------------------------------------------- Previous whole-genome studies of the CTLD superfamily in two invertebrates \[[@B13],[@B14]\] failed to identify any groups of CTLDcps common to both invertebrates and vertebrates. A group of predicted dual CTLD-containing proteins in *Fugu*(F1) may be the first vertebrate group that has detectable homologues in invertebrates. Alternatively, it is possible that none of the *Fugu*F1 group members are in fact orthologous to the invertebrate sequences, as sequence similarities are only moderate (\~30% ID) and the domain architecture is simple and could have evolved independently in different lineages. However, several observations suggest that at least F1 members from *Fugu*and zebrafish and SCARF proteins from *Girardia tigrina*evolved from the same predecessor. First, similarity levels between fish sequences and between fish and planarian sequences are about the same. It is unlikely that the fish sequences are unrelated, which implies that F1 members are evolving quickly, and only major structural features of these molecules are under selective pressure \[[@B49]\]. Second, the CTLDs of planarian and, in all cases at least one CTLD of the fish dual-CTLDcps, contain residues characteristic of Ca/carbohydrate binding. In vertebrates, ability to bind carbohydrates is associated with the oldest CTLDcps groups, and is considered to be an ancestral feature of the CTLD. Indeed, both vertebrate CTLDcp groups that we failed to find in *Fugu*(V and VII) have lost sugar-binding properties. This is also the case for the antifreeze proteins from fish and snake venom CTLDcps, which have only been found in the corresponding clades. Third, similar domain organization (two CTLDs, no transmembrane domain) is also observed in two other known groups of invertebrate CTLDcps: immulectins from various insect species \[[@B71],[@B72]\] and nine proteins from *C. elegans*, classified as group D1 by Drickamer and Dodd \[[@B13]\]. Despite identical domain structure, none of these proteins shows statistically significant homology to the fish F1 group members or their putative homologues from planarian or *Drosophila*. Altogether, this indicates that domain structure alone cannot establish an evolutionary link between the fish and invertebrate sequences. Hence, the suggestive link between the F1 group fish members and the planarian and *Drosophila* proteins is even more interesting. CTLDcp classification update ---------------------------- The existing classification of CTLDcps is generally accepted and popularly used in studies of the superfamily and recently has been updated \[[@B15]\]. The classification divides CTLDcps into monophyletic groups of proteins with identical overall domain architecture based on a combination of structural and phylogenetic information. Although two previous large-scale studies \[[@B13],[@B14]\] showed it to be inapplicable for description of invertebrate CTLDcps, our analysis of the puffer fish genome indicates that it is sufficient to describe the superfamily in all vertebrates, with only minor modifications and some extensions. Our newly discovered CTLDcps, with a few exceptions, do not fit into the existing classification because of their unique domain architecture. We propose several new groups to accommodate the novel CTLDcps which have been found in both higher and lower vertebrates and are supported by cDNA sequences: • XV -- Bimlec (type I transmembrane protein), which in phylogenetic trees is not placed on the same branch as group VIII sequences, has a distinct exon-intron structure of the CTLD region and a neck not similar to the neck region of the group VIII sequences; • XVI -- SEEC, based on unique domain architecture; • XVII -- CBCP, based on unique domain architecture; Additional groups may be required for the sequences not supported by sufficient expression data (NLSLH) and other sequences from the \"unclassified\" group whose presence in higher vertebrates is not clear. Also, clade-specific groups, such as fish antifreeze proteins (AFP), dual-CTLD sequences (group F1) predicted by us and so far identified only in fish, or snake venom CTLDcps which lack orthologues in other vertebrates, are required. It has been suggested previously \[[@B19],[@B48]\] that AFPs belong to group VII based on their domain architecture and exon-intron structure. However, our phylogenetic analysis of an alignment of CTLD sequences from all known human and mouse CTLDcps and 26 different fish CTLD-containing protein sequences identified by searching the NCBI protein database with BLAST, indicates that they constitute a phylogenetically distinct group including all known soluble fish CTLD-containing proteins, except Cyprinidae collectins. As to the exon-intron structure, introns in the group XII (EMBP) CTLDs are at exactly the same positions as in group VII and AFP-like CTLDs, which suggests that all three groups are closely related but does not allow classification of the fish AFP-like sequences to either of the mammalian groups. Interestingly, just like most of the AFPs, mammalian EMBPs contain an atypical WIGL motif with a glycine in the fourth position, a substitution not observed in any other mammalian CTLD we analyzed. Taken together, these observations indicate that in a broader evolutionary perspective the differences between some of the groups including CTLDcps with a very similar domain architecture (VII, XII and AFP; II and V) become less distinct, which makes classification of the \"intermediate\" or \"ancestral\" sequences, equally related to more than one group, problematic. Selective duplication of the *Fugu*CTLDcp-encoding genes and the whole-genome duplication hypothesis ---------------------------------------------------------------------------------------------------- The hypothesis that whole-genome duplications were one of the main driving forces in vertebrate evolution, providing genetic material for increased diversity and progressive development \[[@B73]\], and that there were two rounds of whole-genome duplication in vertebrate phylogeny (the 2R hypothesis) \[[@B73],[@B74]\], is actively debated \[[@B75],[@B76]\]. A more recent whole-genome duplication is suggested for the Actinopterygian branch \[[@B61]\]. Ray-finned fish are the most diverse group of vertebrates, and based on the initial observation that each of the four human HOX gene clusters has two homologues in zebrafish \[[@B77]\] it was suggested that they have undergone an additional round of a whole-genome duplication after the divergence from Sarcopterygian about 430 Myr ago \[[@B61]\]. Analysis of the genome duplication in fish can give a picture of a duplicated genome after 300--400 Myr of evolution and fill the gap between the now generally accepted recent tetraploidizations in plants \[[@B78]\] and yeast \[[@B79]\] and the alleged more ancient duplication(s) of the ancestral vertebrate genome. Although many fish genes are indeed duplicated \[[@B61],[@B77],[@B80]-[@B82]\], it is not clear whether the copies were created by a complete genome duplication (autopolyploidy), merge of different genomes (allopolyploidy), regional duplication, or simply a series of tandem duplications. Attempts to show that ancient tetraploidization (has not) occurred usually involve: (i) searching for an excess of paralogue groups where the number of members is double the number of alleged duplications (i.e. 2 in case of Actinopterygian duplication, and 4 in case of vertebrate duplication, the \"one to four rule\") \[[@B74],[@B76]\]; (ii) showing that a statistically significant number of duplications took place at approximately the same time by molecular clock estimation or synonymous substitution counting \[[@B83],[@B84]\]; (iii) using phylogenetic methods to assess the relation between duplication and speciation events \[[@B61]\]; and (iv) showing that duplicated genes are arranged in paralogous blocks on chromosomes (paralogons) \[[@B62],[@B85],[@B86]\]. We used these approaches to analyze the nature of the observed CTLDcp duplications in *Fugu*. Our results clearly show that tandem gene copying is a mechanism of CTLD family evolution and led to generation of three gene clusters: DC-SIGN-F2 -- DC-SIGN-F5 (4 genes), CRTL1-F1 and CRTL1-F2, and AFPL-F1 and AFPL-F2. Members of the two latter clusters are nearly identical and may be an assembly artifact. Twelve other duplicated genes are not linked in the current assembly and have sequences much more diverged than tandem duplicates. Of the 12 genes only MManR, which has 3 paralogues, is present in more than two copies. We consider this is important evidence in favor of a whole-genome duplication, as sporadic duplications cannot explain such a strong bias towards two-member paralogue groups. Unfortunately, the results of duplication time estimations are less conclusive as they give only a broad timeframe for the possible duplication events of about 300--400 Myr. As in the case of some other fish gene families reported previously \[[@B61],[@B62],[@B87],[@B88]\], molecular phylogeny reconstruction performed by us often indicates that duplications occurred before the divergence between fish and tetrapods. However, this could be an artifact of the method caused by different selection pressures on duplicates. Unfortunately, there is practically no overlap between vertebrate and invertebrate CTLD families, so we could not use invertebrate sequences to refine phylogenetic analysis. To conclude: phylogenetic relationships between CTLD paralogues and estimated duplication time distribution indicate that there was a burst in duplication activity in the *Fugu*genome 300--400 Myr ago. While we cannot determine definitively the nature of the duplications (tandem, regional or whole-genome), a pronounced bias in the number of two-member paralogue groups strongly suggests that there was a single large-scale or whole-genome duplication event in fish. Another interesting observation is that CTLDcp genes were either duplicated, or retained after a large-scale duplication, in a pronounced selective manner. One group (I) is duplicated completely, while in other groups only partial duplications are found. Interestingly, group I (lecticans), which in tetrapods contain four large (\>2000 amino acids) proteins, very similar to each other in sequence and domain structure, is a good candidate for demonstrating the 2 R hypothesis. If the four lecticans arose as a result of the alleged two rounds of the whole-genome duplication early in vertebrate history, the fact that the family was also completely duplicated in fish and retained after the duplication appears very non-random and implies some functional explanation. In the human genome, all four genes encoding lecticans are located on different chromosomes (1, 5, 15 and 19), but it is not clear whether they are linked in *Fugu*. Another group that conforms to the 2 R hypothesis is group VI, which in tetrapods has four members with almost identical domain structure in mammals (Pla2R, MManR, DEC-205 and Endo180). Though in the *Fugu*genome we identified 7 group VI sequences, some of which are fragmented (Figure [1](#F1){ref-type="fig"}), phylogenetic analysis shows that only one member of the family (MManR) was quadruplicated, while others are present in a single copy. Both molecular clock and Ks-based methods date the MManR duplications at approximately the same time as other CTLDcp gene duplications. Phylogenetic trees, built on alignment of the overlapping portions (Figure [1](#F1){ref-type="fig"}) of the complete sequences and three largest fragments (fMManR-F1, fMManR-F2, fMManR-F3) have symmetrical structure, with fMManR-F1, fMManR-F2 and fMManR-F3 forming a separate branch (Figure [4(B)](#F4){ref-type="fig"}). A whole-genome duplication, generating fMManR and fMManR-F1, followed by tandem duplications of fMManR-F1, producing fMManR-F2 and fMManR-F3, can explain this topology. Conclusions =========== We have performed an analysis of the CTLD superfamily composition in *Fugu rubripes*. Although the sequence assembly is in the draft state and lacks physical mapping information and native cDNA sequences that could be used to make and verify gene predictions, the quality of the data is good enough despite these limitations to answer many important questions. Our study demonstrates that all but two groups of CTLDcps present in mammals are also found in fish, that most of the groups have the same composition as in mammals, and that the missing groups are the evolutionarily most dynamic ones involved in physiological processes that may be specific to higher vertebrates. We also identified at least one distinct fish-specific CTLD group, which could be the first known vertebrate CTLD group also found in invertebrates. The compactness of the *Fugu*genome makes it an extremely convenient reference sequence for identification of new genes based on supporting similarity features, and we were able to identify and predict the structure of several new CTLD-containing genes highly conserved between *Fugu*and human. The new sequences are supported by cDNA and EST sequences from databases and have previously unknown domain architectures. We are now characterizing some of these sequences experimentally. We also show that CTLDcp-encoding genes are selectively duplicated in *Fugu*, in a manner that suggests an ancient large-scale duplication event in fish. Methods ======= Corrected gene predictions are made available as a distributed annotation system (DAS) \[[@B89]\] resource \[[@B90]\], which can be viewed in the EnsEMBL genome browser. The data source names for predictions based on assemblies v.2 and v.3 are fugu\_ctld\_1 and fugu\_ctld\_2, respectively. Transcript sequences (in FASTA format) for the CTLDcp-encoding genes created or modified by us (stable IDs starting with ANU) and their translations are also provided in the [additional file 1](#S1){ref-type="supplementary-material"} and [additional file 2](#S2){ref-type="supplementary-material"}, respectively. Searches and gene annotations were done on version 2 of the *Fugu rubripes*genome assembly \[[@B31]\] downloaded from the EnsEMBL web site \[[@B91],[@B92]\]. When the third version of the assembly was released, we mapped gene annotations onto it. Mapping was done on the basis of SSAHA \[[@B93]\] matches in the v.3 assembly for exons predicted on the v.2 assembly. The v.2 assembly is currently accessible at the Singapore IMCB site \[[@B94]\] and on our server \[[@B95]\], which is pre-configured to display the DAS track with our annotations and contains a reference table with hyperlinks for all of the *Fugu*CTLDcp genes discussed. Version 3 of the assembly can be found on the main EnsEMBL web site \[[@B34]\]. The EnsEMBL genome browser can be easily configured to display our gene models as a DAS track. We used a multi-step approach to find genes encoding CTLDs. First, a hidden Markov model (HMM) profile of the CTLD was used to scan a FASTA database of EnsEMBL-predicted genes with the hmmsearch program from the HMMER package \[[@B96]\]. To detect orthologues and paralogues, the set of *Fugu*sequences found was compared with the 95% non-redundant set of sequences of human CTLDcps that could be found in the Entrez proteins database, using the Inparanoid program \[[@B40]\]. All of the 25 orthology links detected by Inparanoid were checked manually. Because of systematic and sporadic errors in EnsEMBL gene predictions, we had to manually revise the structure of each of the 69 genes encoding proteins detected by the HMM-based search. This was done using the Apollo genome annotation software \[[@B97]\] connected to a local installation of the EnsEMBL database. To facilitate annotation, several additional feature tracks were added to the EnsEMBL database: a\) Similarity features detected by GeneWise \[[@B98]\] search of *Fugu*scaffold sequences with a CTLD HMM built in a global alignment mode. This was done to detect well conserved CTLDs while avoiding many false positives. b\) Same as a), but with an HMM built in the local alignment mode; this was done to detect highly conserved fragmented CTLDs; c\) Similarity features detected by a TBLASTN search of *Fugu*scaffold sequences using all known human CTLD sequences; this was done to detect CTLDs that are less conserved; d\) ORFs encoding putative transmembrane (TM) domains. To create this track a database of all possible ORFs longer than 45 bp was produced and translated into protein sequence using the EMBOSS programs. This was then scanned with the TMHMM program \[[@B99]\] to detect ORFs that encode putative TM domains. To verify whether there are CTLDs that were not covered by EnsEMBL gene predictions, we searched for all significant CTLD similarity features detected by GeneWise which do not overlap with any of the genes analyzed in the first stage. This step led to detection of 25 new CTLD-coding genes, including most of the ones that have previously uncharacterized domain organization. At the next stage we analyzed the loci with different CTLD similarity features detected by genewisedb search with a local alignment HMM. Finally, the features identified by BLAST and not overlapping with already detected genes were analyzed. This set of features mostly contained only partial CTLD matches. We translated both the new and already predicted gene CDSs into protein sequences and performed another Inparanoid comparison. Phylogenetic relationships were analyzed with the programs from the Phylip package \[[@B100]\]. ClustalW \[[@B101]\] guiding trees were used for quick phylogeny estimation and in cases where a proper multiple alignment could not be made. BioPerl \[[@B102]\] and EnsEMBL Perl modules were used to automate all stages of the analysis. Domain architectures were analyzed with the SMART web service \[[@B103]\]. To estimate the proportion of substitutions in synonymous sites, we aligned translated sequences of the duplicated CTLDcp-encoding genes with ClustalW, using either whole sequence or sequence for the CTLD-encoding region only, and built nucleotide sequence alignments based on the protein alignments. Ks estimations were performed with four methods: Lynch and Connery \[[@B104]\] and Li \[[@B105]\], both implemented in the ntdiffs package \[[@B104]\]; and Nei and Gojobori \[[@B106]\] and Yang and Nielsen \[[@B107]\], both implemented in the yn00 program from the PAML package. Duplication dating using the calibrated molecular clock approach was performed as in \[[@B83]\]. Alignments of CTLD-containing regions of *Fugu*paralogues and their mammalian orthologues were made with ClustalW. The MEGA2 program \[[@B108]\] was used to build linearized trees from Poisson-corrected distances, p-distances and Gamma-corrected distances by the neighbor-joining method with 1000 bootstrap samplings. The global clock was calibrated using divergence times 96 Myr and 430 Myr for human-mouse and fish-mammal splits, respectively \[[@B32],[@B60],[@B83]\]. List of abbreviations ===================== AFP, antifreeze protein; AFPL, AFP-like; CBCP, [C]{.underline}alx-[β]{.underline} and [C]{.underline}TLD-containing [P]{.underline}rotein; CDS, coding sequence; CETM, [C]{.underline}TLD, [E]{.underline}GF, [T]{.underline}rans[M]{.underline}embrane domain CTLD; C-type-lectin-like domain; CTLDcp; CTLD-containing protein; DAS, distributed annotation system; EMBP, eosinophil major basic protein; EST, expressed sequence tag; HMM, hidden Markov model; MBP, mannose-binding protein; Myr, million years; NLSLH, [N]{.underline}ovel [L]{.underline}-[S]{.underline}eLectin [H]{.underline}omologue; PKD1, polycystic kidney disease protein 1; SEEC, [S]{.underline}CP, [E]{.underline}GF, [E]{.underline}GF, [C]{.underline}TLD; TM, transmembrane. Authors\' contributions ======================= ANZ carried out the bioinformatics studies and participated in the interpretation of its results. JEG conceived the study and participated in the interpretation of its results. Both authors participated in writing the manuscript and approved its final form. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 **Transcript sequences for re-annotated *Fugu*CTLD genes.**The file contains cDNA sequences (in FastA format) for all CTLDcp-encoding genes that were re-annotated by us (sequence identifiers starting with ANU). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 **Protein sequences for re-annotated *Fugu*CTLD genes.**The file contains protein product sequences (in FastA format) for all CTLDcp-encoding genes that were re-annotated by us (sequence identifiers starting with ANU). ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ JEG is supported by the ANU IAS block grant, and ANZ is supported by a PhD scholarship from the ANU. We would like to thank Elia Stupka and the *Fugu*bioinformatics team at the Singapore Institute of Molecular and Cell Biology for help and allowing us access to the pre-release versions of the v.3 assembly annotation. We also thank Anders Krogh for providing us with a copy of the TMHMM program for local use.

PubMed Central

2024-06-05T03:55:47.270076

2004-8-1

PMC514893

Background ========== CXC and CC chemokines are small soluble proteins expressed and secreted by a number of cell types during the initial host response to injury, allergens, antigens, or invading microorganisms \[[@B1]\]. These ligands selectively attract leukocytes to inflammatory foci via facilitation of cellular adhesion, transendothelial migration, chemotaxis and cellular activation. Receptors for chemokines are members of the large family of G-protein receptors that signal via heterotrimeric guanine nucleotide-binding proteins of the Gαi-subclass \[[@B2]\]. Chemokine receptors can be subdivided into specific families based on their specificity for C, CC, CXC, or CX3C chemokine ligands. Three distinct types of receptor binding are currently recognized: (1) chemokine receptors that bind only one chemokine specific ligand; (2) chemokine receptors that bind more than one chemokine often with different binding affinities; and (3) promiscuous chemokine receptors that bind to numerous chemokines \[[@B2]\]. The chemokine receptor CXCR4 binds to the CXC chemokine, CXCL12 and functions as a co-receptor for HIV-1 \[[@B3]\]. CXCR4 is broadly expressed by many cells within the body including cells of the immune and the central nervous system \[[@B4]-[@B7]\]. This receptor mediates the migration of resting leukocytes and hematopoietic progenitors in response to its specific ligand \[[@B8],[@B9]\]. CXCL12-induced chemotaxis is inhibited by pertussis toxin, enhanced *in vitro*by IL-3, and selectively inhibited by soluble ephrin-B receptor. \[[@B10]\]. In addition, proinflammatory stimuli such as lipopolysaccharide, tumor necrosis factor (TNF-α) or interleukin-1 potentiates lymphocyte-and monocyte-, but not neutrophil-mediated CXCL12 responses \[[@B11],[@B12]\]. Furthermore, CXCL12 is an extremely potent *in vitro*and *in vivo*chemoattractant for mononuclear cells and lymphocytes \[[@B13]\]. CXCL12 is expressed in the cells forming Hassall\'s corpuscles and plays a significant role in the elimination of apoptotic thymocytes in normal and HIV-1-infected thymic tissues \[[@B14]\]. In addition to the bone marrow, quantitative PCR analysis has detected expression of CXCL12 in the lymph nodes, lung, and liver \[[@B15]\]. Autocrine and paracrine production of CXCL12 by peripheral blood CD34^+^CD38^+^cells also appears to trigger their transition from G~0~to G~1~and, in conjunction with thrombopoietin, enhances their survival through signal transduction mediated by the PI3K/AKT proteins \[[@B16]\]. Together these data support a role for CXCL12 as a critical factor for cellular growth and differentiation, cellular trafficking, myelopoiesis, and organ vascularization \[[@B17],[@B18]\]. In contrast to CXCL12, considerably less is known about the chemokine CXCL10. CXCR3 (GPR9; CD183), the receptor for CXCL10 also binds the CXC chemokines CXCL9 and CXCL11 \[[@B19]\]. Recent studies of the CNS have suggested that CXCR3 additionally binds CCL21 \[[@B20]\]. CXCL10 is secreted by a variety of cell types, including monocytes, endothelial cells, fibroblasts, and astrocytes. CXCL10 is also a chemoattractant for human monocytes, natural killer and T cells (preferentially Th1 cells), and appears to modulate adhesion molecule expression and function \[[@B21]-[@B23]\]. CXCL10 is expressed in keratinocytes, lymphocytes, monocytes, and endothelial cells during Th1-type inflammatory diseases such as psoriasis and atopic dermatitis, but only at very low basal levels in normal keratinocytes \[[@B24],[@B25]\]. CXCL10 inhibits bone marrow colony formation by CD34^+^cells in the presence of stem cell growth factor (SCGF), colony stimulating factor 2 (granulocyte-macrophage) (CSF2; GM-CSF), or a combination of SCGF and erythropoietin (EPO). Moreover, CXCL10 has antitumor activity *in vivo*and is a potent inhibitor of angiogenesis \[[@B26]\]. This antitumor activity appears to be mediated by the ability of CXCL10 to recruit lymphocytes, neutrophils, and monocytes into inflammatory infiltrates. Moreover, CXCL10 has also been recently shown to be a Ras target gene and is overexpressed by a number of colorectal cancers \[[@B27]\]. Overall, CXCL10 is an important chemokine for mediating delayed-type hypersensitivity responses and a potent regulator of colony formation, angiogenesis, adhesion and cell migration. Alterations in gene expression are important determinants of cellular physiology. As a consequence, the identification, cloning and characterization of differentially expressed genes can provide relevant and important insights into a variety of biological processes. To investigate and compare the similar and distinct genes induced by the chemokines, CXCL12 and CXCL10, in normal physiology, we utilized differential display analysis to identify mRNAs in a Jurkat T cell line expressing endogenous CXCR4 and transfected with human CXCR3 gene. We have identified and cloned several differentially expressed genes displaying both elevated and diminished expression in the context of specific chemokine receptor ligation. The possible relevance of such differential responses within normal immune responses and in normal T-cell physiology is discussed. Methods ======= Differential display -------------------- Total RNA was isolated using the Qiagen RNeasy^®^kit (Qiagen Inc., Valencia, CA) and treated with DNase I (GenHunter, Nashville, TN). Two micrograms of the total RNA was derived from subclone of CXCR3-transfected Jurkat T cells (generously donated by Dr. Thomas Hamilton, Lerner Research Institute, Cleveland, OH) cultured for 24 h in the presence and absence of 1 μg/ml of bioactive CXCL12 or CXCL10 (PeproTech). It should be noted that the CXCR3-transfected Jurkat T cells were subcloned from the original cultures. Subclones of the CXCR3-transfected lines were initially generated at the initiation of these studies so that homogenous CXCR3-bearing cells were available. A single Jurkat subclone was selected and examined for coexpression of both CXCR3 and CXCR4 by flow cytometry (Table [1](#T1){ref-type="table"}). The isolated RNAs were subsequently reverse-transcribed with 400 units of MMLV reverse transcriptase (GenHunter) in three separate reactions each containing 2 uM of a one-base-anchored H-T~11~M (i.e. H-T~11~G, H-T~11~A and H-T~11~C) primer (RNAimage^®^, GenHunter) and 20 uM dNTP for 60 min at 37°C. After heat inactivation of the reverse transcriptase at 75°C for 5 min, 2 μl of each reverse transcription reaction was added to 18 μl of a PCR master mix containing 2 uM of an H-T~11~-arbitrary primer, 1 U Taq polymerase (Qiagen), 2 uM dNTP, and a-\[^33^P\]dATP. Each primer pair was denatured at 94°C for 30 sec, annealed at 40°C for 2 minutes and extended at 72°C for 30 sec for 40 cycles with a final extension for 10 minutes. \[^33^P\]-labeled PCR products were resolved on a 6% denaturing polyacrylamide gel. The autoradiogram was inspected on a light box and differentially expressed bands marked by needle punches. The punched film was carefully oriented on the dried gel and the marked bands excised with a scalpel blade. Glycogen (10 mg/ml), 3 M sodium acetate, and 85% EtOH were added and after overnight storage at -80°C, the DNA was precipitated by centrifugation. Each DNA was subsequently reamplified using the same PCR primer set and conditions except that the dNTP concentration was increased to 20 uM and no isotope was added to the mixture. Reamplified PCR products were resolved on a 1.5% AmpliSize™ (BioRad, Richmond, CA) agarose gel, stained with SYBR^®^Gold (Molecular Probes, Eugene, OR) and extracted from the gel using a QIAEX II kit (Qiagen). Successfully amplified bands were cloned using the PCR-TRAP^®^cloning vector system (GenHunter) and ligated into GH-competent cells. Following transformation, only clones that contain an insert are capable of growing on LB-Tet agarose plates. The cloned insert was subsequently checked by colony-PCR using primers flanking the PCR-TRAP^®^vector and the insert sequenced to identify genes differentially expressed between control and chemokine-treated CXCR3-transfected Jurkat T cells. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Flow cytometric analysis of transfected Jurkat T cell lines ::: \% Positive (MFI) ------------------- ---------- ---------- --------- ------- Jurkat-Neo 98 (138) 99 (118) 2 (5) 4 (4) Jurkat-CXCR3 98 (154) 99 (124) 99 (24) 5 (8) Jurkat-Bcl2 96 (144) 96 (124) 5 (6) 7 (5) CXCR3-, Bcl2-and Neo-transfected Jurkat T cells were stained using IgG FITC-labeled antibodies to CD3, CXCR4, CXCR3, and CXCR2 and analyzed for cell surface protein expression via flow cytometric analysis. The data are expressed as % positive (mean fluorescence intensity). It should be noted that the CXCR3-transfected Jurkat cells expressed low levels of CXCR3 on their cells surface suggesting a lower receptor density in comparison to CXCR4. ::: Quantitative analysis of PCR fragments -------------------------------------- One microgram of DNase I-treated total RNA from control, CXCL12-or CXCL10-treated CXCR3-transfected Jurkat T cells was reverse-transcribed with 200 units of SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA) in a 20 ul reaction for 50 min at 42°C followed by heat-denaturation at 70°C for 15 minutes. Two microliters of the first strand reaction product were amplified in six duplicate reactions using standard PCR conditions (94°C for 1 minute, 63°C for 1 minute and, 72°C for 1 minute for 34 cycles) with sequence specific primers. Individual tubes were removed after 24, 26, 28, 30, 32, and 34 cycles and the concentration of specific PCR product determined using an Agilent 2100 BioAnalyzer and DNA 1000 LabChip (Agilent Technologies, Palo Alto, CA). Each RT-PCR was performed twice for each RNA preparation. Two housekeeping control genes, GAPDH and ribosomal protein L32 (RPL32), were utilized as controls with specific primers producing a 599 bp product for GAPDH and a 233 bp product for RPL32. Western blot analysis --------------------- Cells were lysed in modified RIPA cell lysis buffer (50 mM Tris-HCl, pH 7.4, 1% NP-40, 1% sodium deoxycholate, 0.15 M NaCl, and 1 mM EDTA) with 1 mM phenylmethylsulfonylfluoride (PMSF), 1 mM sodium orthovanadate, 5 μg/ml leupeptin, 2 μg/ml aprotonin and one Complete Protease Inhibitor Cocktail tablet (Roche Diagnostics Corporation, Indianapolis, IN) per 50 ml of buffer. Whole cell protein extract was used directly for Western blot analysis. Protein concentration was determined using Bio-Rad protein assay kit (BioRad). Twenty micrograms of total protein from each sample was separated on a 10% Tris-glycine polyacrylamide gel and transferred onto a polyvinylidene difluoride membrane (Invitrogen, San Diego, CA). Membranes were blocked for 1 h at room temperature in PBS containing 5% non-fat dry milk and 0.1% Tween-20. The membranes were then incubated overnight at 4°C in primary antibody (anti-CA150, anti-thioredoxin, anti-flotillin and anti-ferritin H chain, BD Biosciences Transduction Laboratories, Lexington, KY and Santa Cruz Biotechnologies, Inc, Santa Cruz, CA) diluted 1:1000 in PBS containing 5% non-fat dry milk and 0.1% Tween-20. The membranes were washed in PBS with 0.1% Tween-20 then incubated for one hour in secondary antibody (goat anti-mouse-HRP and rabbit anti-goat-HRP; Santa Cruz Biotechnology, Inc., Santa Cruz, CA). The blots were washed and the proteins detected using the ECL Plus Western Blotting Kit (Amersham Biosciences UK Limited, Buckinghamshire, UK) and X-MAT AR Film (Eastman Kodak, Rochester, NY). Cellular migrations and intracellular calcium mobilization ---------------------------------------------------------- Jurkat T cell migration was examined using a fluorescence-based Transwell chemotaxis assays as previously described \[[@B21],[@B61]\]. CXCR3-and neo-transfected Jurkat T cells were labeled with 10 μg/ml Hoechst 33342 (Molecular Probes) in cRPMI for 30 min at 37°C, and then treated with chol-BCD as described above. The cells were then resuspended in RPMI with 1% FBS to a concentration of 1 × 10^7^/ml. RPMI (0.6 ml) containing 1% FBS with or without 100 ng/ml SDF-1α was added to the bottom wells of the 24-well plate. Transwell chambers with 5 μm pore filters (Corning CoStar, Acton, MA) were then placed into the wells. Cells (1--3 × 10^6^in 100 μl) were then added to the chambers. After 2 h, the migrated cells in the bottom wells were transferred to triplicate wells of a 96-well plate in 150 μl volumes. Hoechst fluorescence was measured on a Fluoroskan Ascent FL fluorescence plate reader (Thermo Labsystems, Franklin, MA) at λ~ex~= 355 nm, and λ~em~= 460 nm. Results are expressed as migration index calculated by subtracting the fluorescence intensity of media alone and comparing the values to the fluorescence intensity (relative number) of cells migrated into the bottom chamber in media alone, which is normalized to a value of 1. Fluorescence values were within the linear range of a standard dilution curve. CXCR3-transfected Jurkat T cells were loaded with the fluorescent indicator, Fura-2AM (Molecular Probes, Eugene, OR), for 30 min, then washed and resuspended in PBS containing calcium and magnesium at 10^6^cells/ml \[[@B61]\]. The ratio of free to bound intracellular calcium was determined by spectrofluorometry by monitoring absorption at 340 nM versus 380 nM and emission at 510 nM. The chemokines, CXCL12α and CXCL10 (Peprotech, Rocky Hill, NJ) were utilized in these experiments at 1 μg/ml. Results ======= CXCR3-transfected Jurkat T cells migrate and mobilize intracellular calcium in response to CXCL12 and CXCL10 ------------------------------------------------------------------------------------------------------------ The cells were found to be functionally responsive to both CXCL12 and CXCL10. As shown in Fig. [1A](#F1){ref-type="fig"}, a subclone of a CXCR3-transfected Jurkat T cell line, which was found by flow cytometry to coexpress both CXCR3 and CXCR4 (Table [1](#T1){ref-type="table"}), specifically migrated in response to CXCL12 and CXCL10 in a dose-dependent fashion. Optimal migration for CXCL12 was noted at 0.5--1 μg/ml, while optimal migration for CXCL10 was observed at 1 μg/ml. In contrast, control neomycin phosphotransferase gene-transfected Jurkat T cells only demonstrated responses to CXCL12. Similarly, CXCR3-transfected Jurkat T cells demonstrated a potent calcium mobilization in response to CXCL12 (1 μg/ml) and a modest response to CXCL10 (1 μg/ml), while neo-Jurkat T cells failed to demonstrate any CXCL10 response (Fig. [1B](#F1){ref-type="fig"}). The modest migration and calcium mobilization observed in response to CXCL10 compared to CXCL12 in this cell line suggests either distinct signaling through CXCR3 or lower cell surface CXCR3 density. Based on the flow cytometric data (Table [1](#T1){ref-type="table"}), the mean fluorescence intensity for CXCR4 is approximately 5-fold greater than that for CXCR3 on this transfected cell line. Similar levels of CXCR4 were expressed on all of these cell lines including a neo control or Bcl2-transfected cell Jurkat line. As expected, CXCR2 failed to demonstrate any staining on these cell lines. While there may be differences in CXCR4 and CXCR3 signaling in these cells, differences in receptor density may influence the chemokine-induced gene expression differences described below and thus cannot be ruled out. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **CXCR3-transfected Jurkat T cells migrate and mobilize calcium in response to CXCL10 and CXCL12.**CXCR3-transfected Jurkat T cells or neo-transfected Jurkat T cells (Panel A) were examined within Transwell chemotaxis chamber for their ability to migrate in response to various concentrations of CXCL10 and CXCL12 as described in the *Methods*. The migration data are expressed as a migration index relative to the number of migrating cells in the absence of chemokine. Panel B shows the mobilization of intracellular calcium within CXCR3-transfected and control Jurkat T cells stimulated with CXCL10 or CXCL12 (1 μg/ml). The data points were collected every 0.48 s and are presented as the relative ratio of fluorescence excited at 340 and 380 nm. Arrows indicate when the chemokine was added to the chambers. The insert within Panel B is a close-up view of the CXCL10 response within CXCR3-transfected Jurkat T cells. We have never observed any calcium mobilization or chemotactic activity by non-transfected or neo-transfected Jurkat T cells in response to CXCL10 (data not shown). :::  ::: Differential display of mRNA expression in CXCL12-and CXCL10-treated T cells ---------------------------------------------------------------------------- In an effort to identify genes, which may be upregulated or downregulated by CXCL12 or CXCL10, we examined Jurkat T cells that that expressed endogenous CXCR4 and that had been transfected with CXCR3 using DDRT-PCR analysis (Figure [2](#F2){ref-type="fig"}, Table [1](#T1){ref-type="table"}). Jurkat T cells were stimulated with either CXCL12 or CXCL10 at a concentration of 1 μg/ml for 24 hr. The dose of 1 μg/ml was selected as this concentration yielded optimal migration for both CXCL12 and CXCL10 in the CXCR3-transfected T cells. Total RNA prepared from normal and chemokine-stimulated Jurkat cells were reverse transcribed into cDNA. The resultant cDNAs were amplified with 45 combinations of the arbitrary and oligo(dT) anchored primers. Seventeen cDNA bands were found to be differentially expressed in CXCR3-transfected Jurkat T cells. Fig. [3](#F3){ref-type="fig"} shows the representative differential display results obtained with six separate primer combinations. Two cDNA fragments, designated as C31.3 (ribosomal protein S25) and A6.7 (thioredoxin) were identified to be differentially expressed by chemokine treated but not in untreated Jurkat cells (Fig. [3A](#F3){ref-type="fig"} and [3B](#F3){ref-type="fig"}). The cDNA fragments were excised from the gel, reamplified, subcloned and sequenced. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Flow chart of the RT-PCR-based differential display procedure.**A detailed description of each step is found in the *Methods.* :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Differential display autoradiography.**Panel A and B show examples of DDRT-PCR autoradiographs. Chemokine treatments are indicated above and the primer combinations below each cell lane. Panel C shows densitometric measurements of the relative gene expression of individual bands after treatment of the transfected Jurkat T cells with 1 μg/ml of CXCL12 or CXCL10 for 24 hrs. :::  ::: Direct sequencing of DDRT-PCR products -------------------------------------- To characterize the sequence identity of DDRT-PCR products, the excised and re-amplified bands were cloned into the PCR-TRAP vector and transformed into GH-competent E. coli. Several colonies were selected to be cultured, plasmids were purified, and the inserts wee subsequently screened by colony-PCR using primers flanking the site of the PCR-TRAP vector. The plasmids containing an insert were sequenced. We successfully sequenced a total of 146 cDNAs. Seventeen mRNAs were differentially expressed post treatment with either CXCL12 or CXCL10, each representing known genes with established functions. Five additional RNAs were also identified defining known genes with unknown functions and nine identified hypothetical or predicted genes of unknown function (Table [2](#T2){ref-type="table"}). Twenty-one genes were upregulated in CXCR3-transfected Jurkat cells following both CXCL12 and CXCL10, four genes displayed a discordant response and seven genes were down regulated by both chemokines (Fig [3](#F3){ref-type="fig"}). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Differentially expressed genes post chemokine treatment. ::: BAND Bp GENE LOCUSLINK GENEACC NAME -------- ----- -------------- ----------- ------------ ------------------------------------------------------ C10.5 AD24 64318 NM\_022451 AD24 protein A09.4 BM-002 51569 NM\_016617 BM-002 hypothetical protein A09.2 DDX1 1653 NM\_004939 DEAD/H (Asp-Glu-Ala-Asp/His) box polypeptide 1 C27.3 DDX30 22907 NM\_014966 DEAD/H (Asp-Glu-Ala-Asp/His) box polypeptide 30 G04.3 DFKZp564M113 none AL049282 Homo sapiens, clone MGC:5564, mRNA A10.3 DFKZp566D193 25847 AL050051 DFKZp566D193 protein A25.1 EIF4B 1975 NM\_001417 eukaryotic translation initiation factor 4B A28.3 EPB41L2 2037 NM\_001431 erythrocyte membrane protein band 4.1-like 2 C10.1 FLJ12876 64767 NM\_022754 hypothetical protein FLJ12876 G01.1 101 FLOT1 10211 NM\_005803 flotillin 1 G09.2 444 FTH1 2495 NM\_002032 ferritin, heavy polypeptide 2 C07.3 329 GHITM 27069 NM\_014394 growth hormone inducible transmembrane protein C10.4 IFITM2 10581 NM\_006435 interferon induced transmembrane protein 2 (1--8 D) A01.2 INVS 27130 NM\_014425 inversin C03.3 KIAA0478 9923 NM\_014870 KIAA0478 gene product G04.6 KIAA0648 23244 AB014548 KIAA0648 gene product G01.4 KIAA1600 57700 AB046820 KIAA1600 protein A04.7 401 LOC51053 51053 NM\_015895 geminin C35.1 LOC51633 51633 NM\_016023 CGI-77 protein G35.1 MAP3K10 4294 NM\_002446 mitogen-activated protein kinase kinase kinase 10 A27.14 MGC10744 84314 NM\_032354 hypothetical protein MGC10744 C15.3 MGC4809 91860 AF308287 serologically defined breast cancer antigen NY-BR-20 A25.3 NCBP1 4686 NM\_002486 nuclear cap binding protein subunit 1,80 kD A27.12 NCBP2 22916 NM\_007362 nuclear cap binding protein subunit 2, 20 kD G11.1 RPL7 6129 NM\_000971 ribosomal protein L7 G30.2 RPS12 6206 NM\_001016 ribosomal protein S12 C31.3 RPS25 6230 NM\_001028 ribosomal protein S25 A11.1 TCERG1 10915 NM\_006706 transcription elongation regulator 1 (CA150) A06.7 157 TXN 7295 NM\_003329 thioredoxin C15.5 VIT1 55519 NM\_018693 vitiligo-associated protein VIT-1 G03.6 WBP11 51729 NM\_016312 WW domain binding protein 11 ::: Sequence analysis indicated that the 401 bp A4.7 cDNA is highly homologous to the gene LOC51053 that is also known as geminin, a cell cycle regulator. The 157 bp A6.7 cDNA fragment is 94% identical to the 501 bp human thioredoxin (TXN) gene. The sequence of the band A11.1 was homologous to the transcription elongation regulator 1 gene (TCERG1) also known as CA150. The 329 bp C7.3 cDNA fragment is 87% identical to the published human gene encoding the growth hormone inducible transmembrane protein (GHITM). The sequence of the band C10.4 was homologous to the interferon-induced transmembrane protein 2 (IFITM2) gene, a member of the 1--8 gene family whose members are strongly induced by both type I (IFNα, IFNβ) and type II (IFNγ interferons). The 102 bp G1.1 cDNA fragment is 97% identical to the human gene encoding flotillin 1 (FLOT1) that is thought to play a role in vesicular trafficking and signal transduction. Sequence analysis indicated that the 444 bp G9.2 cDNA, as well as several other bands, were highly homologous (\>95%) to the gene encoding the iron-storage protein ferritin heavy polypeptide 1 (FTH1). The sequences of the other cDNAs included in Table [2](#T2){ref-type="table"} showed significant homology to published sequences in Genebank. However, it should be noted that in a number of cases, the nucleotide sequence of the cDNA matched named genes about which little is known or matched, in some cases for over 500 bp, the sequence of hypothetical or predicted genes. Several ribosomal proteins including L7, S12 and S25 and ferritin heavy chain (FTH1) were identified several times in our DDRT analysis. In Figure [4](#F4){ref-type="fig"}, these changes in gene expression as assessed by DDRT are more clearly displayed post cluster analysis using an arbitrary densitometric scale of 0--4 for all named genes found by DDRT-PCR. Red indicates upregulation and green down-regulation in response to chemokine treatment. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Comparison of the relative gene expression by CXCL12-versus CXCL10-treated T cells.**Changes in gene expression were displayed using computer programs (Cluster, Tree-View-Eisen) using an arbitrary densitometric scale of 0--4 for all named genes found by DDRT-PCR. Red indicates upregulation, green down-regulation, and black-to-gray means no change. The columns are labeled as C (Vehicle control), S (CXCL12), and I (CXCL10). :::  ::: Identification of differentially expressed RNA transcripts associated with CXCL12 or CXCL10 treatment ----------------------------------------------------------------------------------------------------- As a large fraction of DDRT-PCR products have been shown to yield weak or barely detectable signals in the Northern blot analysis, we sought to confirm the identity and differential expression of these bands via RT-PCR analysis. We focused our attention on a subgroup of the seventeen DDRT-PCR products representing known genes with established functions. Gene specific primers were designed to produce unique amplimers between approximately 150 and 500 base pairs in size. RT-PCR analysis was performed on the Agilent BioAnalyzer 2100 System. The advantage of using this system to examine competitive PCR products lies in the accurate absolute and relative quantitation of each amplified product. Small differences in the amount of amplimer product, which cannot be detected using slab gel analysis, are more easily analyzed on this equipment permitting RT-PCR to be used to measure changes in gene expression. RNA from CXCR3 receptor-transfected T cells, both before and post treatment with CXCL12 or CXCL10 was reverse transcribed in bulk and aliquots RT product amplified by PCR using specific sets of primers. Initially, a 599 bp GAPDH amplimer was utilized as a housekeeping gene. However, it was noted that GAPDH significantly up-regulated post treatment of the cells with either CXCL12 or CXCL10. To address this issue, a 233 bp ribosomal protein L32 amplimer was utilized as a housekeeping gene for the subsequent comparison of gene expression levels. In Figures [5](#F5){ref-type="fig"} (CXCL12) and [6](#F6){ref-type="fig"} (CXCL10), each of the seven genes (FLOT1, GEM, GHITM, FTH1, IFITM2, TXN and TCERG1) examined were found to be either up-regulated or down-regulated as predicted by the differential display autoradiogram bands in Fig. [3](#F3){ref-type="fig"}. However, there was little, if any, relationship between the intensity of the DD band and the quantity of PCR product subsequently detected. These quantitative RT-PCR data confirm our DDRT data and support CXCL12-and CXCL10-mediated gene regulation in the Jurkat line. It should be noted that given that we are utilizing CXCR3-transfected Jurkat T cells in these studies, it is quite possible that differences in CXCR3 and CXCR4 receptor density and signal molecule association may influence the genes induced in response to these chemokines. Moreover, as high doses of chemokine were utilized in the stimulation cultures (1 μg/ml), these supraphysiologic concentrations may also differentially influence the observed differences in gene expression. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Quantitative RT-PCR measurement and verification of CXCL12-induced gene expression.**A detailed description of the procedure employed is found in the *Methods*. Each transcript was normalized to the expression of RPL32 in the same PCR reaction. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Quantitative RT-PCR measurement and verification of CXCL10-induced gene expression.**A detailed description of the procedure employed is found in the *Methods*. Each transcript was normalized to the expression of RPL32 in the same PCR reaction. :::  ::: Additional studies were performed using primary human T cells derived from several different donors to verify the differential CXCL12-and CXCL10-induced gene expression observed in Jurkat T cells. However, despite the clear differences in gene expression in Jurkat T cells after chemokine stimulation for 24 hr in culture, primary resting and anti-CD3/CD28 mAb-activated human T cells demonstrated variable and non-reproducible expression of the geminin, thioredoxin, DDX1, GHITM and TCERG1 gene products (data not shown). This variability may have more to do with the activation state of the primary T cells being utilized and the chemokine receptor expression and density on primary T cells versus the CXCR3-transfected Jurkat T cell subclone. More detailed studies examining CXCL12-induced gene expression in primary human T cells at various stages of activation are the focus of current studies. Moreover, studies using CXCL10 on primary human T cells are quite difficult as the expression of CXCR3 on resting human T cells is quite low to non-existent (\>5% on CD3+ T cells with MFI between 5--12) and may be selectively expressed on certain cell subsets. CXCR3 studies in normal human T cells would require an activation of T cells using IL-2 or Th1 polarizing stimuli and thus may not be a valid comparison with the Jurkat T cells utilized in these studies. Protein expression in Jurkat T cells post CXCL12 treatment ---------------------------------------------------------- To further confirm the expression of several of these genes, Western blot analysis was subsequently performed on total cell lysates of CXCL12-or gp120 IIIB-treated Jurkat T cells. The results shown in Fig. [7A](#F7){ref-type="fig"} demonstrate that, similar to its gene expression, TCERG1/CA150 levels increased in Jurkat T cells cultured with CXCL12 or the HIV glycoprotein, gp120 IIIB, over a 24 hr time period. This expression was found to be CXCR4-dependent as neutralizing antibody to CXCR4 (but not control mouse IgG) inhibited the CA150 increase in response to CXCL12 and gp120 IIIB. This CA150 increase was inhibited by the addition of pertussis toxin, a Gα1 inhibitor. Given that gp120 IIIB binds to both the CD4 and CXCR4 molecules on the surface of human T cells, the similar results between CXCL12 and gp120 IIIB treatment suggests an active signaling role for gp120 IIIB through CXCR4. Figure [7B](#F7){ref-type="fig"} demonstrate that TCERG1 (CA150) and thioredoxin were both increased post CXCL12 treatment. In addition, although not identified in our DD studies, the signaling protein, interferon regulatory factor-1 (IRF-1), was also examined on our gels and increased significantly within Jurkat T cells post CXCL12 treatment. Similar to its gene expression, the gene product, flotillin-1, was found to be decreased post CXCL12 treatment in several experiments; however, these results were found to be highly donor variable (data not shown). We believe that this variability was most likely due the use of total cell lysates instead of membrane preparations. In addition, despite examining numerous lysates preparations and blots, we were unable to detect ferritin heavy chain expression by Western blot in any of these studies. ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **CA150/TCERG1, thioredoxin and IRF-1 protein expression by Jurkat T cells post CXCL12α treatment.**Jurkat cells were treated for 24 hrs with either CXCL12 (A, B) or gp120 IIIB (A) at 1 μg/ml in the presence or absence of control mouse IgG, mouse anti-human CXCR4 mAb, or pertussis toxin (PTx; 200 ng/ml) at 37°C. After incubation, the cell pellets were isolated, counted, washed, and subsequently lysed with the detergent. Protein determinations were then performed. Samples were loaded at 20 μg per lane on a 10% polyacrylamide gel. After electrophoresis, the gels were transferred using a transfer apparatus to an immobilon membrane and stained for CA150/TCERG1 (A, B), thioredoxin (B) or IRF-1 (B) expression via Western blot analysis (shown in panel B for each of these proteins versus control). The results are expressed as fold change versus control expression (post background subtraction). :::  ::: Discussion ========== Studies on the alteration of gene expression following chemokine-receptor ligation and the identification of genes that are differentially expressed can provide relevant and important insight into a variety of biological processes and disease etiologies. To date, numerous approaches, model systems, and techniques have been used to search and identify chemokine-modulated genes \[[@B28]-[@B33]\]. In the present study, we used the PCR differential display method to screen genes and compare the changes that occur following CXCL12/CXCR4 and CXCL10/CXCR3 ligation. Both the CXCR3 and CXCR4 chemokine receptors are broadly expressed in many tissues and ligation to their specific chemokines is known to result in downstream signaling through several different pathways such as Ras, and PI3 kinase. PI3 kinase and JAK/STAT signaling pathways activated by CXCL12/CXCR4 ligation play roles in lymphocyte chemotaxis in response to these signals \[[@B34]-[@B37]\]. Although less well characterized, the interaction of CXCR3 with its ligands, in this case CXCL10, results in increased chemotaxis and activation of the Ras/ERK cascade as well as stimulation of Src phosphorylation and kinase activity and increased activity of phosphatidylinositol 3-kinase and its downstream pathway Akt \[[@B38]\]. Despite knowledge of the activation of multiple cytokine-induced signaling pathways, an understanding of the transcriptional mechanisms whereby CXCR3 and CXCR4 ligation regulates and mediates cellular change remains largely unknown. In the present study, we have identified 31 cellular genes that were either up-or down-regulated in CXCR3-transfected Jurkat T cells following treatment with either CXCL12α or CXCL10. Suzuki and colleagues \[[@B30]\] have recently examined gene expression in Jurkat T cells treated with 380 ng/ml of CXCL12 (a dose that demonstrated optimal Jurkat migration in their hands) for various time periods up to 12 hours in the presence or absence of serum using cDNA microarray gene analysis. The arrays utilized were 2140 cDNA microarray with 1847 unique genes <http://nciarray.nci.nih.gov/cgi-bin/gipo>. Many of the genes identified in this study are associated with detoxification, DNA repair, apoptosis, migration, T cell receptor signaling and interferon signaling. While many of the chemokine-induced genes observed in our study are not found on the arrays used by Suzuki et al. \[[@B30]\], we did identify several genes and proteins with similar functional associations as this group, namely interferon-induced transmembrane protein 2 and interferon regulatory factor-1 (interferon-associated genes), thioredoxin and ferritin (detoxification/redox), growth hormone inducible transmembrane protein (apoptosis), flotillin-1 (cell signaling and migration) and geminin (cell signaling/division). It should be noted that we do believe there are differences in the systems utilized by our groups as we failed to confirm using real time RT-PCR and flow cytometry several of the genes identified by Suzuki et al. Differences in the cell lines (bulk line vs. subclone vs. transfected), culture conditions, serum status, dose of chemokine utilized (380 ng/ml vs. 1 μg/ml) and receptor density may account for such disparity. Regardless of these differences, the relationship between the genes identified in our current study and those within the Suzuki study and their role in chemokine biology and function remains to be determined. Although not previously linked to chemokine receptor-ligand signaling, expression of several of these genes such as interferon-induced transmembrane protein 2 (IFITM2) and growth hormone inducible transmembrane protein (GHITM) are recognized to be a part of the interferon signaling system that includes Janus kinases and their downstream target STAT proteins \[[@B39],[@B40]\]. IFITM2 is a member of the large 1--8 gene family whose members are strongly induced by both type I (IFNα, IFNβ) and type II (IFNγ) interferons \[[@B41],[@B42]\]. However, additional information regarding the molecular function of this protein remains unknown. Likewise, information about the molecular function of GHITM is also quite sparse. GHITM displays some similarity to the testis-encoded transcript (TEGT). While the function of TEGT is also unknown, its amino acid sequence predicts a 26.5 kDa integral membrane protein with seven potential transmembrane domains suggesting a possible receptor function \[[@B43]\]. Interestingly, TEGT is 100% homologous to BAX-inhibitor 1 (BI1) \[[@B44],[@B45]\]. Studies of cells overexpressing BI1 have shown its role as a regulator of cell death pathways controlled by BCL2 and BAX \[[@B44]\]. GHITM has also been referred to by several other names including DERP2, My021, PTD010 and HSPC282. DERP2 is a novel protein originating in human hair papilla cells that has an effect on regulating the growth of hair. The logic behind relating these seemingly inane associations is further supported by knowledge that hair follicle development requires Sonic hedgehog expression \[[@B46]\] that is essential for CXCL12 signaling in the CNS \[[@B47],[@B48]\]. The transcriptional cofactor, CA150, whose gene is now designated transcription elongation regulator 1 (TCERG1) is capable of repressing transcription from many viral and cellular promoters whose initiation is dependent upon the presence of a TATA box \[[@B49]\]. CA150 represses RNA polymerase II (RNAPII) transcription by inhibiting the elongation of transcripts. CA150 is a transcriptional co-activator of HIV-1 Tat and therefore is likely to regulate many cellular genes involved with cell signaling, proliferation and differentiation \[[@B50]\]. A portion of the CA150 molecule contains six FF domains and this region appears to directly bind to the phosphorylated carboxyl-terminal domain of the largest subunit of RNAPII. WW1 and WW2 functional domains are also found in CA150 near the FF domains and appear to fine-tune the repression of transcription through their association with the ubiquitous splicing-transcription factor SF1. At present, CA150 is believed to bind to the phosphorylated C-terminal repeat domain of RNA polymerase II of the elongating RNAPII with SF1 targeting the nascent transcripts \[[@B51]\]. CA150 also has been found by DD to be up-regulated in all-trans retinoic acid (ATRA)-induced apoptosis of H9 and SR-786 T cell lymphoma cell lines \[[@B52]\] and to be significantly increased in striatal and cortical brain tissue from individuals with the neurodegenerative disorder Huntington\'s disease (HD) \[[@B53]\]. Interestingly, a small subset of HD patients with early onset of symptoms have a mutation in the region of the HD gene that increases its binding to CA150 leading to the suggestion that CA150 may interfere with the transcription of genes essential for neuronal survival. Despite these findings, the role of CA150 in T cell activation and survival is currently unknown. Geminin (LOC51053) is a 25-kDa protein that inhibits DNA replication and is degraded during the mitotic phase of the cell cycle \[[@B54]\]. This protein has generated considerable interest due to its critical role in replication licensing \[[@B55],[@B56]\]. In order to successfully replicate, eukaryotic cells must assure that their chromosomes are duplicated only once in each cell cycle. A process called \"licensing\" assures that chromatin can only undergo another round of replication after it has passed through mitosis. Geminin inhibits DNA replication by accumulating during metaphase, binding to and inactivating CDT1 and then undergoing degradation during the metaphase to anaphase transition \[[@B54],[@B57]\]. It is hypothesized that geminin may have evolved to couple S-phase regulation to growth and development signals \[[@B55]\]. As geminin is a powerful negative regulator of the cell cycle, it also may function as a tumor suppressor protein. Flotillin-1 (FLOT1), also known as Reggie-2, encodes a caveolae-associated, integral membrane protein \[[@B58]\]. Caveolae are small indentations on the plasma membrane that are involved in vesicular trafficking and signal transduction. Purified caveolin-rich membranes are enriched for a variety of lipid-modified signaling molecules such as G proteins, Src family kinases, Ras and nitric oxide synthetase \[[@B59]\] and also are populated with members of several families of integral membrane proteins \[[@B60]\]. Recently, flotillin 1 and flotillin 2 have been recognized to comprise a second family of integral membrane caveolin proteins. The function of these flotillins has not been determined but their expression levels are independent of the other caveolin family members. In at least some cell types, caveolins and flotillins are capable of forming hetero-oligomeric complexes and are believed to play a role in receptor signaling \[[@B60]\]. Movement of CCR5 and CXCR4 molecules into lipid rafts is important in the maintenance of receptor conformation and through this mechanism rafts modulate the binding and function of receptors \[[@B61],[@B62]\]. Flotillin 1 and 2 along with stomatin are the major integral protein components of erythrocyte lipid rafts \[[@B63]\]. Although flotillins are major components of caveolae, these proteins may also be components of lipid rafts in other differentiated cell types such as adipocytes, endothelial cells, fibroblasts and immune cells. Thioredoxin (TXN) is a 12-kDa oxidoreductase enzyme containing a dithiol-disulfide active site. It possesses a variety of biological functions including the ability to modulate the DNA binding activity of the ligand-activated transcription factor aryl hydrocarbon receptor (AHR) as well as the activity of several other transcription factors including general transcription factor IIIC, NF-κB, and AP-1 \[[@B64],[@B65]\]. Reactive oxygen species generated through cellular metabolism can function as cellular second messengers through the regulation of numerous signal transduction pathways. Thioredoxin protects cells against TNF-induced cytotoxicity, general oxidative stress and is able to scavenge free radicals \[[@B66]\]. As increasing evidence accumulates that oxidative stress plays a crucial role in many age-associated diseases and various neurodegenerative disorders, the importance of regulating and maintaining cellular redox status by intracellular redox-regulating molecules such as thioredoxin becomes important to maintain tissue homeostasis \[[@B67]-[@B69]\]. Ferritin is a highly conserved iron-binding protein that in its cytosolic form is composed of 2 subunits, ferritin H and ferritin L, each encoded by a distinct gene. Depending on many factors including tissue type, redox status, and the inflammatory state of a given cell or tissue, the ratio of H to L subunits varies greatly \[[@B70]\]. Several proinflammatory cytokines including TNFa, IL-1a (but not IL-1β, IFNg and IL-6) has been shown to transcriptionally induce ferritin heavy chain (FTH1) expression \[[@B71],[@B72]\]. TNFa regulation of FTH1 is through its binding to the p50 and p65 subunits of NFκB \[[@B73]\]. In addition, expression of FTH1 appears to be regulated by insulin, IGF-1 and thyroid hormone \[[@B74],[@B75]\]. Cytokines also regulate the post-transcriptional modification of ferritin possibly through their ability to induce iNOS \[[@B74],[@B75]\]. As noted by Torti & Torti \[[@B70]\], the pathways that link ferritin gene expression with cell stress and altered growth regulation are just beginning to be explored and based on present knowledge are very complex and multifaceted. Conclusions =========== In the current report, we do not reveal a specific physiological role for the gene expression changes that have been observed by DDRT-PCR post CXCL12 or CXCL10 signaling. As noted above, many of the more fully characterized genes have multiple interactions utilizing a number of distinct signaling pathways (e.g., JAK/STAT, AP-1) that are frequently utilized by other cytokine family members. A more detailed understanding of the various genes differentially induced by chemokines via ligation of their cell surface receptors or during the chemotactic process should provide some insight into the process of cell migration and activation and may identify novel targets for therapeutic intervention. Competing interests =================== None declared Authors\' contributions ======================= JEN, RJS, LS, DB, VDD, EMS and DDT performed the experiments. JEN and DDT prepared the figures and wrote the paper. DDT also supervised the work and edited the manuscript. All authors have read and approved the final manuscript. Abbreviations ============= differential display, DD; growth hormone inducible transmembrane protein, GHITM; GEM, geminin; transcription elongation regulator 1, TCERG1; thioredoxin, TXN; DEAD/H box polypeptide 1, DDX1 Acknowledgment ============== The content of this publication does not reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products or organizations imply endorsement by the U.S. Government.

PubMed Central

2024-06-05T03:55:47.277218

2004-8-5

PMC514894

Background ========== *Salmonella*express flagellar (H), polysaccharide (O) and capsular (Vi) antigens which determine strain pathogenicity and therefore variation of these antigens has formed the basis for *Salmonella*serotyping. The Kauffmann-White scheme, first published in 1929, divides *Salmonella*into more than 2500 serotypes according to their antigenic formulae. Within these, 46 O antigen groups are recognised by *Salmonella*serotyping. O antigen synthesis and assembly is encoded by the *rfb*gene cluster which typically contains 12 open reading frames, and ranges in size between serotypes, from approximately 8 kbp to 23 kbp. The variation of O antigens is not due to individual gene sequence variation, but rather to different sets of genes \[[@B1]\]. Approximately 20,000 repeating flagellin proteins polymerise to form the flagellar filament. The ends of the protein are conserved and responsible for the hairpin shape of the subunit while variation in the central region generates the antigenic diversity. Most serotypes exhibit diphasic flagellar antigen expression by alternately expressing two genes, *fliC*(phase 1) and *fljB*(phase 2) which encode flagellins of different antigenicity. *Salmonella*serotyping methods recognise 63 distinct phase 1 flagellar antigenic factors and 37 phase 2 flagellar antigenic factors although the latter are not always present. Some antigenic factors, denoted by square brackets in formulae, may be present or absent without affecting serotype designation. Serotyping methods are stable, reproducible and have high typeability, yet there are several drawbacks, particularly the dependence on availability of antisera considering the ethics, cost and quality control measures necessary to maintain such a supply. Pulsed-field gel electrophoresis (PFGE) \[[@B2],[@B3]\] is currently the bench-mark for molecular subtyping of *Salmonella*, however it is best used in combination with plasmid profiling and ribotyping for strain discrimination for epidemiological purposes \[[@B4]\]. Other approaches include fluorescent amplified fragment length polymorphism (FAFLP) \[[@B5]\] and multi-locus enzyme electrophoresis (MLEE) \[[@B6]\] which sample genomic DNA and provide a view of genetic diversity between strains and partially group some serotypes, but on the whole do not group or identify serotypes. Multi-locus sequence typing (MLST) has been used to discriminate between *Salmonella*strains by sampling variation in a set of housekeeping genes which precludes antigen encoding genes \[[@B7]\]. In 1993, Luk *et al*\[[@B8]\] published a length heterogeneity PCR (LH-PCR)-based method that targeted genes only associated with particular O antigens (A, B, C2 and D), while a more recent study by Fitzgerald *et al*\[[@B9]\] developed a serotype specific PCR assay targeting a single O serotype (O:6,14). Several studies have used a molecular approach to discriminate between particular flagellar serotypes (9, 11--12). *FliC*fragment restriction patterns using a dual enzyme combination allowed differentiation of flagellar types b, i, d, j, l,v, and z~10~but r and e,h nor \[f\],g,m, \[p\], g,p, and g,m,s could be separated using this technique \[[@B10]\]. Hong used restriction fragment patterns of *fliC*and *fljB*for serotyping of poultry *Salmonella*but could not distinguish *S*. Enteritidis from *S*. Gallinarum and *S*. Dublin \[[@B11]\]. Design of a multiplex-polymerase chain reaction (multiplex-PCR) to identify 1,2, 1,5, 1,6, 1,7, 1,w, e,n,x and e,n,z~15~second-phase antigens has been reported \[[@B12]\]. Peters and Threlfall reported *fliC*restriction fragment length polymorphism (RFLP) profiles were not specific enough to differentiate between certain serotypes \[[@B13]\]. To date no studies have attempted a universal molecular serotyping approach. Relevant publicly available sequence data is incomplete, as is epitope mapping information about specific serotypes, therefore approaches are currently being explored to characterise the expressed antigen or the encoding genes as an alternative to traditional serotyping. For *fliC*, evidence from antibody binding studies suggests that sequences of \~300 nucleotides of the central variable region of flagellin correlate with serotype \[[@B14]\] and differences in amino acid sequence can be associated with differences in antigenic specificity. Comparative sequencing has distinguished some *salmonella* serotypes or biotypes \[[@B15]-[@B18]\]. Previous studies have provided full gene sequence for 19 phase 1 flagellar types. The need for a robust single molecular technology to discriminate different serotypes is clear, however sequence data representing all 63 recognised phase 1 flagellar types is incomplete. The aim of this study was to generate full gene sequences for representatives of the majority of phase 1 flagellar serotypes with a view to identifying serotype-specific motifs. These were then used to design a short sequence- or single nucleotide polymorphism-(SNP) based assay targeting characteristic motifs using pyrosequencing. This technology is based on sequencing by synthesis; four nucleotides are added step-wise to a primer-template mix. Incorporation of a nucleotide i.e. extension of the DNA strand, leads to an enzymatic reaction resulting in a light flash. A pyrogram is produced from which the template DNA sequence is deduced. The assay was validated on a panel of 55 strains to initiate a DNA sequence based approach for serotyping *Salmonella enterica*. Results and discussion ====================== Alignment of 106 *fliC*sequences generated in this study and 32 phase 1 flagellin sequences previously published (see Methods section), representing 35 phase 1 flagellar serotypes revealed a clear division of sequences into two groups. Representative sequences are aligned in [Additional file 1](#S1){ref-type="supplementary-material"}. A tree indicating the relatedness of these sequences generated from translated DNA sequence supported this division with a 100% bootstrap value (Figure [1](#F1){ref-type="fig"}). Sequences encoding phase 1 flagellar antigens exhibiting antigenic factors \"g\" or \"m,t\" are referred to as members of the g-complex and the *fliC*sequences of this group clustered exclusively with the non-motile strains Gallinarum and Pullorum on the tree (Cluster I, Figure [1](#F1){ref-type="fig"}). The level of amino acid sequence homology within Cluster I sequences was 90.05%. Sequences not encoding the antigenic factors \"g\" or \"m,t\", formed the second group of sequences (Cluster II), referred to here as the non-g complex. Lower levels (80.3 %) of amino acid similarity were observed within Cluster II. Sero-specific polymorphisms were identified within the central variable region where consensus sequences of Cluster I and Cluster II diverged, between amino acid positions 160 -- 407 (based on amino acid numbering system of the sequenced strain of *S*. Typhimurium (AE008787) represented here as sequence type Typhimurium\_a). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Protein distance tree of H1 antigens. The tree displays the inferred amino acid sequence distances between full H1 antigens. The label displays the H1 antigen, and the *Salmonella*serotype or subgroup from which the sequence was obtained. *Esherichia coli fliC*-*H7*sequence was used to root the tree. Bootstrap values are displayed at major nodes. Sequences labelled with \_a, \_b or \_c indicate an H1 allele found to be encoded by multiple sequences (Additional file 2). :::  ::: *Salmonella fliC*sequences were conserved at their termini and variable in the central region between serotypes \[[@B16],[@B18]\] and clustered according to allele. Amino acid and nucleotide positions described here-in are with reference to the sequenced strain LT2. It was apparent from the alignment of sequences generated in this study that two assays were required, one encompassing Cluster I strains and one for Cluster II. Multiple alignments were created for each cluster and regions of the *fliC*gene containing sero-specific polymorphisms were identified at nucleotide positions 917 -- 933 and 739--749 in Cluster I and Cluster II respectively (Figures [3](#F3){ref-type="fig"} and [4](#F4){ref-type="fig"}). PCR primers were designed to amplify the target region in each sequence (see below). One multiplex PCR was developed for each group containing a mixture of specific primers. All primers designed for short sequence assays in this study are shown in [Additional file 3](#S3){ref-type="supplementary-material"} and the testing algorithm is shown in Figure [5](#F5){ref-type="fig"}. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Sequence motifs at target g. The assay for g-complex strains detected 17 bp of sequence commencing at nucleotide position 917. Fifteen sequence types were identified and differentiated between H1 serotypes. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Sequence motifs at target non-g. The assay for non-g complex strains detected 9 bp of sequence commencing at nucleotide position 739. Sixteen sequence types were identified among non-g complex strains tested. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Algorithm for identification of unknown isolates. Times given are approximate for 96 samples using methods described. :::  ::: Summary of *fliC*sequence variation within the g-complex -------------------------------------------------------- All polymorphisms within the g-complex sequences analysed are displayed in Figure [2](#F2){ref-type="fig"} The target region (highlighted) was selected because it conferred multiple sero-specific amino acid substitutions and was variable at the DNA level. In the 17 bp nucleotide sequence assayed, 15 sequence types were identified (Figure [3](#F3){ref-type="fig"}). This region was assayed against the test panel of 17 *Salmonella*strains belonging to the g-complex and was able to exclusively identify sequence motifs corresponding to phase 1 flagellar serotypes. The serotypes not differentiated by this assay (\[f\],g,m, \[p\], g,m, g,m,s and g,m, \[p\],s or non-motile Gallinarum) were known from full sequencing to be identical at the target region. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Amino acid polymorphisms among *fliC*of g-complex strains. Alignment displaying polymorphic codons only of g-complex *fliC*genes. Codon numbering is based on LT2 sequence, a slash is used where codons fall between LT2 codons in alignment. Highlighted area indicates region analysed in g-complex assay. :::  ::: Amino acid differences between g-complex strains identified by full sequencing ------------------------------------------------------------------------------ The following polymorphisms located in *fliC*of the g-complex are likely to be involved in specific epitope formation: two amino acid sequence types were observed in 25 *fliC-\[f\],g,m, \[p\]*sequences obtained from *Salmonella enterica*serovar Enteritidis strains. Twenty-three *S*. Enteritidis strains demonstrated complete conservation in their DNA sequence (B16, B18, JTCM02 and 20 phage type 4 strains (Enteritidis\_b)). The sequence of B17 was congruent with published *S*. Enteritidis (M84980) (Enteritidis\_a), and exhibited a single amino acid (Ser\>Gly at 302) substitution compared to sequence type Enteritidis\_b. Published *S*. Othmarschen (U06455) *fliC-g,m*, *\[t\]*sequence inferred the same amino acid sequence as Enteritidis\_a but exhibited a silent mutation at the DNA level. As the *fliC*sequence for these two serotypes was identical it was apparent that the sequence included here represented an *S*. Othmarschen strain in which the t factor was absent. Published *S*. Gallinarum sequences demonstrated 100% DNA homology to Enteritidis\_b except for a SNP encoding a stop codon in M84975. *S*. Pullorum and *S*. Gallinarum are non-motile as they do not express flagella. Antisera to the g factor antigen react strongly with induced-motility *S*. Pullorum culture, indicating that g epitopes are expressed in these cells \[[@B19]\]. This correlates with our sequence data as S. Pullorum clusters with g,m sequences (Figure [1](#F1){ref-type="fig"}). Biotype-specific polymorphisms for *S*. Pullorum were observed at amino acid position 91 and 323. Molecular identification of *S*. Pullorum and *S*. Gallinarum would be of considerable benefit as standard serotyping cannot differentiate these two serotypes. *FliC-g,q*was differentiated from all other g-complex sequences by an Asp\>Gly serotype-specific polymorphism observed at position 284 for *S*. Moscow. A Thr\>Ala substitution at residue 304 conferred by a single nucleotide polymorphism (SNP) was identified between sequences of g,m and g,p, congruent with a previous report \[[@B20]\], and forms the basis for differentiation of these two serotypes. DNA polymorphisms, but no inferred amino acid substitutions, were observed between strains exhibiting g,m,s and g,m, \[p\],s. The p factor was not coded for by the *fliC*sequences of these strains. *S*. Essen *fliC-g,m*was distinct from other g and m coding sequences by an Asp\>Asn substitution at 283. *fliC-g,p,s*could be differentiated from *fliC-g,p*by a Thr\>Ala substitution at 254. A motif of two amino acids at positions 302 and 307 was common to *S*. Derby, *S*. Agona, *S*. Adelaide, and *S*. Berta which exhibit phase 1 flagellar antigenic factors \"f\" and \"g\". This motif was exclusive to these serotypes. DNA sequence variation at corresponding positions allowed *S*. Derby and *S*. Agona to be distinguished from *S*. Adelaide and *S*. Berta. *FliC-g,z*~51~; and *fliC-m,t*with *fliC-g,m,t*each form distinct clusters (Figure [1](#F1){ref-type="fig"}). Summary of *fliC*sequence variation within the non-g complex ------------------------------------------------------------ Sequence conservation within alleles that did not encode g or m,t antigenic factors was demonstrated by 97.8 -- 99.1% homology and 80.35% homology was measured in the complex. The high level of variability between alleles in this group did not allow association of specific amino acids to epitope formation that was possible with the g-complex sequences. The quantity and distribution of polymorphic bases observed in this group (specified below) meant that there was a choice of regions that could be used for differentiation. Following testing of four possible regions, the region encompassing amino acids 248--250 was selected for use in the final non-g assay. Each serotype had a unique motif at the target region except *fliC-l,v*and *fliC-l,z*~13~which shared a sequence type (Figure [4](#F4){ref-type="fig"}). Some amino acid sequences were not identical within non-g alleles, including i, r, d, e,h, a and z~4~,z~23~([Additional file 1](#S1){ref-type="supplementary-material"}). A previous study of *fliC-i*sequences reported no variation in a 260 bp region among seven Typhimurium strains \[[@B17]\]. Six full *S*. Typhimurium *fliC*s and a fragment spanning nucleotides 434--1090, corresponding to amino acids 159 -- 400, of a further 20 *S*. Typhimurium strains were sequenced. Three distinct DNA sequences which resulted in translated differences in the expressed peptides were observed within the serotype. Sequence type \"Typhimurium\_a\" was detected in 18 strains, identical to the sequenced strain LT2. Sequence type \"Typhimurium\_b\" was detected in four strains and was differentiated by a SNP at 768, conferring a 256 Glu\>Lys substitution. Sequence type \"Typhimurium\_c\" conferred a Glu\>Lys substitution and an Ala\>Thr substitution at 263 and was found in two strains: 571896 and 571913. Strains 571896 and 571913 were phage type DT104 however, other strains tested did not conform to recognised phage typing patterns so no assured correlation could be made with phage type or other phenotype. *S*. Choleraesuis sequence (*fliC-c*) differed from that published (AF159459) at one nucleotide, conferring amino acid substitution of Thr \>Ser at codon 99. *FliC*sequences of nine *S*. Heidelberg strains were identical, consistent with the results of a previous report \[[@B18]\]. The published sequence for *fliC-r*of *S*. Rubislaw (X04505) differed from *S*. Heidelberg at three amino acids. The *S*. Muenchen sequence determined in this study differs in twelve amino acids to the published *S*. Muenchen (X03395), and differed in 25 amino acids from the *S*. Duisberg sequence in this study. *S*. Anatum, *S*. Newport and *S*. Saintpaul exhibit factors e,h in their phase 1 flagellar. Amino acid sequence was conserved in two strains of *S*. Saintpaul but distinct for each serotype due to four amino acid substitutions at codons 192, 213, 238, 356. *S*. Brandenburg and *S*. Panama exhibit l,v in the phase 1 antigen, no inferred amino acid differences were detected. *FliC-l,v*sequences clustered with *fliC-l,z*~13~(Figure [1](#F1){ref-type="fig"}). *FliC*from three strains exhibiting the z~4~antigenic factor in phase 1 flagellar were sequenced. Cluster analysis grouped these sequences together in the non-g group although they contain regions of sequence similar to g-complex strains (amino acid positions 96 -- 164). Z~4~,z~24~is distinct from z~4~,z~23~and z~4~,z~23~sequences varied within the serotype at seven amino acid positions: 235, 237, 239, 242, 253, 351 and 369. The complex mosaic nature of *fliC*is evident from analysis of amino acid alignment of sequences in particular strains from subgroups in the SARC collection (see Materials and Methods). Molecular serotyping assays --------------------------- By comparison of amino acid sequences coding for antigens of the different serotypes, sero-specific motifs were identified. Individual regions of *fliC*were selected for the g-group and non-g group to provide unique sequence for as many serotypes as possible, while keeping the assay simple to perform and analyse. Two multiplex PCRs were developed for the production of *fliC*amplicon of g-complex strains and *fliC*amplicon of non-g strains. Sero-specific motifs in each amplicon were consequently identified by sequencing-by-synthesis. G-complex assay --------------- Fifteen sequence types were identified in the 17 bp of nucleotide sequence assayed (Figure [3](#F3){ref-type="fig"}). Twenty-seven strains were tested and each produced a recognised sequence motif which differentiated between serotypes. Serotypes would be fully resolved through the detection of further polymorphisms, for example g, \[s\],t and g,t can be separated through additional detection of a A\>G change at nucleotide position 777 conferring amino acid Ser\>Gly substitution specific to g,t. Non-g assay ----------- Fourteen sequence types were identified in the 9 bp of nucleotide sequence assayed (Figure [4](#F4){ref-type="fig"}). Thirty strains were tested, each producing a recognised sequence motif allowing separation of serotypes. Serotypes l,v and l,z~13~gave the same motif at the target region but could be separated by nucleotide substitution A\>G at position 783 conferring a Thr\>Ala change. The stability of the targeted polymorphisms in *Salmonella*phase 1 flagellar antigens was demonstrated through testing on a panel of 55 isolates. The SNP responsible for the antigenic difference between serotypes g,m and g,p was within the target region and so could be differentiated by the assay. The amino acid substitution that separated *fliC-g,p,u*was also encoded within the sequence assayed. Antigens i, r, c, d, b, e,h, k, a, z~41~, z, z~10~, z~4~,z~23~, z~4~z~24~, g,q, g,m,p, g,p,u, \[f\],g,t, g,z~51~and biotype *S*. Pullorum gave unique motifs, l,v and l,z~13~shared a motif. Some serotypes for which certain factors may be present or absent (denoted by square brackets in antigenic formulae) were not separated from similar serotypes: \[f\],g,m, \[p\], g,m and g,m, \[p\],s; \[f\],g,m, \[p\] and g,m, \[t\]; g, \[s\],t and g,t although these could be separated by other DNA polymorphisms as discussed. Two motifs were observed for k, each specific to *S*. Thompson and IIIb. Two motifs were observed for d, specific to *S*. Duisberg and *S*. Muenchen / *S*. Schwarzengrund. Published sequence data for *fliC-m,t*, from serotypes *S*. Banana, *S*. Oranienburg and *S*. Pensacola were included in assay design. The polymorphic region targeted by the assay is predicted to differentiate m,t sequences from other g-complex antigens, and also differentiate *S*. Pensacola from *S*. Banana and *S*. Oranienberg. Strains exhibiting factors m,t were not available for testing. Conclusions =========== A high level of sequence homology between *fliC*genes of g-complex strains was observed. Data produced for this study is congruent with a previous report of g-complex sequences \[[@B16]\]. The genetic basis between distinct antigens in this group of sequences can be a single amino acid substitution. Specific motifs could be identified as the genetic basis for particular antigenic differences and hence their involvement in epitope formation and stability among strains inferred. Full gene sequences were distinct for each antigen analysed in this study. Furthermore, analysis of multiple representatives revealed that some antigens were encoded for by multiple sequences. In these cases DNA sequence based methods are more discriminatory than traditional serotyping methods which do not recognise these as distinct antigens. Assays were designed such that an unknown strain could be identified in respect of its phase 1 flagellar antigen in two steps. The specific PCR acted as the first level of identification and the resultant amplicon was used for the pyrosequencing assay. A positive PCR indicated which of two Pyrosequencing assays to apply. Each assay was uniform in that only one mix of pyrosequencing primers and one dispensation order was needed. All the strains tested were successfully amplified by PCR. As some analyses have been performed on unpublished data, exhaustive testing of the assay will be performed to confirm specificity and typeability of all recognised serotypes. Molecular serotyping will incorporate the desirable properties of serology (typeability, reproducibility, epidemiological significance) together with the advantages of DNA analysis (ability to automate, labour saving, serum independent). Antisera production and associated quality control measures would be unnecessary for a DNA sequence based method. Time-consuming flagellar phase reversal to identify both flagellar antigens is not necessary at the genetic level. Other advantages include reduced labour costs, rapid results in comparison to traditional serotyping methods. DNA sequence data is highly portable and easy to interpret. The method described was easily automated by use of the vacuum preparation tool for the DNA strand separation step and could be further automated by use of robotics for PCR set-up. Result output included a pyrogram, raw text and confidence level; automation of data analysis could be achieved by use of a computer script to screen at a set confidence level and cross-check results against a database of recognised motifs. With the capability to identify approximately 97% of phase 1 flagellar antigens from medically important *Salmonella*strains occurring in England and Wales, the assay can be used now as an economic screen of unknown isolates and alleviate the burden on routine serotyping work. A scheme including the phase 1 flagellar assay and complementary assays for phase 2 flagellar and polysaccharide antigens is currently being piloted and based on incidence data of the top 50 serotypes from 2003, it is anticipated that the scheme will provide a complete molecular serotype for around 80% of isolates and confident prediction of 76% of the remainder. Future work ----------- Alternative sero-specific polymorphisms identified in this study could be exploited by similar assays to allow further separation when antigens did not give unique pyrograms. The alliance of the *fliC*assay to a *fljB*and *rfb*assay would allow the full antigenic formulae of *Salmonella*serotypes to be determined. Common phase 2 flagellar antigens will be selected for sequencing and together with published data will be analysed for sero-specific motifs and a short sequence assay designed with the approach described in this study. In 1993, Luk *et a*l \[[@B8]\] outlined a simple length heterogeneity PCR for identification of *Salmonella*major serogroups A, B, C2, and D. They based their PCR on the presence/absence of genes or sequence polymorphisms within shared genes. Essentially, only serogroups A and D possess a gene to synthesise tyvelose but serogroup A genes carry an early stop codon and do not produce the sugar itself. Only groups B and C2 possess a gene to synthesise abequose but the sequences are distinct. We have also designed a preliminary pyrosequencing assay to distinguish these serogroups based on amplification and short sequences of these genes (data not shown). In summary, epitopes are conformational and it is difficult to determine which amino acids would interact from a linear sequence. However, in the g-complex sequences some amino acid changes could be identified as responsible for differences in antigenic factors because variation was minimal. There is no common factor among the non-g antigens and the sequences are much more heterogenous; there are too many substitutions to draw conclusions about epitope specific sequences. Epitope mapping could be used to further investigate epitopes responsible for antigenic specificity. Methods ======= Bacterial strains ----------------- Strains exhibiting the different phase 1 flagellar antigenic factors were selected from *Salmonella*Reference Collections A, B and C obtained from the University of Calgary. Multiple isolates of *S*. Enteritidis phage type 4, and *S*. Typhimurium phage type DT104 plus a panel of serotyped strains were gratefully received from the *Salmonella*Reference Laboratory, Health Protection Agency, Colindale ([Additional file 2](#S2){ref-type="supplementary-material"}). DNA preparation, PCR and sequencing ----------------------------------- MagNA Pure instrument and Total Nucleic Acid Extraction Kit 1 (Roche, East Sussex). PCR reactions contained 1X PCR buffer, 20 pmoles of FL\_START2, 20 pmoles rFSa1 \[[@B21]\], 1 U Taq polymerase, 0.25 mM of each dNTP, 4 mM MgCl~2~(Sigma-Aldrich, Dorset). PCR amplification of the *fliC*gene was performed with an 9700 GeneAmp PCR System (Applied Biosystems, Cheshire): 35 cycles of 95°C for 60 sec, 50°C for 60 sec, 72°C for 30 sec followed by a 7 min final extension at 72°C. PCR products were purified with Qiaquick spin columns (Qiagen Ltd, West Sussex) and quantitated by gel electrophoresis using Ready-to-Run pre-cast gels (Amersham Biosciences, Buckinghamshire). Fifty to one-hundred nanograms of the purified PCR product was used for cycle sequencing, with specific primers ([Additional file 3](#S3){ref-type="supplementary-material"}) and the CEQ DTCS dye terminator kit (Beckman Coulter, Buckinghamshire). Excess dNTPs were removed from sequencing reactions using GenClean, a 96-well plate format gel filtration system (Genetix Ltd, Hampshire). Sequencing reactions were run on a CEQ 8000XL capillary sequencer (Beckman Coulter). Primers were designed on generated sequence aided by Eprimer3 \[[@B22]\] in a primer walking approach to complete sequencing of the full gene. Sequences generated have been submitted to GenBank (Accession numbers AY649696-AY6497242). Sequence analysis ----------------- Data were analysed and assembled using SeqMan, a component of the DNA Star software package. Multiple alignments were created using BioEdit (Tom Hall, North Carolina State University). Phylogeny inference package Phylip (Joe Felsenstein, University of Washington) was used to compute a distance matrix from protein sequences and build trees illustrating the relatedness of *fliC*sequences. Some previously published sequences were included ([Additional file 3](#S3){ref-type="supplementary-material"}). Polymorphisms postulated to be serotype specific were identified from the alignments of full *fliC*sequences; in-house programme MOP-UPs \[[@B23]\] identified motifs and designed primers to user-specified groups of sequences in the alignment (Anthony Underwood, Health Protection Agency, London). Assay design ------------ Two multiplex PCRs were designed to amplify polymorphic regions of both g- and non-g complex *fliC*sequences. Amplicon sizes were approximately 316 bp for g-complex strains, 170 bp -- 250 bp, (size varied according to serotype) from non-g strains. The order of nucleotide dispensation was tailored to enable the first two dispensations to act as negative and positive controls. DNA preparation and Pyrosequencing ---------------------------------- A 1 μl loop of cells was boiled in 100 μl of sterile distilled water for 10 min at 95°C. One microlitre of lysate was used for PCR. Two PCRs were run in parallel to amplify fragments of the *fliC*gene. Fifty microlitres of PCR product was prepared containing 1 U Taq polymerase, 0.25 mM of each dNTP, 4 mM MgCl~2~. PCR for amplification of g-complex strains used three forward primers: 100 pmol GPYRO-A; 12.5 pmol GPYRO-B; and 12.5 pmol GPYRO-C; and 125 picomoles of reverse biotinylated primer G-REV. PCR for amplification of non-g strains used 14 forward primers (NON-G-PYRO-A, NON-G-PYRO-B etc.) in equal concentrations. Thirteen biotinylated reverse primers (NON-G-REV-A, NON-G-REV-B etc.) were used in equal concentrations. In each 50 μl reaction 125 pmol of mixed forward primer and 125 pmol mixed reverse primer was used. Primer sequences are detailed in [Additional file 3](#S3){ref-type="supplementary-material"}. Thermocycling was performed with an Applied Biosystems 9700 GeneAmp PCR System using a touch-down programme: initial denaturation step of 94°C for 2 min; followed by 17 cycles of 94°C for 20 sec, 66°C (-1°C per cycle) for 30 sec, 72°C for 30 sec; followed by 20 cycles of 94°C for 20 sec, 54°C for 30 sec, 72°C for 30 sec. The excess primers were removed using a filter plate and vacuum system (Genetix Ltd, Hampshire) before visualising the PCR products on the Ready-To-Run agarose system (as previously). Biotinylated single-stranded DNA was immobilized on streptavidin-coated sepharose beads (Amersham Biosciences, Buckinghamshire) with binding buffer. The mixture was agitated at 1400 rpm for 10 min at room temperature. Single stranded DNA bound to beads was isolated from the mixture using a series of wash steps for 5 seconds each in turn, 70% ethanol, 0.2M NaOH and washing buffer. Ninety-six samples were prepared in 2 minutes by automation of strand separation step using a vacuum preparation tool (Pyrosequencing AB, Uppsala, Sweden). A combination of pyrosequencing primers was used for each assay; 20 primers for the non-g complex assay, and three for the g-complex assay ([Additional file 3](#S3){ref-type="supplementary-material"}) into which DNA was eluted. Pyrosequencing primers were annealed to single-stranded DNA on the beads by heating to 80°C for 2 minutes and allowed to cool slowly. Single stranded binding protein, enzyme mix, substrate mix and dNTPs (Pyrosequencing) were added sequentially by the instrument according to the programmed dispensation order. Authors\' contributions ======================= CM carried out the sequencing, constructed the multiple alignments and designed the assays. TP carried out the serotyping. TP advised on *Salmonella*serotyping and provided strains. CA CM SG TP and JL participated in the design of the study. CA conceived of and coordinated the study. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 Amino acid alignment. Amino acid alignment of 106 *fliC*gene sequences representing 32 H1 alleles. Sequences labelled with \_a, \_b or \_c indicate an H1 allele encoded by multiple sequences (Additional file 2). Codon numbering is in reference to the sequence of Typhimurium\_a which represents sequenced strain LT2. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 3 Primers used for PCR and Pyrosequencing. Orientation of the primer is represented by F (forward) or R (reverse) and approximate position is given as nucleotide distance from 5\' end of *fliC*\*These primers were also used as pyrosequencing primers ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 Sequences used in this study. Sequences labelled with \_a, \_b or \_c indicate an H1 allele encoded by multiple sequences ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ The authors would like to thank the Bioinformatics section of the Genomics, Proteomics and Bioinformatics Unit at the Health Protection Agency for their support, particularly Dr. A Underwood. Linda Ward is thanked for her support, advice and *Salmonella*strains used in this study.

PubMed Central

2024-06-05T03:55:47.281103

2004-8-6

PMC514895

Background ========== Most of the existing computational methods for RNA secondary structure prediction fold an already completely synthesized RNA molecule. This is done either by minimizing its free energy (e.g. done by MFOLD \[[@B1]-[@B3]\] and by the programs of the VIENNA package \[[@B4]-[@B8]\]) or by maximizing the probability under a model whose parameters can incorporate a variety of different sources of information, e.g. comparative information, free energy and evolutionary information (e.g. \[[@B9]\], TRNASCAN-SE \[[@B10]\], PFOLD \[[@B11],[@B12]\] and QRNA \[[@B13]\]). All of these programs, including those that predict folding pathways by folding an already synthesized RNA sequence \[[@B14],[@B15]\], therefore disregard the effects that co-transcriptional folding may have on the RNA\'s functional secondary structure. They essentially aim to predict the *thermodynamic RNA structure*, i.e. the secondary structure that minimizes the free energy of the molecule. However, theoretical studies of RNA molecules \[[@B16]\] indicate that the thermodynamic structure of even moderately long RNA molecules need not necessarily correspond to the *functional structure*which confers the desired functionality within the organism to the RNA molecule. RNA molecules are known to fold as they emerge during transcription \[[@B17],[@B18]\]. Transcription is a directed process of variable speed, during which the 5\' end of the RNA molecule is synthesized before its 3\' end. Hydrogen-bonds at the 5\' end of the RNA molecule can thus form earlier in time than hydrogen-bonds involving the 3\' end of the molecule. The thus emerging secondary structure elements can be transient or not, depending on their stability, their formation times and the availability and stability of competing alternative pairing partners. The directedness and also the speed of transcription can influence both the folding pathway and the functional secondary structure of the RNA molecule. We call this phenomenon *sequential*or *co-transcriptional folding*and call the resulting secondary structure the *kinetic structure*of the RNA molecule. Co-transcriptional folding leads to the formation of temporary secondary structure elements \[[@B18],[@B19]\]. The time that it takes to form and replace these transitory structure elements may successively narrow down the set of accessible folding pathways and may thereby guide the folding towards an ensemble of secondary structures which contains the desired functional secondary structure. However, these temporary secondary structure elements can also have distinct biological functions, e.g. in viroids \[[@B19]\] and as initial sites for protein anchoring during pre-mRNA transcription \[[@B20]\]. Based on experimental and theoretical investigations, Harlepp et. al. \[[@B21]\] and Isambert et. al. \[[@B22]\] found that temporary structures may form during transcription. All these results suggest that temporary secondary structure elements may play an important role in the correct folding of RNA sequences. The speed of transcription also has an effect on folding which can be investigated by varying the nucleoside triphosphate concentration \[[@B19]\] or by transcribing RNA genes with viral polymerase T7 which has faster elongation during transcription than bacterial polymerases \[[@B23],[@B24]\]. Both decreasing and increasing the natural speed of transcription can yield inactive transcripts \[[@B23],[@B24]\]. Recent *in vitro*investigations of the *Tetrahymena*ribozyme \[[@B25]\] show that its co-transcriptional folding *in vitro*is twice as fast as the refolding of the entire RNA molecule under the same conditions and that both lead to the same functional folding. Moreover, they find that the co-transcriptional folding *in vitro*is still much slower than *in vivo.* Among the multitude of biochemical processes which are known to occur transcriptionally \[[@B26],[@B27]\], some processes act in order to prevent the mis-folding of RNA molecules. RNA chaperones are proteins which are believed to help refold mis-folded RNA structures by promoting intermolecular RNA-RNA annealing through non-specific interaction \[[@B28]\]. Without RNA chaperones, moderately long GC-rich helices have dissociation half-times of up to 100 years \[[@B29]\]. This time can be significantly reduced by RNA chaperones, which preferentially bind stretches of unfolded RNA and thereby decrease the kinetic barrier between the correct and incorrect secondary structure elements \[[@B28]\]. Specific RNA-binding proteins are also known to promote RNA folding by either guiding its folding or stabilizing its correct structure \[[@B30],[@B31]\]. The hnRNP proteins non-specifically bind pre-messenger RNA and help in the splicing process \[[@B32]\]. RNA sequences can also promote the proper folding of other RNA sequences. It is known, for example, that the temporary interaction with highly conserved leader sequences of bacterial rRNA-operons is needed for the proper formation of 30S ribosomal subunits and the maturation of 16S rRNA \[[@B33],[@B34]\]. All these experimental and the few theoretical findings suggest that co-transcriptional folding may play an important role in the correct folding of RNA molecules. They also show that the functional structure may only be a transient one which is available during a certain time span and that the functional structure need not correspond to the structure which would dominate the ensemble of structures after an infinite time span. Little is known whether co-transcriptional folding is mainly governed by the specific or non-specific binding of proteins (or other molecules) which target the emerging RNA or whether the primary structure of the RNA molecule itself conveys the desired properties to guide its own correct co-transcriptional folding. In this paper, we propose several statistics in order to detect, if and how co-transcriptional folding influences RNA sequences. Using these statistics, we show that the effects of co-transcriptional folding are widespread in RNA genes. Methods ======= Theory ------ We want to show that an RNA sequence is organized in such a way to help the formation of the functional secondary structure during transcription. We aim to support this hypothesis by detecting two different features: • **Possible competitors of helices in the functional structure are suppressed.**When the 3\' end of a helix that is part of the final secondary structure emerges during transcription, the number of possible competitors for the 5\' part of the helix should be as low as possible in order to promote the formation of the correct helix. • **The folding pathway is engineered.**During transcription, several temporary helices are formed which may guide the folding process. We investigate these features using several statistics which are based on the known primary and secondary structures of our RNA sequences. A crucial point in investigating these features is to define a set of statistics that have expectation of zero in the *H*~0~case, when we suppose no co-transcriptional folding. However, verifying that these statistics have an expectation value of zero in the *H*~0~case cannot simply be achieved by analyzing random sequences. Indeed, even generating random sequences is not trivial. First, it is hard to reliably predict the minimum free energy structure for the randomized sequences as most secondary structure prediction algorithms discard pseudo-knots and, even without pseudo-knots, predict only on average about 70 % of the base-pairs correctly. In addition, there is no guarantee that the secondary structure with the lowest free energy would correspond to the functional one. Second, even if the random sequences are generated by a shuffling algorithm which keeps the given secondary structure fixed, it cannot be guaranteed that the fixed structure remains the correct one for the new primary sequence. Generating random sequences therefore provides no straightforward solution for obtaining a *H*~0~statistics with expectation value zero. We circumvent this problem by studying pairs of statistics, where both statistics have the same, unknown expectation value in the *H*~0~case and where one statistics has a bias away from the *H*~0~expectation value in case of co-transcriptional folding, while the other statistics is not affected by co-transcriptional folding. By studying the difference of these two statistics, we thus gain a new statistics with expectation value zero in the case of no co-transcriptional folding and an expectation value larger or smaller than zero in the case of co-transcriptional folding. The statistics (which we will define in detail below) measure the presence of alternative helices which compete for at least one base-pair with the helices of the known secondary structure. These competing alternative helices are required to consists of at least *min*~*stem*~= 9 consecutive base-pairs of type {G - C, C - G, A - U, U - A, G - U, U - G} and are calculated by a dynamic programming procedure in which the known primary and secondary structure of the RNA is fixed, see Figure [1](#F1){ref-type="fig"} for the definition of a competing, alternative helix. We checked that we obtain qualitatively similar results for smaller and larger *min*~*stem*~values (data not shown). While calculating all helices of at least *min*~*stem*~length, we test which of these helices constitute competing alternatives to helices of the known secondary structure and record each such competing case in one of our statistics. These alternative helices may be part of a pseudo-knotted structure and we do not discard them. As each of the two bases *i*and  of a base-pair in a known helix can have a competing alternative base-pairing partner within an alternative helix and as this alternative partner can either be found 5\' (before), 3\' (behind) or between the two strands of the known helix, all cases can be classified into six different classes. Of these six, we discard the two classes where the alternative helix falls between the two strands of the known helix as this un-paired loop region is typically too short to accommodate an alternative helix of at least *min*~*stem*~length. The remaining four classes, see Figure [2](#F2){ref-type="fig"}, can be sub-divided into two *cis-*and two *trans-*alternative classes, depending on whether the known base-pairing partners lie between the alternative base-pairing partners *(trans)*or not *(cis).*The four statistics 3\'*cis, 3\'trans, 5\'cis*and *5\'trans*that we use correspond to these four classes. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Definition of a competing, alternative helix.**Pictorial definition of a competing, alternative helix. The known base-pair between sequence positions *i*and  has to have at least two other directly adjacent base-pairs within the known secondary structure (right) and the competing, alternative helix has to contain an alternative base-pair between sequence positions *i*and *c*(*c*is the competitor of ) which has to be contained within a helix of minimum stem length (left). :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Definition of the statistics.**Pictorial definitions of the four configurations 3\'*cis*, 3\'*trans*, 5\'*cis*and 5\'*trans*which correspond to the four statistics used to measure the directedness of RNA folding. Sequence positions *i*and  form a base-pair within the known secondary structure. Sequence position *c*is an alternative base-pairing partner for *i*(but according to the base-pairing rules therefore not for ) within a competing, alternative helix of a minimum length *min*~*stem*~. See the text for more explanation. :::  ::: It is important to note that even without co-transcriptional folding, the destabilizing effects of competing *cis-*and *trans*-alternative helices are not necessarily the same as the stacking energies are not symmetric with respect to the 5\' → 3\' direction of the RNA sequence \[[@B3]\]. In addition, alternative *cis*-pairing partners are closer to the known pairing partners than *trans*-pairing partners and may thus lead more easily to incorrect helices. We may therefore compare only *cis*-competitors with other *cis*-competitors and *trans*-competitors with other *trans*-competitors. This yields two possible comparisons: 3\'*trans*versus 5\'*trans*and 5\'*cis*versus 3\'*cis,*see Figure [2](#F2){ref-type="fig"}, with which we can measure the effects of co-transcriptional folding. We proceed as follows to detect if co-transcriptional folding takes place: For every RNA sequence of the data set, we detect events of type 3\'*cis, 3\'trans, 5\'cis*and *5\'trans*, where an alternative helix competes with a known helix. Each such event is given two different weights, see Table [1](#T1){ref-type="table"} for an overview of definitions: (1) a weight of 1/ (*d·log*(*l*)), where *d*is the distance between the two competing helices and *l*is the length of the sub-sequence 5\' or 3\' of the known helix on which the competing helix falls, or (2) a weight of \|*G*\| / (*d*·log(*l*)), where the former weight is multiplied by the absolute value of the free energy *G*of the competing, alternative helix. The factor 1/*d*gives alternative helices that are far away from the known helix a smaller weight than closer ones. The factor 1*/log (l)*accounts for the fact that *log (l)*is proportional to the expected sum of 1*/d*statistics for a sub-sequence of length *l*(i.e. the integral ). The free energy factor *G*in the second type of weights gives stable alternative helices which have a larger impact on the folding pathway a greater weight than helices which are easily unfolded. Statistics derived from weights of type 1*/(d*log(*l*)) are denoted by an index *p*(for plain) and those of type \|*G*\| / (*d*·log(*l*)) by an index *g*(for free energy). By summing the weighted counts for each of the four classes of events, we thus arrive at eight different scalar values which characterize each RNA sequence: *3\'Trans*~*x*~*, 3\'Cis*~*x*~*, 5\'Trans*~*x*~and *5\'Cis*~*x*~for *x*∈ {*p,g*}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Definitions of the different statistics. Definitions of the different statistics used. *i*and denote the sequence positions of a base-pair in the known structure, *c*is an alternative pairing partner for *i*(but according to the base-pairing rules therefore not for ), *L*is the length of the RNA sequence, *N*is the number of sequences in the data set and the index *x*indicates the type of weight used. Please refer to the text for a description of how alternative pairing partners are calculated. ::: *x* *p*plain weights *g*free energy weights --------------------------------------------------------------------------------------- ---------------------------------- ----------------------------------------- *3\'cis*~*x*~ 1/((*c*- *i*) log(*L*- *i*)) \|*G*~*ci*~\|/((*c*- *i*) log(*L*-*i*)) *3\'trans*~*x*~ 1/((*c*- ) log(*L*- )) \|*G*~*ci*~\|/((c - ) log(*L*- )) 5\'*cis*~*x*~ 1/((*i*-*c*) log(*i*)) \|G~*ic*~\|/((*i*-*c*) log(*i*)) 5\'*trans*~*x*~ 1/(( - *c*)log()) \|G~*ic*~\|/(( - *c*) log()) *cis*~*x*~ 5\'*cis*~*x*~- 3\'*cis*~*x*~ *trans*~*x*~ 3\'*trans*~*x*~- 5\'*trans*~*x*~ 3\'*Cis*~*x*~ Σ~\#3\'*cis*~3\'*cis*~*x*~ 3\'*Trans*~*x*~ Σ~\#3\'*trans*~3\'*trans*~*x*~ 5\'*Cis*~*x*~ Σ~\#5\'*cis*~5\'*cis*~*x*~ 5\'*Trans*~*x*~ Σ~\#5\'*trans*~5\'*trans*~*x*~ *Cis*~*x*~ 5\'*Cis*~*x*~- 3\'*Cis*~*x*~ *Trans*~*x*~ 3\'*Trans*~*x*~- 5\'*Trans*~*x*~ where *x*∈ {*p,g*}, *y*∈ {3\'*Cis*, 3\'*Trans*, 5\'*Cis*, 5\'*Trans*, *Cis*, *Trans*} ::: We can now define the two statistics which are capable of measuring the two main types of asymmetry within each RNA sequence: *Cis*:= 5\'*Cis*- 3\'*Cis* *Trans*:= 3\'*Trans*- 5\'*Trans* which can calculate for both types of weights. Without co-transcriptional folding, the expectation value of these two statistics is zero. Co-transcriptional folding induces two types of asymmetries by suppressing the number of alternative helices which compete with the final helices (indicated by an increased number of  configurations, see Figure [2](#F2){ref-type="fig"}) and by promoting the formation of transient helices which guide the correct folding (indicated by an increased number of  configurations). Both types of effects are indicated by an expectation value larger than zero for the respective statistics. Without co-transcriptional folding, the introduced statistics have an expectation of zero, moreover, the distributions should be symmetric. The number of positive cases *(pos)*thus follows a binomial distribution with parameter *p =*0.5 and the statistic  where *n*is the number of all cases, approximately follows a standard normal distribution. If this value is sufficiently positive, we have to reject the hypothesis that co-transcriptional folding is not encoded within RNA genes. Data ---- All 16S rRNA, 23S rRNA as well as Group I and Group II type intron sequences with completely known secondary structures were downloaded from the Comparative RNA Web (CRW) Site \[[@B35],[@B36]\], resulting in 304 16S rRNA, 84 23S rRNA, 15 Group I intron and 6 Group II intron sequences from three main taxonomical units (Archea, Bacteria, Eukaryotes) and two organelles, see Table [2](#T2){ref-type="table"}. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Composition of the two data sets. ::: Taxonomic unit all 16S rRNA 23S rRNA Group I Group II ---------------- ----- ---------- ---------- --------- ---------- Data set A Archea 28 22 6 0 0 Bacteria 277 232 45 0 0 Eukaryotes 41 35 6 0 0 Chloroplasts 6 6 0 0 0 Mitochondria 9 9 0 0 0 Sum 361 304 57 0 0 Data set B Eukaryotes 15 0 0 15 0 Bacteria 5 0 5 0 0 Chloroplasts 5 0 5 0 0 Mitochondria 23 0 17 0 6 Sum 48 0 27 15 6 ::: Organellar 23S rRNA sequences frequently contain Group I introns and recent research revealed that the 23S rRNA of several hyperthermophilic bacteria also have Group I intron \[[@B37]\]. Other species only rarely have introns in rRNA genes, however, some 16S rRNA introns are known \[[@B38]\]. rRNA genes in bacteria are encoded in the so-called rrn-operon (see for example \[[@B39]\]). The canonical order of rRNA genes in the rrn-operon is 16S-23S-5S, but some exceptions to this rule are known. In *Vibrio harvey,*the order is 23S-16S-5S \[[@B40]\], but not in *Vibrio cholerae*\[[@B41]\] and *Vibrio parahaemolyticus*\[[@B42]\], whose 16S rRNA sequences were downloaded from the Comparative RNA Web Site. We divided the gathered sequences into two sets: data set A which consists of all RNA sequences that are thought to correspond to the originally transcribed sequence units and data set B which contains all those RNA sequences that are known to differ from the originally transcribed sequence units. Data set B thus contains the Group I and II intron sequences, organellar and hyperthermophilic bacteria 23S RNA sequences. As we neither know the sequence nor the secondary structure of the original transcript units from which the sequences of data set B were derived, we are limited to detecting the effects of co-transcriptional folding within these shorter sequences. We expect this to be much more difficult than in sequences that correspond to the originally transcribed sequence units as co-transcriptional folding introduces long range effects which are harder to detect the shorter the investigated sub-sequence gets. See Table [2](#T2){ref-type="table"} for a detailed overview of the composition of each data set. Results ======= We calculated the *3\'Cis*~*x*~*, 3\'Trans*~*x*~*, 5\'Cis*~*x*~and *5\'Trans*~*x*~values for both types of weights, i.e. *x*∈ {*p,g*}, for each sequence in the two data sets. From these values we then derived each sequence\'s *Cis*~*x*~and *Trans*~*x*~values, again for both *x*types of weights. Their distributions are shown in Figure [3](#F3){ref-type="fig"}. Averaging over the values of all sequences in each of the two data sets resulted in the final values shown in Table [3](#T3){ref-type="table"}. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Distribution of *Cis*and *Trans*values.**Distribution of *Cis*and *Trans*values for the sequences of data sets A and B and both types of weights (plain (p) or free energy based (g)). The area under each curve has been normalized to one to allow a direct comparison between the two data sets. :::  ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Average values for different statistics. Final values of the different statistics which were obtained by averaging the values of each sequence in the data set. The error shown is the standard deviation. ::: dataset --------- --------------- --------------- --------------- --------------- ---------------- --------------- A 0.215 ± 0.009 0.461 ± 0.032 0.285 ± 0.009 0.382 ± 0.032 0.070 ± 0.004 0.079 ± 0.026 B 0.298 ± 0.040 0.562 ± 0.086 0.296 ± 0.043 0.521 ± 0.075 -0.003 ± 0.015 0.041 ± 0.082 dataset A 2.916 ± 0.106 6.236 ± 0.431 3.710 ± 0.111 5.134 ± 0.354 0.794 ± 0.061 1.102 ± 0.384 B 3.392 ± 0.406 7.033 ± 1.050 3.362 ± 0.456 6.380 ± 0.954 -0.030 ± 0.184 0.653 ± 1.253 ::: The first thing to note in Figure [3](#F3){ref-type="fig"} is that all distributions follow approximately a symmetric distribution, thus confirming our theoretical considerations, and that the distributions of data set B are always shifted towards lower values with respect to the corresponding distributions for data set A which are always centered around average values larger than zero. The mean values of *Cis*and *Trans*in Table [3](#T3){ref-type="table"} are positive for data set A for both types of weights, indicating the influence of co-transcriptional folding, whereas they are closer to zero or even negative in the case of data set B. A *Cis*value larger zero means that configurations of type  outnumber those of type , see Figure [2](#F2){ref-type="fig"}. The formation of potential transient helices involving base-pairs between *c*and *i*that can later yield to the final secondary structure element containing the base-pair between *i*and  thus seems to be encouraged. However, these transient structure elements may not be too stable if they are to guide rather than impede the proper folding. The presence of transient helices could thus be further substantiated by showing that these transient helices are less stable than the final helix. In contrast to the  configuration, the competing *ic*helices in the  case are suppressed as they lie 3\' of the final  helix and thus emerge later in time during co-transcriptional folding. A *Cis*value larger than zero can therefore be explained by the presence of temporary helices which may guide the formation of the final, functional secondary structure during co-transcriptional folding. A *Trans*value larger than zero means that  configurations are less frequent than  configurations, see Figure [2](#F2){ref-type="fig"}. In the  configuration, both *c*and  are competing pairing partners for *i*as they both emerge before *i*during transcription. This may lead to the formation of wrong *ci*helices, whereas the order of pairing partners in the  configuration has a lower risk of mis-folding due the *c*emerging only after the  and thus only after the  helix could have already formed. In addition, *3\'Trans \> 3\'Cis*in Table [3](#T3){ref-type="table"} can be interpreted as a stabilization of the final, functional secondary structure. Imagine that the hydrogen bounds of the  or  helix temporarily break up. In the case of the *3\'Trans*configuration, the pairing partners come in the order  along the RNA sequence, whereas they come in the order  in the *3\'Cis*configuration. In the  order, the *c*part is in vicinity to the *i*part, so the possibility of ending up with a wrong refolding due to a *ic*helix is larger than in the  case. Overall, we can thus conclude from the average values in Table [3](#T3){ref-type="table"}, that the sequences of data set A are tailored towards co-transcriptional folding, whereas we cannot reliably detect the effects of co-transcriptional folding within data set B. We detected co-transcriptional folding in data set A by showing that the final secondary structure is actively stabilized *(3\'Trans \> 3\'Cis),*that the formation of temporary helices may guide the structure formation and that these helices may thus be used to actively engineer a folding pathway *(Cis \>*0) and that secondary structure elements which may interfere with the formation of the final, functional secondary structure during co-transcriptional folding are suppressed *(Trans \>*0). In order to quantify the influence of co-transcriptional folding further, we calculated two statistics, a t-test for the hypothesis that the given statistics have an expectation value of zero as well as the p-value of the number of positive cases for our two co-transcriptional folding indicators, see Table [4](#T4){ref-type="table"}. The high p-values for data set B imply that the presence of co-transcriptional folding is not well supported in this data set. However, the corresponding indicators strongly support co-transcriptional folding within data set A. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Statistical significance of results. p-values of t-test for the hypothesis that the final values in Table 3 have an expectation value of zero as well as the p-values for the hypothesis that the number of positive cases follows a binomial distribution with parameter 0.5. ::: dataset A B --------- ----------- ----------- -------- -------- \< 0.0001 \< 0.0001 0.5733 0.6137 \< 0.0001 \< 0.0001 0.5650 0.6137 0.0012 \< 0.0001 0.3093 0.8068 0.0021 \< 0.0001 0.3011 0.5000 ::: Discussion ========== Recent experimental studies \[[@B23],[@B24],[@B19]\] have shown that the proper speed of transcription helps the correct folding of RNA molecules. In addition, theoretical studies \[[@B16]\] indicate that the functional structure of an RNA need not correspond to the minimum free energy structure, even for moderately long RNA molecules. These findings suggest that co-transcriptional folding may play a decisive role in the formation of functional RNA structures. Although our statistics are able to reveal two general effects of co-transcriptional folding within data set A, we cannot conclude that they would be powerful enough to serve as a reliable indicator of co-transcriptional folding for single RNA sequences, as some of the sequences in data set A may not correspond to the originally transcribed sequence units. In addition, all of our statistics consider only a first order effect of co-transcriptional folding by studying alternative helices for the known helices, but do not take higher order effects into account as e.g. alternative helices of alternative helices etc. Based on computer simulations, H. Isambert et. al. \[[@B43]\] conjecture that pseudo-knotted motifs are common in co-transcriptional folding. Pseudo-knotted structures are explicitly included in our statistics, as the corresponding calculations naturally allow for alternative helices which are part of a pseudo-knot and as we do not reject them. Conclusions =========== To summarize, our findings show that co-transcriptional folding is a guiding principle in the formation of functional RNA structure and that it can influence both the primary and potential secondary structures of an RNA molecule. This has several implications. Current algorithms for RNA secondary structure prediction can probably be improved by adopting co-transcriptional folding as a guiding principle rather than only free energy minimization. This may hopefully provide the extra information needed to be able to reliably detect RNA genes \[[@B44]\]. Several groups have already come up with computer algorithms which attempt to fold an RNA sequence co-transcriptionally \[[@B45]-[@B48],[@B22]\]. These findings also have implications for computational methods which infer the phylogeny of RNA sequences, as these consider only co-evolution within the base-pairs of the functional helices, but discard any information due to the conservation of folding pathways and may hence mis-estimate evolutionary times. Similar arguments hold for all comparative studies that aim to detect functional secondary structure elements, since co-evolution of nucleic acids does not necessarily imply that these nucleic acids are base-paired in the final functional secondary structure. As evolution probably not only selects for the correct functional secondary structure, but also for a suitable folding pathway, it should be possible to detect the effects of co-transcriptional folding also in a comparative way. Most importantly, co-transcriptional folding should lead to a better understanding of *how*RNA sequences fold. This should in turn enable us to also understand why some RNA sequences mis-fold and fail to function properly in the organism. Even though protein folding is known to differ in many respects from RNA folding, they also have some features in common \[[@B49]\]. One of the obvious similarities is that both proteins and RNA sequences are synthesized in a directional process. It would thus be interesting to investigate if protein folding is also influenced by *co-translational folding.* In this study, we neither attempted to study the effects that co-transcriptional folding may have on sequences that are transcribed together (e.g. genes in an operon) nor to study the influence that the binding by proteins or RNA sequences or RNA editing may have on the co-transcriptional folding pathway and the final, functional RNA structure. This will almost certainly require more refined investigation methods, but we hope that this study provides enough insight and motivation to start to tackle these exciting questions. Authors\' contributions ======================= I.M.M. proposed this work and contributed the main idea for the statistics. I.M. selected the data and evaluated the statistical significance of the results. Both authors shared the programming tasks and the writing of the manuscript. Acknowledgments =============== I.M.M. acknowledges support from EPSRC grant HAMJW and MRC grant HAMKA. I.M. is supported by a Békésy György postdoctoral fellowship.

PubMed Central

2024-06-05T03:55:47.283683

2004-8-6

PMC514896

Background ========== The emergency department (ED) is intended to treat medical urgencies or emergencies, but a large proportion of visits are due to problems that could be treated in the primary care setting \[[@B1],[@B2]\]. ED services are available 24 hours a day while primary care facilities have limited service hours. In the Israeli health system patients can be referred to the ED by their family practitioner, or by other community health providers, or be self referred. Recently many out of hours community based services have been established but without a significant reduction in visiting rates to in-hospital ED. The visit to the ED constitutes a brief, yet an important point in the continuum of medical care. In today\'s era of cost effectiveness and increasingly competent primary care physicians, ambulatory investigation, treatment and follow-up have largely replaced prolonged and costly hospitalizations \[[@B3],[@B4]\]. The ED visit however, remains a cross-road which may mark a sudden change in the patient\'s medical condition. In many cases it may result in introducing new medications, withdrawing others and recommendation of further investigations. The family practitioner is the one expected to coordinate and carry out the treatment and follow-up. The new information given from the ED should be effectively delivered to the family practitioner, the modality usually used is the discharge letter. The continuation of treatment between hospital departments and the primary care physician had been issued in several studies using discharge letters audit \[[@B5]-[@B7]\]. Raval et. al. assessed the adequacy of the discharge summary in reporting important investigative results and future management plans in patients hospitalized and discharged with a diagnosis of heart failure \[[@B5]\]. They found substantial inadequacies in communicating to the community physician that may have implications for continuity of care and subsequent clinical outcome. Wilson et. al. examined the reliability, effectiveness, accuracy and timeliness of hospital to general practitioner information transfer by discharge summaries. In a retrospective audit of 569 patient discharge summaries and related medical records they found that summaries written for patients discharged from hospital were estimated to be received by the patient-nominated general practitioner in 27.1% of cases \[[@B6]\]. Bolton et al assessed the quality of communications between hospitals and general practitioners. The general practitioner\'s(GP) name was recorded in 88% of audited records. Few inaccuracies were detected in the medications recorded in the discharge summaries, and on contrary to Wilson et al 77% of discharge summaries were received by the GP \[[@B7]\]. The continuity of care between the ED and the primary care physician had been assessed for children with asthma \[[@B8]\] but we did not find data about the continuity of care for adults. To evaluate the continuity of care after ED visits, we evaluated the ED referral and discharge letters, their content, and the documentation of the ED visit in the patients\' primary care files. We have focused on discharges from the internal medicine ED. We expected that in these cases the patients would be followed up by their family practitioner. Methods ======= The study was conducted in the district medical center (Kaplan), serving more than 500,000 inhabitants, and in 12 primary care clinics (32 family practitioners), of The Clalit Health Services in the Rehovot region, Israel. In Israel the entire population have a national health insurance by law and each citizen can choose to be a member of one of four HMOs. Every member of the Clalit Health Services, the largest HMO in Israel, is registered to a single family physician, and have a medical record in his physician\'s clinic. Visits to the emergency department are regulated in the national health insurance law. A referral by a physician or by ambulance is free of charge, but this referral should be with a referral letter and not by a phone call to the ED or to the patient. A self referral may cost the patient a co-payment of up to 100 USD. We reviewed retrospectively all the charts of the ED visits for a period of one month, excluding the visits to the pediatric and the gynecologic-obstetrics EDs (see flow-chart 1). 5,898 visits documented that month, resulted in 4,256 discharges and 1,642 hospitalizations. There were 1,564 discharges from the general ED. Trauma, surgery and orthopedics accounted 2,209 discharges and the rest 483 were from other specialties (ophthalmology, ENT, dermatology etc.). Inclusion criteria were: visit to the general ED, age above 18 years, discharge to the community (not hospitalized) at that visit, living and getting medical care in a family medicine group practice in the Rehovot region. Visits due to accidents, trauma, surgery, orthopedics, ENT, ophthalmology and other specialities were excluded from the study. The 1,564 discharges from the general ED were reviewed and 359 were found to be eligible to this study. Two physicians reviewed independently each ED medical chart. Data extracted included: age and gender of the patient, attendance date and hour, self referral, or a referral by a physician and the final diagnosis in the discharge letter. In cases of referrals the content and format of the referral letter were assessed, including: hand writing quality and whether the referring physician referred the patient with a specific question (for example: rule out new onset angina pectoris, suspected pneumonia, please make a chest X ray etc.). Continuity of communication and care: The primary care files of ED visitors were retrieved and checked for the existence of the ED discharge letter and comments about the visit in the follow-up chart. If the discharge letter and / or any comment on the ED visit in the follow-up chart had been found the case was defined as \"a case with good continuity of care\". The cases in which the family physician was the referring physician we looked for documentation of the encounter prior to ED attendance. Visits to the ED were divided into [\"working hours\" visits]{.underline} -- when the visit took place during working hours of primary care clinics in the community (Sunday to Thursday from 08:00 to 20:00, and Friday 08:00 to 14:00), and [\"out of hours\" visits]{.underline} when primary care clinics in the community are closed (from 20:00 to 08:00 weekdays, and weekends from 14:00 on Friday to 08:00 on the following Sunday). [A referral letter]{.underline} was defined as \"any document written by a medical authority in the community prior to the index ED visit\", including referrals from family practitioners, other practitioners in the community and arrival by an ambulance. [A recurrent visit to the ED]{.underline} was defined as a patient\'s visit to ED within less than two weeks from a previous visit with the same complaint, when in both cases the patient was not hospitalized. [Diagnoses at discharge]{.underline} were coded for a specific diagnosis and for the system involved according to the ICPC coding system. [Statistical analysis:]{.underline} Data was analyzed using distribution analysis and χ^2^tests to investigate the association between categorized variables. Student\'s t tests were used to analyze continuous variables. The analysis was performed using the SPSS package. Results ======= During the study period there were 359 ED visits that were eligible to be included in the study (table [1](#T1){ref-type="table"}, flow chart 1). . 214 (59.6%) visited the ED during the \"working hours\" of primary care clinics, 28 (7.8% of all visits) were recurrent visits to the ED. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Data on 359 visitis to the Emergency Department ::: All visits (359) Referral letter (196) Self referral (163) P Value\* -------------------------------------- ------------------ ----------------------- --------------------- ----------- Gender   Women - 192 (53.5%) 106 (54%) 86 (52.7%) NS   Men -- 167 (46.5%) 96 (46%) 77 (47.3%) [Age]{.underline} (years, mean ± SD) 54.1 ± 18.7 55.1 ± 19.0 52.9 ± 18.4 NS Age distribution  \<45 127 (35.4%) 68 (34.7%) 59 (36.2%) NS  46--65 105 (29.3%) 53 (27%) 52 (31.9%)  66--75 63 (17.5%) 38 (19.4%) 25 (15.3%)  \>75 64 (17.8%) 37 (18.9%) 27 (16.6%) Visit time  Working hours -- 214 (59.6%) 143 (73%) 71 (43.6%) \<0.0001  Out of hours -- 145 (40.4%) 53 (27%) 92 (56.4%) \* -- p Value for comparison between patients with referral letter and self referrals ::: Out of all ED visits only 196 (54.6%) patients had a referral letter, the rest were self-referrals. Referral letters were mainly from the family practitioner (147/196, 75%), 14 (7%) from other practitioners in the community, and 35 (18%) of referrals were by ambulance. The referral letters from the community were legible in 43.4% (70/161), 46.5% (75/161) were barely legible and 10% (16/161) illegible. In only 25/161 letters (15.7% of the referrals) a specific question was asked by the referring physician and in another 32 (20.2%) there was only a general question. The main diagnostic groups according to the ICPC were: respiratory (15.7%), digestive system (18.1%), musculo-skeletal (15.2%) and cardio-vascular (11%). In 9.6% of the cases the discharge letter did not contain a specific diagnosis and the diagnosis fell in the \"general\" category. The most common specific diagnoses were: chest pain (5.9%), abdominal pain (3.9%), other respiratory tract infections (3.7%), asthma (3.1%), back pain (2.8%), COPD exacerbation (2.8%) headache (2.5%), nephrolithiasis (2.5%), vertigo or dizziness (2.5%) and gastroenteritis (2.3%). The \"out of hours\" visitors tended to be younger (52.2 ± 17.5 vs. 55.4 ± 19.5, p = NS) (table [1](#T1){ref-type="table"}). A third of the \"working hours\" visits (71/214) were self referrals as opposed to 63.5% (92/145) of \"out of hours\" visits (p \< 0.0001). Table [2](#T2){ref-type="table"} compares the referring practitioners according to ED visiting hours. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Comparison between the source of referral, in 196 visits according to ED visit hours\* ::: The referring practitioner \"Working hours\"\*\* \"Out of hours\" Total ---------------------------- ----------------------- ------------------ ------- The Family practitioner 125 (87.5%) 22 (41.5%) 147 By ambulance 16 (11%) 19 (36%) 35 Other practitioner\*\*\* 2 (1.5%) 12 (22.5%) 14 Total 143 (100%) 53 (100%) 196 \* -- p \< 0.0001 \*\* -- \"working hours\" visits -- when the visit was during the working hours of primary care clinics in the community \*\*\* -- private practitioners, and out of hours commmunity emergency clinics ::: In 147 cases the reffering physician was the family physician, documentation of the ED referral was found in 32% (47/147) of primary care files. The ED discharge letter was found in 50% (179/359) of the primary care files. A follow-up visit was documented in only 31% (111/359). Neither follow up visits nor discharge letters were found in 43% of the files (153/359). No associations between clinic characteristics (size, place) or family practitioner qualification and ED visit documentation was found. Discussion ========== The Emergency Department (ED) acts as a link between community and hospital based medicine. In Israel a patient who needs non elective admission to a hospital unit must pass through the ED, either with a referral note from a medical practitioner, or as a self referral. Most ED visits, however, do not result in hospitalization, and many could be regarded as primary health care problems \[[@B1],[@B2],[@B9]\]. These patients are discharged directly from the ED to the community and further care of the family practitioner. A visit to the ED is generally not prompted by a benign complaint; The most common reasons for referral include, chest pain, asthma exacerbations and nephrolithiasis, subsequent follow up by the family practitioner can be vital. It was found that most children do not have outpatient follow-up after an ED asthma visit \[[@B8]\]. However, those patients that present for outpatient follow-up have an increased likelihood for repeat ED asthma visits, and this visit should be a key opportunity to prevent future ED asthma visits. The increasing role played by the ED in treating primary care problems has been discussed in a number of recent articles \[[@B9]-[@B11]\]. One aspect, which is important to the ED team, is the logistics and manpower needed to optimize the treatment of these non-urgent patients in ways that will not interfere with emergencies yet providing them adequate care. It is unclear whether the capability and quality of primary care services in the ED should be improved and compete with the community family physicians. This is true especially in Israel where there is a universal national health insurance and every patient can have a personal family practitioner. The continuity of comprehensive management is expected from the family practitioner, and is gaining importance nowadays \[[@B12]\]. To achieve this goal the communication between health care providers who treat the patient is mandatory. In the case of the ED visit, where we found many self referrals and referrals from other physicians, it becomes even more important. The modes of communication are the referral letter and the discharge letter. We have found that the referral letter can be improved both in style (printed instead of illegible hand writing) and content (the referring physician should define and clarify the reasons for referral and his expectations). These problems exist in discharge letters as well \[[@B13]\]. Documentation in the primary care file was poor, only one third of referrals were documented and less than 60% of discharges. This figure is between the 27%--77% that was found by others \[[@B6],[@B7]\]. A possible bias is that some follow-up visits were to specialists. But in the case of discharge from the general medicine ED we presume that most patients were advised to return to their family physicians. It is well known that medical notes are poor in other areas, Miller et al \[[@B14]\] found documentation of only 15% of prescriptions given by family practitioners. They explained one of the causes for this discrepancy as the need of double writing (both the prescription itself and in the medical notes). By introducing carbon copy prescriptions, they achieved an 82% documentation rates in patients\' files. Opila \[[@B15]\] found documentation in out patient medical records greatly improved after employing quality control and a feedback system. With the introduction of computerized medical files in primary care clinics in our region, the need of \"double writing\" will disappear, and this in turn should dramatically improve documentation rate of referrals and discharges to the ED; particularly if a computerized reminder system is used to encourage follow up of referrals by the family practitioner. Limitations ----------- Israeli health care system works in regard to ED use quite different from the US and other countries. Likewhise these results may not automatically be generalized to other health care systems. This study described the written communication between the emergency department and the primary care physician, which is the first and mandatory step in establishing continuation of care. This is only one of the four dimensions of continuity of care in family practice: chronological, geographical, interdisciplinary, and interpersonal \[[@B16]\]. Each of these dimensions may influence the quality of care and be evaluated and studied. Further study is needed to prove the link between documantion of ED visit and good contuniuity of care. Large scale prospective intervention studies are needed to prove that continuity of care between ED and the primary care physician improves outcome and saves money. Conclusion ========== ED visits may have important implications for the patient and his family practitioner. The high rate of ED self referrals together with low documentation rates of ED visits in the primary care charts result in poor continuity of care of ED visitors. Competing interests =================== None declared. Authors\' contributions ======================= All authors read and approved the final manuscript. VS Conceived and designed the study, participated in the collection, analysis and interpretation of data and drafted the manuscript. KE Participated in the statistical analysis, interpretation of data and draft of the manuscript.OY participated in the design of the study, data collection and interpretetio hand draft of the manuscript. SN participated in the design of the study, interpretation of data and draft of the manuscript. All authors have read and approved the final manuscript. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Emergency department (ED) visits that included in the study :::  ::: Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2296/5/16/prepub> Acknowledgments =============== We want to thank Dr Gavriel Plotkin for his participation in the data collection.

PubMed Central

2024-06-05T03:55:47.286671

2004-8-7

PMC514897

Background ========== The use of prescribed psychotropic drugs by older persons has been a subject of interest for several decades \[[@B1],[@B2]\]. Psychotropic drugs are defined as substances that act directly on the central nervous system, affecting mood, cognition, and behavior, and usually include anxiolytics, sedatives and hypnotics, antidepressants, neuroleptics, anticonvulsants, and stimulants. The topic continues to draw researchers\' attention for several reasons, including (1) the high prevalence of older users (especially of benzodiazepines) and their typically long-term consumption, (2) their special vulnerability to drug induced iatrogenesis, (3) the discrepancy between rates of mental disorder and rates of drug use among older people, and (4) inappropriate prescribing. The prevalence of psychotropic drug use among community-dwelling older persons (usually defined as those 65 years and older) varies from about 20% to 48% \[[@B3]-[@B5]\]. More than half of them take psychotropics for six months or longer, and are therefore considered long-term users \[[@B4],[@B6],[@B7]\]. For example, 69% of Canadian elders using benzodiazepines have done so for at least a year \[[@B3]\]. Long-term use is contraindicated because benzodiazepines lack effectiveness beyond a few weeks or months of sustained use for their principal indications, the relief of insomnia and anxiety \[[@B8]-[@B17]\]. Since aging increases the likelihood of drug accumulation and intoxication, older persons are particularly vulnerable to adverse effects from psychotropic drugs \[[@B18]\]. Notable harmful consequences of psychotropic drug use include memory impairment, psychomotor slowing, delirium, falls with a risk of hip fracture, automobile accidents, and psychiatric hospitalizations \[[@B19]-[@B21]\]. Although more psychotropic drug users are found among older persons than any other age group, the prevalence of mental disorders appears to be lower among older than younger adults. This has been shown for major depression, anxiety disorders and sleep disorders \[[@B6],[@B22]-[@B30]\]. Other observations, such as the small rate of hospital admissions for psychiatric reasons among older persons (1.1%) \[[@B4]\] compared to their high rate of psychotropic drug use relative to younger adults\' rate, confirm these findings. The inappropriate prescription of psychotropic drugs may account partially for the discrepancy between low levels of distress and high levels of drug use among older persons. Inappropriate prescriptions include questionable drug combinations (such as two benzodiazepines), excessive treatment duration, and drugs contraindicated for use by older people (such as long half-life benzodiazepines). Authors estimate that one-fifth to one-half of psychotropic drug prescriptions to the elderly are \"inappropriate\" \[[@B4],[@B16],[@B31],[@B32]\]. These findings suggest that older persons\' psychological well-being is not the principal determinant of their psychotropic drug use. In the present article, we review recent empirical studies in order to identify other factors that may account for psychotropic drug use among older persons. A large body of work bears on this age group and their use of psychotropic and other drugs \[[@B33],[@B34]\]. Earlier reviews examined prevalence rates \[[@B35]-[@B39]\], the issues of long-term use \[[@B40]\] and dependence \[[@B41],[@B42]\], as well as various models of drug use \[[@B43],[@B44]\]. However, to our knowledge none focuses on community-dwelling psychotropic drug users. Consequently, the present review critically examines various factors associated with such drug use among this population. Each factor is discussed in terms of the empirical support it has received, of hypotheses put forth to explain its association with drug use, and of suggestions for future research. Methods ======= Reports were selected on April 20, 2001, from the following bibliographic databases: *MedLine, Cumulative Index to Nursing & Allied Health Literature*, *Psychlit*, *Eric*and *Sociological Abstracts*for the years 1990 to 2001. Various keywords were used. Selected reports had to: (1)constitute either a peer-reviewed journal article or a government publication published in English or French; (2) present specific retrievable empirical results concerning older persons, regardless of other age groups studied (some reports \[[@B16],[@B39],[@B45]\] were rejected because of this criterion); (3) report on non-institutionalized, community-dwelling participants at the time of the study; and (4) provide specific results on at least benzodiazepine or antidepressant drug use, regardless of other drug classes studied (some reports \[[@B29],[@B46],[@B47]\] were rejected because of this criterion). A total of 61citations met these criteria in the databases. Eliminating overlaps left only 32 separate reports, all of which were retrieved and are included in this review. Their reference lists were also consulted, but no other study was identified by this mean. The 32 publications report on 30 different studies conducted in ten different countries. Unless indicated otherwise, each report is treated as a separate study. Most reports originated from the United States (31%) and Canada (28%), followed by France (12.5%), Sweden (9.5%), and 3% each (one report) from Australia, Austria, Ireland, the Netherlands, Spain, and the United Kingdom (Table 1 \[see [additional file 1](#S1){ref-type="supplementary-material"}\]). Of 26 reports that disclosed a source of funding, 16 (61.5%) identified government agencies, 6 (23%) government-industry or government-foundation partnerships, and 4 (15.5%) industry. Most studies (62%) collected their data during the 1990s, but several did so during the 1980s (34%) and 1970s (3%). As far as could be determined, the median time from end of data collection to publication was 6 years (range: 1 -- 12 years). Most reports (24, 75%) presented multivariate data analyses, usually multivariate logistic regression, where the independent contribution of various variables to the variance in psychotropic drug use could be ascertained while controlling for the contribution of other variables. Although variables with statistically significant associations (p \< .05) to drug use that remained in the regression equation were often considered \"predictors,\" strictly speaking only an association is demonstrated in this fashion, and the direction of causality can rarely be established. This is especially so when a theoretical model is not specified prior to the data analysis, as was the case in the vast majority of reports. The eight remaining studies conducted bivariate analyses solely, and most examined three or less variables. In the following review, associations of sociodemographic factors and life conditions with psychotropic drug use are examined first. A factor was deemed well supported by the review studies when 70% of the results from all studies that considered this factor were statistically significant and pointed in the same direction. Second, methodological and conceptual issues relevant to the study of psychotropic drug use among the elderly, and salient in the reviewed studies, are discussed. Results ======= Psychotropic drug use and sociodemographic characteristics ---------------------------------------------------------- ### Prevalence of psychotropic drug use Weighted average prevalence rates of psychotropic drug use were estimated from all reports having collected data from probability samples or entire populations (n = 22). In longitudinal studies, data from the last year of the study were used. Average prevalence of any psychotropic drug use from 9 studies was 29.0% (range 11.8% -- 42.5%). For drugs identified in the reports as benzodiazepines, minor tranquilizers, anxiolytics, or sedative-hypnotics, the average prevalence in 13 studies was 21.5% (range 6% -- 43.8%). In 11 studies with data on antidepressants, the average prevalence was 6.9% (range 2.3% -- 14%). Finally, for neuroleptics, it was 3.1% (6 studies, range 1% to 6%). ### Age Older people are more likely than any other age group to use any type of medication \[[@B29]\]. However, it has been observed that medication use decreases in those over 75 years \[[@B4]\]. Some suggest that this occurs because: doctors exercise more caution when prescribing to the oldest old \[[@B48],[@B49]\], survivors into advanced age are healthier and thus use fewer medications \[[@B4],[@B50]\], the elderly face fewer stressful events \[[@B50]\]. The observation of a decline in psychotropic drug use with very advancing age is not universal \[[@B51]\]: Blazer et al. \[[@B6]\] and Mamdani et al.\[[@B49]\] found that it reaches its peak prevalence among those older than 85 years. As Table 2 \[see [additional file 2](#S2){ref-type="supplementary-material"} \] shows, 22 studies carried out 23 tests of the association between age and drug use, but only 8 (35%) of the results showed an age-specific trend (5 found an increase and 3 a decrease). This low percentage prevents firm conclusions about any age trend in regard to psychotropic drug use among older people. ### Gender Over the past three decades, numerous hypotheses have been proposed to explain the higher prevalence of psychotropic drug use among women than men: women are more inclined to reveal their emotional problems to their doctor\[[@B35],[@B36],[@B52]-[@B55]\], to request prescriptions explicitly \[[@B36],[@B54],[@B55]\], to hold more positive views of psychotropic drugs \[[@B50]\]. Some authors have suggested that men prefer to use alcohol rather than prescribed drugs to deal with emotional problems \[[@B56],[@B57]\]. Graham and colleagues \[[@B42]\] failed to support this last hypothesis in a sample of 826 older persons. Other authors suggest that since women live longer than men, they experience more effects of aging, losses and health problems, all of which increase their likelihood of using a psychotropic \[[@B37],[@B49],[@B55],[@B58]\]. Physicians might be more willing to prescribe a psychotropic drug to a woman than to a man \[[@B49],[@B54],[@B55],[@B59]\]. Also, women visit physicians more often, increasing their chance of receiving a prescription \[[@B36],[@B55],[@B57]\]. However, in studies using population-wide databases, Brown et al.\[[@B60]\], Jorm et al.\[[@B55]\], and Weyerer and Dilling \[[@B45]\] controlled both the number of health problems and physician visits and showed that women still used more psychotropic drugs than men. Kirby et al. \[[@B53]\] found the same result when controlling for mental health status. Contradictory results are reported by Mayer-Oakes et al., \[[@B61]\] and Swartz et al., \[[@B62]\] who showed the use of benzodiazepines to be associated not to gender but to physical health status, poorer in women than in men. In this review, 73% of the relevant studies support the established finding that women are more likely to use psychotropic drugs than men. In two studies, the relationship with gender holds only among those aged between 65 and 74 years, and in a third, the results are opposite. However, the studies shed little light on reasons for this disparity. Only two studies identified as a secondary objective to look at the gender issue in relation to psychotropic drug use \[[@B42],[@B49]\]. A few studies tested hypotheses to examine the role of physical health status and mental health status relative to gender among seniors, but none received consistent support \[[@B42],[@B45],[@B53],[@B55],[@B60]-[@B62]\]. No other hypothesis was directly tested in this body of studies. Some authors borrowed hypotheses from studies with middle-aged adults to discuss their results \[[@B49],[@B53],[@B55],[@B63]\]. Authors in other studies which found a statistically significant relationship between gender and psychotropic drug use did not discuss or interpret the association \[[@B6],[@B48],[@B55],[@B60],[@B64]-[@B74]\]. In sum, many studies confirm that more women than men use psychotropic drugs, but no single compelling explanation for this difference among the aged emerges from this body of studies. ### Race The role of culture, ethnicity, and race has received little attention so far in the literature on older persons\' psychotropic drug use. Explaining cultural, ethnic, or racial disparities in use of health services is a complex endeavor, requiring analysis of predisposing, structural, access, and other variables in interaction \[[@B75]-[@B78]\]. In this review, two studies from Canada compared likelihood of use among French and English speakers, one finding an association with French speakers, albeit in a non-probability sample \[[@B79]\]. In addition, six studies from the United States compared the prevalence rate of psychotropic drug use between Whites and African-Americans. All studies but one controlled for income or education, and all found that Whites are significantly more likely to use psychotropic drugs. These results concur with many findings showing differential use of health services along racial and ethnic lines in the United States \[[@B75]\]. Interpretations of the race-specific findings from these six studies focused mostly on professional-and individual-level determinants. Brown et al. \[[@B60]\] suggested that doctors prescribe fewer psychotropic drugs to older African-Americans by prudence, since the former are more sensitive to some drug effects. Other authors proposed that different prescription patterns result from differences in the expression of psychological distress \[[@B62],[@B80]\] or from the lower rate of depression among African-Americans \[[@B6]\]. The smaller proportion of African-Americans and other minorities in the United States with private health insurance also may play a role, especially with regard to newer, more expensive drugs \[[@B81]\]. In that country, more studies are needed to test these hypotheses and others involving structural and access variables, also with other significant minority groups such as Latinos. Beyond racial or ethnic status, studies are needed to understand whether distinct cultural attitudes independently predict psychotropic drug use among older people after various sociodemographic and structural variables are controlled. ### Marital status It has been suggested that older widows would be more likely to use psychotropic drugs because of the distress of bereavement \[[@B82]\]. Eleven of 32 studies examined psychotropic drug use in relation to marital status, but only 4 (36%) confirmed an association between drug use and not being married (single, widowed, divorced, or separated). A spouse\'s death, especially if preceded by long illness, is likely to represent a significant stressful event accompanied by insomnia, anxiety or depression for any person and this can lead to psychotropic drug use. However, strictly speaking, stressful events or psychological distress rather than marital status as such would be implicated, and perhaps only on a relatively short-term basis. In summary, marital status so far is not revealed as a significant factor accounting for psychotropic drug use in older people. ### Socioeconomic status Living in deprived environments and having an unskilled occupation, low income and little formal education are variables associated with psychological distress among adults, and it might be assumed that the same would hold among older people. However, the studies reviewed fail to support this association. Socioeconomic status (SES) is typically operationalized by education, income, and occupation. As Table 2 \[see [additional file 1](#S1){ref-type="supplementary-material"} \] shows, 13 studies examined one of more of these variables. Lower education was significantly associated with drug use in only 3 of 11 results (27%). Lower income was significantly associated with drug use in 3 of 7 results (43%). Two studies examining occupational status before retirement found an association: one found that the self-employed were less likely to use psychotropics \[[@B83]\], the other, using a non-probability sample, found more blue-collar workers among users \[[@B79]\]. In sum, most studies did not support the impact of lower educational level or lower income on psychotropic drug use among older persons. The sparse findings regarding occupation are difficult to interpret. While the importance of SES as a predisposing or mediating influence on drug use might be more established in general population studies, it may be less relevant in the older age group because of a lower rate of psychological distress and greater access to income security plans and retirement pensions that prevent extreme poverty. ### Insurance status In the United States and until recently in Canada, older persons benefited from Medicare or provincial health insurance plans without medication coverage. Insurance coverage has been shown to influence psychotropic drug use by the elderly \[[@B84]\]. Five studies in this review (four from the USA and one from Canada) examined this issue in six different tests but only two (33%) supported a relationship between having insurance coverage for medication and being more likely to use psychotropic drugs. Besides the fact that the most frequent psychotropic drugs used by older persons during the time period covered by the 32 studies in this review (for example, benzodiazepines like lorazepam or tricyclic antidepressants) were available in inexpensive generic versions, no compelling hypothesis exists to account for these results. However, newer psychotropics, such as selective serotonin reuptake inhibitors which reached peak usage during the late 1990s, have been marketed at much higher prices, and the impact of insurance coverage on their use might be shown to differ in future studies. ### Proximity to health centers No study examined meso-level variables related to SES, such as census-tract or neighborhood poverty, except proximity to health centers. The use of health services, including medications, may be increased by proximity to such services \[[@B85]\]. Close or easy availability of health services augments medical consultations (and consequently the request of medication or its prescription). This hypothesis is supported in six of seven studies (85%) in the present review. Life conditions --------------- ### Stressful events Stressful events may increase people\'s likelihood of using psychotropic drugs. Only three studies examined this relationship, each reporting a different result: a positive association \[[@B65]\], a negative association \[[@B50]\], and no association \[[@B79]\]. The precise role of stressful events within a model of psychotropic drug use by older persons clearly needs to be elucidated. To explain the mixed results, future research might consider the transformation of stressors over time: stressful events may impact significantly on the initiation of psychotropic drug use, but may recede to a negligible level in long-term consumption. A longitudinal design is needed to examine this relationship more adequately, but unfortunately none of the 12 longitudinal studies in this review explored stressful events in relation to psychotropic drug use. ### Illnesses and other medications A recent qualitative synthesis of the literature concludes that the pathway to psychotropic drug use among older people typically includes the presence of organic disease \[[@B86]\] (and loneliness, see ahead). Since elderly people suffer from more diseases than younger people, they might use more psychotropic and of course non-psychotropic medications. In another vein, some anxious elderly users of psychotropic drugs might worry excessively about their health and be more likely to consult physicians and receive other medications \[[@B85]\]. In this review, nine of 16 studies (56%) found a significant association between the presence or the number of physical illnesses and psychotropic drug use. In their conceptual model, Gustaffson et al. \[[@B69]\] suggest that affliction by a new disease constitutes a stressful event that worsens an older person\'s mental health status, leading to the need (personally or professionally perceived) for a psychotropic drug. Nonetheless, the results overall are equivocal, perhaps because researchers neglected to consider both the type and the duration of health problems. For example, living with high blood pressure and being struck with congestive heart failure are dissimilar experiences. Typically, researchers counted only the presence or number of illnesses and did not distinguish between individuals who have been living with a disease for a long period of time and those recently experiencing it. As actually measured, the illness factor did not clearly discriminate between older users and non-users of psychotropic drugs. Taking into account both the nature and the duration of illnesses in future studies might better elucidate the relationship between health status and psychotropic drug use. Four of six studies (60%) examining the relationship between psychotropic drug use and use of other medications found a significantly positive association, but the direction of causality remains unclear. ### Health perception Health perception--the self-evaluation of one own\'s health--has been shown to be more strongly associated with psychotropic drug use than actual diagnosis of disease \[[@B33]\]. Some researchers see here an indirect relationship, where poor health perception negatively influences the mental health status of a person, which in turn leads to the request for psychotropic drugs \[[@B65],[@B69]\]. Previous studies support the correlation between mental health status and health perception \[[@B87]-[@B90]\]. In this review, seven of eight studies (87.5%) examining the association between health perception and drug use found a positive relationship. However, two thirds of the reviewed studies were cross-sectional, preventing conclusions about cause-effect relationship, and typically more than 80% of the psychotropic drugs used by their subjects were benzodiazepines. This raises the possibility that some subjects had poor health perception as a result of benzodiazepine consumption. Iatrogenic effects of long term benzodiazepine use may worsen health and functional capacity, hence health perception \[[@B73],[@B91]\]. The present findings confirm that health perception has a place within a model of psychotropic drug use among the elderly, yet leave its precise role unclear. A longitudinal study would help to elucidate this role. ### Social support It has previously been proposed that low social support is associated with psychotropic drug use among older people \[[@B92],[@B93]\]. In 10 studies here reviewed, various scales operationalized constructs identified as social support, social network, social relationships, family relationships, and social isolation. A variable of living alone or with others was included in two of these studies, and subjective reports of loneliness in two studies. However, in 13 tests of the association between some of these variables and drug use, 9 results (69%) failed to observe any significant relationship. All five tests of association with \"social support\" and all three tests of association with \"social relationships\" found no relationship with drug use. The single association of \"social relationships\" was positive, as was one of two of \"living alone.\" Overall, results are counter-intuitive since one would expect that without significant social support to aid in time of difficulties, a person might be more likely to seek medical help, increasing the chances of receiving a drug prescription. Either social support is a minor strand in the tapestry of psychotropic drug use, or the concept of social support may be inadequately operationalized by standard scales. These may need to measure, beyond frequency of social contacts and residential status, subjective judgments of loneliness and of opportunities for social intimacy. Whereas a summed scale score indicating \"social isolation\" was not associated with benzodiazepine use in one study \[[@B61]\], in both studies where the older person\'s feeling of \"loneliness\" as such was elicited, its association with psychotropic drug use was found to be significant after controlling for other variables, including \"social support\" scores \[[@B69],[@B70]\]. Rather than the presence or extent of social support, a subjective feeling of loneliness may influence, directly or indirectly, psychotropic drug use \[[@B86]\]. ### Mental health status The association between a diagnosis or rating of mental disorder or psychological distress and the use of psychotropic drugs is evidently anticipated. In 18 studies, 22 tests of this association were carried out, and it was significant in 17 (77%). An intriguing observation is that two of the four studies that failed to support the association included only the antidepressant drug class \[[@B67],[@B71]\]. While the association between mental health and psychotropic drug consumption is well supported, it is worth noting that the magnitude of the association varies considerably across studies. The following two pairs of longitudinal and cross-sectional large-sample studies using multivariate logistic regression analysis illustrate this variability: benzodiazepine users were 2.13 times more likely than non-users to report depressive symptoms at 10-year follow-up \[[@B6]\], and 6.7 more likely to report emotional or nervous problems \[[@B63]\]. Psychotropic drug users were 2.78 times more likely to report depressive symptoms at 6-year follow-up \[[@B68]\], and 4 times more likely to report anxious or depressive symptoms \[[@B94]\] (Table [3](#T1){ref-type="table"}). We discuss ahead how this variability may stem from methodological differences, especially in the measurement of the key variables in the associations. ::: {#T1 .table-wrap} Table 3 ::: {.caption} ###### The association between mental health variables and psychotropic drug use ::: **Study** Results ---------------------------------------- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ^65^Allard et al. (1995), n = 500 In bivariate analysis, drug use negatively correlated with morale (r = -0.32, p \< 0.001) ^6^Blazer et al. (2000), n = 4,162 In multivariate logistic regression, benzodiazepine use significantly associated with depressive symptoms at 10-year follow-up (OR: 2.13, p \< 0.05) ^68^Dealberto et al. (1997), n = 2,812 In multivariate logistic regression, drug use significantly associated with depressive symptoms at 6-year follow-up (OR: 2.78, p \< 0.001) ^69^Gustafsson et al. (1996), n = 421 In LISREL model, drug use positively correlated (0.63) with poor mental health (chi^2^= 3.38, p = 0.33) ^94^Paterniti et al. (1998), n = 1,389 In multivariate logistic regression, drug users 4 times more likely to report anxious and depressive symptoms than non-users (OR: 4.0, CI: 2.5--6.5) ^63^Gleason et al. (1998), n = 5,181 In multivariate logistic regression, benzodiazepine users 7 times more likely to report emotional or nervous problems than non-users (OR: 6.66, CI 5.1, 8.7) ^74^Taylor et al. (1998), n = 5,222 Depressive symptoms multiply by 3 the likelihood of using hypnotics and by 5.1 the likelihood of using anxiolytics. Anxious symptoms multiply by 4.2 the likelihood of using hypnotics and by 3 of using anxiolytics. ::: ### Sleep In the general population, Ohayon and Caulet \[[@B28]\] have clearly shown that a sleep disorder, such as primary insomnia diagnosed according to *DSM-III-R*criteria \[[@B95]\], is associated with the use of psychotropic drugs. These authors also noted that complaints of poor or inadequate sleep were associated with drug use. In the present review, all five studies (100%) that evaluated this latter association among older persons supported it. Although as mentioned earlier they do not report more sleep problems than younger persons, the presence of sleep complaints appears to play an important role in their psychotropic drug use. ### Medical consultations The number of annual visits to a doctor\'s office has an obvious influence on the number of prescriptions since it has been shown that up to 75% of visits by older people end up with a prescription \[[@B96],[@B97]\]. All seven studies (100%) having examined this factor in this review supported its association with psychotropic drug use. Because no drug can be prescribed without the active participation of a physician, the physician\'s role is undoubtedly critical. However, it might differ over time. It has been previously suggested \[[@B98]\] that the first prescription of a benzodiazepine drug and the renewal prescription are two separate phenomena: the doctor might be more directive during the visit leading to the first prescription, whereas the older patient might be more directive to ensure its renewal. Thus, the physician\'s role in psychotropic drug use among community-dwelling older persons might best be modelled according to the duration of use. In previous research, several characteristics of physicians, notably a high number of consultations, have been significantly associated with the likelihood of prescribing psychotropics to older patients \[[@B99]\]. ### Summary of findings on sociodemographic factors and life conditions Among the 16 factors identified in this review of 32 empirical reports, only the variables of language and stressful events were examined in less than five different studies. Gender, age, and mental health status were most frequently examined. The variables of race, medical consultations, proximity to health centers, sleep complaints and health perception are significantly associated with drug use in all or almost all studies incorporating them. Gender and mental health status are associated with drug use in over 70% of studies, while physical illness and number of medications are associated in slightly over half of relevant studies. However, significant associations with age, marital status, socioeconomic status, and social support are observed in only 27% to 36% of studies in which these factors are examined. Methodological issues in the study of psychotropic drug use by the elderly -------------------------------------------------------------------------- Conflicting results could be explained partially by methodological and conceptual characteristics of many of the studies reviewed. The following discussion focuses on study design, sample, concept definitions and measurements (particularly of mental health status and psychotropic drug use), as well as distinction between short-and long-term psychotropic drug use. Table 1 \[see [additional file 1](#S1){ref-type="supplementary-material"} \] lists these and other methodological characteristics of the studies reviewed, as well as summaries of findings on prevalence of psychotropic drug use. ### Design Descriptive methods (100%) using cross-sectional data (62%) predominate in this body of recent reports of psychotropic drug use in the elderly. Cross-sectional data are vulnerable to sample bias, especially when respondents are not randomly selected, as was the case in 9 of 30 (30%) studies. Given the greater costs involved, the proportion of longitudinal research (38%) seems respectable. Possibly because of cost factors, these studies examined fewer variables in relation to psychotropic drug use. However, besides highlighting the phenomenon of long-term use of psychotropic drugs by the elderly \[[@B55]\], they provided clear demonstrations of the deleterious effect of psychotropic drug consumption on cognitive capacities \[[@B19]\] and of how nursing home admission increases psychotropic drug prescriptions \[[@B5]\], for example. Despite their greater cost and demanding logistics, longitudinal designs are essential to understand more accurately how and why an older person begins, ceases, continues, and modulates psychotropic drug use. Finally, the absence of qualitative studies, which usually allow for a better grasp of elderly users\' own perspectives on their use, is noteworthy; most studies so far have reflected only the researchers\' points of view. ### Samples Virtually all samples (29 in the 30 distinct studies) were composed of older consumers and non-consumers of psychotropic drugs. Comparisons between these two groups allow researchers to identify what differentiates users from non-users of psychotropic drugs. However, one notices from this review that researchers do not always succeed. One possible reason may be that short-and long-term users are combined in one group. Indeed, in almost every study it appears that investigators are comparing long-term users with non users. About 20% to 30% of community-dwelling older persons use psychotropic drugs \[[@B3],[@B84]\], with over two-thirds having been users for more than a year, and over one half for more than one year. Among the very old, up to 93% have consumed for more than a year \[[@B100]\]. Thus, the characteristics of users with less than six months of use are not well known. In one longitudinal study, the one factor that most accurately distinguished long-term users of psychotropic drugs from non-users was having been a user of the drug at the first measurement three years earlier \[[@B68]\]. The researchers found that older users were 15 times more likely to be users 3 years later -- 71 times more likely in the case of antidepressant drugs -- whereas the fact of being depressed increased the likelihood of drug use at follow-up by 4.7 times. Thus, if most psychotropic drug use among the elderly transforms into long-term use, the process of transformation itself requires more observation and explanation, again by means of longitudinal research. Immediate problems leading to the first prescription of a psychotropic drug, such as psychological distress or insomnia, might have little relevance among long-term users. Possibly, drug dependence has developed among some long-term users and sustains long-term consumption \[[@B41]\]. The six-month cut-off period habitually used to distinguish short-from long-term consumption in studies may be too long if dependence is taken into account, as patterns of physiological and psychological dependence may already have become established within such a time frame. However, current studies have not taken into account the dependence issue. For instance, one can wonder how withdrawal symptoms are influencing beginning and long-term users of psychotropic drugs. Fear of withdrawal symptoms might explain long-term use of a certain percentage of elderly users. In one previous study, 71% of middle-aged psychotropic drug users wanted to stop consuming these medications, but half of them feared stopping because of withdrawal symptoms \[[@B47]\]. It seems desirable to increase knowledge on this aspect of the phenomenon among older people, given the well documented potential of benzodiazepines to provoke dependence. It also appears that over the past decade, the use of mixed samples of users and non-users of psychotropic drugs has provided little information about older consumers themselves. Although comparison is essential for understanding, perhaps more studies should target older drug users exclusively. For instance, we know of no study examining mental health outcomes of older psychotropic drugs users over time. One might deem it illogical if no studies had been conducted on the long-term effects of antihypertensive medication on blood pressure. Grad \[[@B101]\] made the same observation when commenting on the few empirical studies of the impact of long-term use of benzodiazepines on the quality of sleep of community-dwelling older persons. ### Mental health status, distress, and psychological well-being The concepts of mental health, psychological distress, depression, and social support are often studied in relation to psychotropic drug use, but operationalized differently across studies. For instance, in the body of studies here reviewed, mental or emotional state is generally measured by standardized, self-rated scales that focus on the manifestation or experience of various symptoms of psychological or bodily distress, or by structured diagnostic interviews aiming to identify mental disorders according to official diagnostic criteria. In all, at least 12 different instruments or scales, 4 different systems of diagnostic criteria (ICD-9 and ICD-10, DSM-III-R and DSM-IV), and 3 unstructured open questions were used in 18 studies to measure concepts identified as depression (12 studies), anxiety (6 studies), psychological distress, mental disorders, life events (3 studies each), psychiatric syndromes (2 studies each), and morale, nervous or emotional disorders, melancholy, nerves, emotional condition, dementia, and sleep problems (1 study each). This variability in concepts and instruments surely impacts the variability of observed relationships between drug use and mental health status (Table [3](#T1){ref-type="table"}). In addition, it is relevant to ask whether the sole use of scales measuring psychiatric symptoms is appropriate to understand psychotropic drugs use among older people. As discussed, the rate of mental disorders, including sleep problems, is lower among older than middle-aged adults, and largely outpaces their use of psychotropic medications. This suggests that many older people are prescribed psychotropic drugs for mild to moderate psychological difficulties that would not qualify as DSM mental disorders, and that might not produce significant impairment in daily functioning. For example, in two studies \[[@B61],[@B79]\], while drug users displayed more depressive symptoms than non-users according to the Center for Epidemiologic Studies Depression scale (CES-D), their scores did not reach the standard threshold of 16 points necessary for a diagnosis of depression. Results such as these suggest that scales less oriented to symptoms, distress, or disorder, such as measures of psychological well being that focus on so-called \"positive\" psychological dimensions such as self-acceptance, autonomy, relationships, and purpose in life \[[@B102]-[@B104]\], should supplement traditional-type scales. Psychological well-being is not merely the reverse of psychological distress, and the relationship of psychological well-being to symptoms of distress is complex \[[@B105]\]. No scale of psychological well-being was used in any of the reviewed reports. However, in one recent study of older persons, Guerette \[[@B106]\] found that among several measures of mental status, measures of psychological well-being produced the strongest associations with benzodiazepine use. ### Measures of psychotropic drug use Some conflicting or ambiguous results across studies can be attributed to how psychotropic drug use should be measured. Two issues are discussed here: what drug classes to include, and what time period of use to consider. Depending on the study, one or more different classes of drugs are included among the substances measured as psychotropic medications, although, for example, indications for the prescription of benzodiazepines differ from those for the prescription of neuroleptics. Using a composite variable of any psychotropic drug use when examining the association with relevant sociodemographic factors or life conditions probably affects the observed association and consequently our understanding of the phenomenon. It would seem better to determine separately what variables are associated with benzodiazepine, antidepressant, and other drug use. Benzodiazepines warrant careful attention, as these drugs may be classified as anxiolytics, sedatives, hypnotics, and anticonvulsants--and we would not expect logically the correlates of anticonvulsant use in the elderly to resemble those of hypnotic use. In addition, researchers rarely specify which drugs they classify as \"minor tranquilizers,\" or how they distinguish between \"anxiolytics\" and \"sedatives.\" These definitional issues have long existed in pharmacoepidemiology and we should not expect them to be solved in these studies, but without assurances that drug categories do not overlap, measures of the dependent variable occasionally remain imprecise. The average prevalence of neuroleptic drug use among community-dwelling older persons, estimated at 3.1% from six studies with probability samples, deserves comment in this respect. It has been common wisdom that such drug use has been rare except in institutionalized samples. The studies also confirm that the main psychotropic drugs prescribed to seniors are benzodiazepines and antidepressants. However, most of these studies, conducted in the early 1990s, probably missed the increased popularity of atypical neuroleptics. These drugs have so far enjoyed a reputation as less toxic drugs than conventional neuroleptics, encouraging their prescription among the older age group \[[@B98]\]. Several problems associated with benzodiazepine use among seniors are not associated with neuroleptics. However, all neuroleptics carry their own substantial risks of adverse effects, such as tardive dyskinesia and cognitive dysfunction (conventionals) and weight gain and diabetes (atypicals) \[[@B107],[@B108]\]. In future studies, investigators will probably need to examine the prevalence of these drugs\' use among the elderly. The category of reversible cholinesterase inhibitors, also known as anti-dementia, anti-Alzheimer\'s, or \"cognition enhancers,\" is also unmentioned in any of the reviewed studies. Introduced in the mid-1990s, drugs such as donazepil and rivastigmine are increasingly prescribed to community-dwelling elderly showing subtle signs of dementia and simple forgetfulness (often termed \"mild cognitive decline\") according to a preventive ethos \[[@B109]\]. The second measurement issue to consider relates to the period of time for which data are collected, a source of confound noted by other reviewers \[[@B41],[@B42],[@B110],[@B111]\]. In the 30 studies, temporal windows varied widely, from \"current use\" (4 studies), \"current and past use\" (1 study), \"regular use\" (2 studies), two days (2 studies), one week (1 study) two weeks (3 studies), one month (3 studies), three months (5 studies), and one year (7 studies). In two studies this period is undefined. When drug utilization is measured based on more than a week, one risks inflating the prevalence rate by including consumers who no longer use the drugs. Conversely, one risks omitting occasional users when referring to the last two days only \[[@B45]\]. Graham and Vidal-Zeballos \[[@B42]\] suggested asking respondents about short-term consumption (past 2 days and \"recent use\") and about long-term use (past year). This suggestion does not however solve problems of inaccurate self-report. Studies have found that older persons frequently hide or deny their use of psychotropic drugs \[[@B112]-[@B114]\]. To avoid these pitfalls, researchers in six studies determined psychotropic drug use according to municipal, provincial, state or national prescription databases held by government agencies or health maintenance organizations. The number of prescriptions, however, can greatly differ from actual drug consumption \[[@B16]\]. For example, about 50% of the elderly do not comply with their prescription regimen \[[@B115],[@B116]\]. and an unknown proportion share their pills with relatives or friends \[[@B85]\]. Finally, researchers in at least 10 studies measured drug use by inspecting medication containers at the home of the respondent \[[@B83]\]. This method confirms that drugs have been obtained but leaves the compliance problem unresolved. A good rationale exists nonetheless to encourage this method, probably the most reliable way to get as close as possible to actual consumption \[[@B55],[@B63],[@B83],[@B111]\]. Given that one third of studies in this review used this method suggests that its cost may be relatively acceptable. Interestingly, only two studies used telephone surveys. Owing to their relatively lesser cost than personal interviews, and since the development of reliable technologies, telephone surveys using random-digit dialing and random-select dialing are increasingly used to contact respondents in health related surveys. Conclusions =========== Without being exhaustive, the present literature review is comprehensive and the range of reports, methods, samples, and geographical locations provides a reasonably solid base to support most recommendations. This review was limited to empirical studies. To better understand the phenomenon of psychotropic drug use among community-dwelling elderly, it is desirable that historical, psychological, sociological, political and regulatory, as well as medical aspects of the phenomenon be considered. While some factors are clearly associated with psychotropic drug use in the reviewed studies (such as race, proximity to health centers, sleep complaints, and health perception), few investigators test specific hypotheses to account for the associations. For example, sparse work focuses on cultural factors that might explain drug use disparities between Whites and African-Americans. Similarly, with most drug users being long-term and most drug treatments for insomnia losing their effectiveness with long-term use, the strong association between sleep complaints and drug use needs thorough examination. One of the least studied aspects of the phenomena in the reviewed studies concerns the role of health care professionals. Physicians, nurses, and pharmacists all interact significantly with community-dwelling older persons, especially with those who take psychotropic medications. Only through a health care professional such as a physician and a pharmacist would the vast majority of older people obtain a psychotropic drug to relieve a sleep problem. Recently, the Internet and mail-order pharmacies without face-to-face interaction have facilitated individuals\' access to prescription drugs. Each professional\'s role, as mentioned, is also likely to alter as the individual\'s phase of consumption transforms from short-to long-term. Researchers must face the challenge to incorporate variables related to health care professionals\' attitudes and behaviors, as well as to new modes of distribution of psychotropic drugs to consumers, in their study of the phenomenon. Approximately one third of community-dwelling older persons use psychotropic medications. If the rate of psychiatric disorders among a population serves as a guideline, then obviously older persons\' use of psychiatric drugs far outpaces these drugs\' standard indications and has extended into areas where drugs have little documented effectiveness. Viewed in this light, the ubiquitous phenomenon of long-term psychotropic drug use should evoke concern and caution. The discipline of nursing can definitely contribute to the rational use of these agents among older people. As suggested in this review, researchers do not still fully grasp the dynamics of psychotropic drug use among this population and creative and rigorous research from several disciplines, and from interdisciplinary perspectives, is needed. However, nurses concerned by the problem of the overuse of medication and its adverse consequences can already implement and evaluate programs to educate older people and allied health care and social service professionals about the risks of psychotropic drugs and alternatives to drugs for the management of everyday anxiety, loneliness, depression, and especially insomnia. Nurses can also actively implement and evaluate drug withdrawal programs aimed at long-term users who have had difficulty in withdrawing, and especially at short-term users who might soon be trapped into dependency and thus long-term use. Conversely, diverse patterns of psychotropic drug use undoubtedly exist among older persons, and positive patterns of use, emanating from users\' own experiences and discoveries, need to be documented and disseminated. Competing interests =================== None declared. Authors\' contributions ======================= PV and DC conducted the literature review and drafted the manuscript. SL and JC revised the literature review and subsequent drafts. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1472-6955/3/3/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional File 1 Characteristics of 32 empirical reports on psychotropic drug use among community-dwelling older persons, 1990--2001. Reports on psychotropic drug use among community-dwelling older persons. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 Factors associated with psychotropic drug use among community-dwelling older persons in 32 empirical reports, 1990--2001. Factors associated with psychotropic drug use among community-dwelling older persons. ::: ::: {.caption} ###### Click here for file ::: Acknowledgement =============== The authors thank the Conseil Québécois de la recherche sociale, the Faculty of nursing of Laval University and of University of Montreal, the University institute of social gerontology of Quebec, the Canadian nurses foundation, the research group on social aspects of health and prevention and the College of health and urban affairs of Florida International University for their financial support.

PubMed Central

2024-06-05T03:55:47.288919

2004-8-13

PMC514898

Introduction ============ Electroconvulsive therapy (ECT) is often regarded by the general public as a controversial procedure for the treatment of mental disorders. This is despite evidence of its safety and efficacy \[[@B1]\], and its benefit over anti-depressants in patients resistant to conventional medications and those with life threatening conditions such as catatonia and depressive stupor. The evidence suggests that in unipolar depression ECT has better efficacy when compared with older tricyclic antidepressants and monoamine oxidase inhibitors, as well as newer drugs such as paroxetine \[[@B2]\]. Notwithstanding the efficacy of ECT, its use is declining in some countries \[[@B3]\], while in a few others, including Italy -- where ECT was first introduced in 1938 by Cerletti and Bini -- it is prohibited. Aside from political reasons and public pressure, the declining trend in ECT use could be the result of the introduction of more effective antidepressants. A further possible explanation for the reduction in ECT use may relate to the concern over adverse effects of the procedure. There are a number of short-term side effects including headache, nausea and, sometimes, brief confusion. However, the main side effect of concern is memory impairment for past events (retrograde amnesia) and for current events (anterograde amnesia) that can last for several months after a course of ECT treatment. Some of these side effects are substantially reduced by advances in safety and the introduction of controlled-current ECT machines. The utilisation of muscle relaxants, anaesthetics and resuscitation equipment, and electroencephalographic monitoring during the application of ECT are considered now considered routine. In addition, ECT guidelines issued by the UK National Institute for Clinical Excellence \[[@B4]\] restrict the use of ECT only to patients with severe symptoms to which \"an adequate trial of other treatment options has proven ineffective\" (p. 5). The risk associated with ECT has also been reduced with the introduction of refined ECT procedures, such as \"maintenance ECT\" or \"unilateral ECT\" (uECT) \[[@B5]\]. It has been suggested that unilateral treatment significantly reduces side effects, especially memory disturbances \[[@B6],[@B7]\]. Despite the well-documented efficacy of unilateral over bilateral ECT, current practice still favours bilateral treatments \[[@B8],[@B9]\]. Unilateral treatment, for the majority of patients, entails that electrodes are placed over the non-dominant, right hemisphere. Given that memory impairment could be reduced by unilateral electrode placement and the fact that placement of electrodes to the dominant hemisphere may cause a greater disturbance in memory compared to non-dominant uECT, determination of cerebral dominance appears to be critical \[[@B10]\]. It is important to note that cerebral dominance here equates to speech dominance, including a lateralised capacity of the cortex to be the locus of language-specific memory traces \[[@B11]\]. Avoiding the stimulation of the speech area will therefore reduce speech dysfunction after ECT. Traditionally, the routine clinical determination of cerebral dominance has been through the assessment of hand, foot and eye dominance. It certainly is an easy and inexpensive approach, but it does not ensure accuracy. Unilateral ECT and cerebral dominance ===================================== The practice of determining cerebral dominance from handedness appears to mirror Broca\'s view that a person\'s handedness is opposite to hemispheric language specialisation. This, however, is incorrect, since there is no \"mirror-image\" cortical language organisation in left-handers. Several attempts to improve cerebral dominance assessment by introducing additional clues such as handwriting posture (i.e. inverted or hooked style versus non-inverted) and familial sinistrality have not substantially improved the prediction as to determination of cerebral dominance \[[@B12],[@B13]\]. For example, the use of hand writing posture to determine speech dominance has been shown to be completely invalid \[[@B14]-[@B16]\]. A great majority of left-handers have also an ipsilateral functional specialisation for language (i.e. left hemispheric, as in the majority of right-handers). Although right-handers are more clearly lateralised than left-handers in this regard, a certain proportion of right-handers have language localised in the right-hemisphere. This has been confirmed by various techniques, ranging from the old and invasive procedures such as the intracarotid sodium amytal test and ECT, to the new and more sophisticated techniques such as functional magnetic resonance imaging (fMRI) and functional transcranial Doppler sonography (fTCD). The pooling of empirical data from a number of studies \[[@B17]-[@B27]\] which used both old and new, non-invasive techniques to determine cerebral dominance for language is shown in Table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Percentages of right- and left-handers with speech localised predominantly in the left, right hemisphere, or bilaterally, according to different studies and techniques ::: Study Right-handers Left-handers -------------------------------- --------------- -------------- ---- ---- ---- ---- Milner, 1975\* 96 0 4 70 15 15 Rossi & Rosadini, 1967\* 99 1 0 40 10 50 Pratt & Warrington, 1972^\#^ 99 0 1 \- \- \- Warrington & Pratt, 1973^\#^ \- \- \- 70 7 23 Geffen et al., 1978^\#^ 92 0 8 67 0 33 Geffen & Traub, 1979^‡^ 84 9 7 61 15 24 Springer et al. 1999^&^ 94 6 0 \- \- \- Pujol et al. 1999^&^ 96 4 0 76 14 10 Szaflarski et al. 2002^&^ \- \- \- 78 14 8 Hund-Georgiadis et al. 2002^&^ 94 0 6 47 12 41 Knecht et al. 2000^†^ \- \- 4 \- \- 27 \* intracarotid sodium amytal test ^\#^ECT ^‡^dichotic listening test ^&^fMRI ^†^fTCD ::: From Table [1](#T1){ref-type="table"} one can see that if the hemisphere for uECT treatment were solely ascertained from handedness assessment, then a small proportion of right-handers and a much larger proportion of left-handers would have treatment administered to the dominant hemisphere. One could also see from it that about 3% of right-handers and 25% of left-handers have speech localised in the right hemisphere. This represents the error rate percentage in both groups if uECT was administered to all patients on the right side of the cranium. However, the overall error rate is lower since the incidence of left-handedness is low, and is likely to be in the range of 6.4% to 12.5% \[[@B28]\]. A strict application of the \"mirror-image\" cortical organisation (i.e. considering left-handers as right-hemisphere dominant and therefore performing left-sided uECT) is even more destructive, illogical, and would increase the error rate. For example, in the survey of the use of ECT by psychiatrists in New Zealand \[[@B9]\], 20% of respondents reported using uECT depending on handedness. Adverse effects caused by determining speech dominance on the basis of handedness would be lower if right-sided ECT was always administered, thus making handedness assessment unnecessary. Given the additional risk of uECT treatment on the dominant hemisphere, which is even more disruptive than bilateral ECT \[[@B29]\], correct identification of cerebral dominance appears to be crucial. The importance of identification of cerebral dominance prior to electrode placement has been highlighted by a number of authors \[[@B6],[@B10],[@B28]\], but routine ECT practice has remained unchanged. From intracarotid injection to transcranial sonography ====================================================== Until recently, an accurate determination of speech dominance prior to a course of ECT treatment was possible only through invasive procedures such as intracarotid sodium amytal test \[[@B30]\], also known as the Wada tests, and through the administration of ECT itself -- the ECT Test \[[@B10]\]. Lateralisation of language capacity using the Wada test is based on the temporary anaesthesia of one half of the brain. The subject in the study receives sodium amytal -- a short-acting anaesthetic -- into (usually) the left carotid artery. This causes the left hemisphere to be temporarily rendered dysfunctional. As a result, if this were the patient\'s dominant hemisphere, the subject\'s language capacity -- primarily speech production -- is affected. Conversely, injecting sodium amytal into the right carotid artery leaves this language capacity intact. By using this technique it is possible to identify precisely which hemisphere hosts language, which is considered necessary for patients who are to go through neurosurgical procedures. Although accurate, the use of Wada procedure in a normal healthy population is generally considered as unsuitable. Using ECT for the determination of cerebral dominance is, as mentioned previously, associated with adverse effects and therefore may not be entirely appropriate, although Weiner \[[@B31]\] suggests giving left and right side ECT alternately followed by the administration of a simple verbal performance test and then continuing treatment with the side associated with the better result. The advent of sophisticated and non-invasive technologies during the 1980s and 1990s has enabled a non-invasive approach to the assessment of speech dominance. One of the most elegant, mobile, and cost effective methods for determining cerebral dominance for speech is functional transcranial Doppler sonography (fTCD). fTCD is increasingly used in both clinical and research settings and is a new and robust technique based on the same principles as fMRI. Subjects in studies using this method are asked to generate as many possible words within 5-second periods after a letter presented on the computer screen cues for word generation. Basically, fTCD measures cerebral blood flow velocity which corresponds to brain activity. The physical foundation for this technique is quite old and is based on the work of the Austrian mathematician and physicist, Christian Doppler (1803--1853), who discovered that the change in pitch results from a shift in the frequency of the sound waves. This means that the speed of a physical object (i.e. blood) can be estimated by measuring the rate of change of pitch. To complete the fTCD procedure, the additional sound produced through the arteries is required. Recently, it has been argued that fTCD can reliably replace the Wada procedure in patients undergoing brain surgery \[[@B32]\]. The validity of fTCD has been established by comparing fTCD with the Wada test, which is considered as the ultimate (gold standard) test of cerebral lateralisation for speech. Several independent studies \[[@B33]-[@B35]\] have found highly significant correlations between these two methods. A high agreement between fTCD and fMRI has also been identified \[[@B36]\] for the assessment of cerebral speech lateralisation. Conclusion ========== This review of the literature on ECT and cerebral dominance provides evidence that fTCD represents a valid and safe alternative to invasive techniques for identifying speech lateralisation. It seems therefore, that fTCD, notwithstanding costs, could be used as a standard procedure prior to uECT treatment to determine cerebral dominance, thereby further reducing cognitive side-effects of ECT and possibly making it more acceptable to both patients and clinicians. Competing interests =================== none declared.

PubMed Central

2024-06-05T03:55:47.293082

2004-8-12

PMC514911

Patterns of drug prescription ============================= Despite tremendous advances in the science and technology of drug development, as well the emergence of guidance and consensus building among scientists, many clinicians, pharmacists, and consumers remain uninformed regarding the scientific basis of establishing bioequivalence, the generic-drug approval process, and the issues related to individualizing therapy in general \[[@B7],[@B8]\]. The consequence may be drug dosing errors: overdosing or underdosing of drugs, resulting in the occurrence of harmful effects or the nonoccurrence of the expected treatment benefit. Recent information \[[@B9]\] indicates that doctors are not consistently prescribing proven treatments at recommended doses, and at times they are not prescribing proven treatments at all. A decrease in dose may decrease the efficacy (*relative risk reduction*\[RRR\]) of therapy and thereby decrease the treatment\'s *net benefit*. Not prescribing an agent will effectively nullify the potential benefit to individuals, and when repeated frequently enough, failure to prescribe the agent will significantly decrease the benefit to the population as a whole. At other times, doctors tend to prescribe a drug more generally than clinical trials dictate. The treatment of a population with lower *outcome prevalence*(OP) decreases net benefit and may lead to *harm*. Overdosing may increase treatment-related harm, and underdosing may erode efficacy; both will result in diminished treatment benefits. Finally, noncompliance on the part of the patient may lead to a decrease in efficacy and a requisite decrease in net treatment benefit. If a patient reduces the dose without totally eliminating the drug, the risk of non-dose-related side effects of treatment may remain. The relationship between the terms mentioned above, which govern treatment success, can be expressed mathematically as follows \[[@B10]-[@B12]\]: Net Benefit = RRR \* OP - Harm The graphical representation in Figure [1](#F1){ref-type="fig"} allows for a series of observations that expand our understanding of the benefits and risks of treatment. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Basic relationships. Net benefit is plotted as a function of outcome prevalence. The *line*represents the relationship with the assumption of a relative risk reduction of 30% and treatment harm of 0%. The *line*is the maximum net benefit that can be attained at any given outcome prevalence. The *x-intercept*, or benefit threshold, represents the outcome prevalence at which net benefit will accrue to individuals and the population as a whole. The point of maximum benefit occurs when the outcome prevalence is 100%; at this point, if harm is absent, the net benefit or efficiency of treatment equals the relative risk reduction or efficacy of that treatment. Benefit decreases proportionately as a function of outcome prevalence. :::  ::: The difficulties in drug prescription that are mentioned above are not related to the physicians\' training or experience but result instead from difficulties in relying with confidence on label claims of efficacy, safety, and interchangeable use of new drugs \"within class.\" The common denominator among these problems seems to be an insufficiently well crystallized knowledge base regarding the proper use and interpretation of general terms such as *equivalence/similarity, pharmaceutical equivalence/therapeutic equivalence*, and *bioequivalence/bioavailability*. Problems are further compounded by the increasing use of generic drugs and the interpretation of such terms as *prescribability and switchability*. Term definitions ================ *Bioavailability*(BA) indicates a measurement of the rate and amount of therapeutically active drug that reaches the general circulation and its presumed site of action \[[@B13]\]. Bioequivalent drug products (BE) -------------------------------- Bioequivalence is the absence of a significant difference in the rate at which, and the extent to which, the active ingredients in pharmaceutical equivalents become available at the site of drug action in the body when administered under similar experimental conditions in an appropriately designed study. A product may also be considered bioequivalent to an innovator product if (a) the difference in rate of drug absorption between the two products is intentional and (b) no significant difference is found in the extent of absorption of the two products when they are evaluated under similar experimental conditions \[[@B13]-[@B17]\]. *Bioequivalence requirement*refers to a requirement, imposed by the Food and Drug Administration (FDA), of in vitro and/or in vivo testing of specified drug products that must be satisfied as a condition of marketing \[[@B17]\]. *Pharmaceutical alternatives*are drug products that contain the identical therapeutic moiety, or its precursor, but not necessarily in the same amount or dosage form as the same salt or ester. Each such drug product individually meets either the identical or its own respective compendial or other applicable standards of identity, strength, quality, and purity, including potency and, where applicable, content uniformity, disintegration times, and/or dissolution rates \[[@B17]\]. Pharmaceutical equivalence -------------------------- To be considered pharmaceutically equivalent, two drug products must (a) contain identical amounts of the same active ingredients in the same dosage form, (b) be formulated to meet the same compendial or other applicable standards of quality and purity, and (c) generally be labeled for the same indications. However, pharmaceutical equivalents may differ in the excipients (e.g., flavors, preservatives) that they contain, as well as in their shape, scoring, packaging, and in certain circumstances, their labeling \[[@B17]\]. *Average bioequivalence*involves assessment of pharmacokinetic parameters such as area under the curve (AUC) and peak concentration (Cmax), as well as calculation of a 90% confidence interval for the ratio of the averages of these parameters for the two products that are compared, usually a test product (T) against a reference product (R). The calculated confidence interval should fall within a conventionally established BE limit of 80% to 125% for the ratio of the product averages \[[@B18]\]. The clinical judgment underlying this BE limit is that a test product with BA measures outside this range would be denied market access. However, in specified circumstances, clinical judgment permits widening or narrowing of the BE limit (e.g., 90% to 111% for narrow-therapeutic-range drugs and drug products). Since the implementation of the FDA Bioavailability and Bioequivalence Requirements in 1977, the assessment of bioequivalence has been a subject of continuous debate \[[@B16],[@B20]-[@B22]\]. This controversy has led to modifications in the bioavailability/bioequivalence regulations and guidelines. Despite these modifications, the process of assessing bioequivalence continues to evolve as scientific consensus emerges on many of the issues driving the debate \[[@B23]\]. The average BE reflects comparison of population averages and therefore fails to assess the subject-by-formulation interaction variance (i.e., the variation of the averages in particular individuals). In contrast, the newer *population*and *individual*approaches reflect differences in the objectives of BE testing at various stages of drug development. These differences are embodied in the concepts of *prescribability*and *switchability*(interchangeability) \[[@B19]-[@B22]\]. These concepts underscore the difference between the population and individual bioequivalence approaches. Population bioequivalence assesses total variability of the measure in the population, and it becomes important when physicians are initially prescribing a medication and they need to rely on the average performance of the drug product \[[@B24]\]. In contrast, the most important consideration for individual bioequivalence rests on the assurance that products deemed bioequivalent can be used interchangeably in the target population (i.e., they exhibit switchability) \[[@B25]\]. In addition to the comparison of averages, the individual bioequivalence approach compares within-subject variabilities and assesses subject-by-formulation interaction. It offers flexible equivalence criteria based on the individual therapeutic window and variability of the reference drug product. Furthermore, it allows scaling criteria for highly variable/narrow-therapeutic-range drugs and promotes the use of subjects from the general population in bioequivalence studies. *Prescribability*refers to the clinical setting in which a practitioner prescribes a drug product to a patient for the first time. In this setting, the prescriber relies on the understanding that the average performance of the drug product has been well characterized and relates in some definable way to the safety and efficacy information from clinical trials. *Switchability*refers to the setting in which a practitioner transfers a patient from one drug product to another. This situation arises with generic substitution, as well as with postapproval changes by an innovator or a generic firm in the formulation and/or manufacture of a drug product. Under these circumstances, both the prescriber of the drug and the patient should be assured that the newly administered drug product will yield safety and efficacy comparable to that of the product for which it is being substituted. However, such a switch may, in fact, occur without the patient\'s or clinician\'s knowledge, and this concern is addressed in equivalence studies designed to minimize the risk to the patient in both situations. Although average bioequivalence is the recommended parameter for most bioequivalence studies, the current FDA guide titled Statistical Approaches to Establishing Bioequivalence \[[@B26]\] recommends that population and individual bioequivalence also be evaluated in some cases. Understanding of this process is enhanced by the outline in Figure [2](#F2){ref-type="fig"}, which illustrates the classical exposure-response relationship that might assist in adjusting dosages and dosing regimens in the presence of influences on pharmacokinetics (PK) by demographic factors (e.g., age, gender), intrinsic factors (e.g., impaired organ function), or extrinsic factors (e.g., concomitant medication, food intake). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Exposure-response relationships. Relationships between (1) drug substance and drug product, (2) exposure expressed as dose or systemic exposure on log scale, and (3) positive (efficacy) or negative (toxicity) outcomes. These outcomes may be measured by clinical end points, surrogate endpoints, or biomarkers. The relationships between exposure and outcomes define the optimal dose and therapeutic window. The term *change*introduces the concept of equivalence in outcomes before or after a specified change (eg, generic substitution, postapproval manufacturing change). CMC, Chemistry, manufacturing, and controls; BA/BE, bioavailability and bioequivalence; PK, pharmacokinetics; PD, pharmacodynamics. :::  ::: The outcomes (clinical benefit, reflecting the \"response\" component of the relationship), can be measured as clinical or surrogate endpoints or as biomarkers. The relationship between exposure and outcomes, expressed as dose- or concentration-response curves that have efficacy and toxicity levels at their extremes, define the optimal dose and the therapeutic window \[[@B27]\]. Theoretically, these curves should be generated in different individuals for developing prescribability criteria, and in the same individuals for developing switchabiliy criteria. Premarketing risk-benefit assessment ==================================== The approach described above is a simplified framework of successive steps to be taken in phases I and II of the development of any drug, with the aim of generating data for drug labeling (i.e., data on dose-response \[effectiveness and toxicity\] relations of the new agent and how these depend on patient characteristics). The dose regimen of such a new agent, explored and established through phases I and II, is to be demonstrated in phase III as \"safe and effective\" for the claimed indication. Proper study design and research methodology, as well as appropriate statistical analysis, should be applied to ensure that the drug\'s effectiveness can be substantiated. Further, the outcome derived from two such randomized clinical trials should document that the estimate of the true treatment effect favors the new drug over the reference drug and that the toxicity of the new agent does not exceed acceptable limits. In other words, a first risk/benefit assessment would favor the new drug. Having come that far, a question justifiably arises: what logic leads to the conclusion that the new drug is likely to be effective in future patients? That is, are the treatment benefits generalizable to an actual clinical population? The question is extended to *external validity*, meaning the extent to which the conclusions of a study would hold true for other persons, in other places and at other times. Examples from most therapeutic areas indicate that approved drugs may not work as expected when applied in broad community populations, in real-world settings, and among diverse practitioners operating under real-world constraints \[[@B28]\]. In such instances, years of effort and huge investments fail to meet the intent of providers, the expectations of consumers, or the demands of healthcare payers. Conventionally, the dose-response trials mentioned above, exploring the behavior of biomarkers both cross-sectionally between individuals treated with different doses of an agent, and longitudinally within individuals as doses (or concentrations) change with time, supply the necessary information on the drug\'s pharmacologic action (the so-called *empiric confirmation*at a conventional *α*level). However, despite all their methodological rigorousness, these trials may not entirely eliminate factors that influence bioavailability in earlier development stages and internal validity in later-phase studies (randomized controlled trials \[RCTs\]). Bioavailability may be influenced by • Patient-related factors, such as concurrent diseases, differences in first-pass metabolism, interactions with concomitant medications, diet, circadian biorhythms, the influence of fed-versus fasted-state physiologic conditions, and gastrointestinal factors (e.g., pH, motility, blood flow, bacterial flora) \[[@B29]-[@B32]\]. • Product-related factors, such as: physical and chemical properties of the drug (e.g., solubility, degree of ionization, crystalline forms, chemical form, isomers), as well as variables related to manufacturing, formulation, or both (e.g., coatings, compression force, particle size, presence or absence of excipients) \[[@B30]-[@B33]\]. To control for as many variables as possible, most bioequivalence trials are conducted with healthy volunteers as subjects, \[[@B19],[@B34],[@B35]\] and in real patients only in circumstances wherein the use of volunteers would be unethical (e.g., tests with cytotoxic drugs) or when assessment of bioequivalence is based on pharmacodynamic and/or clinical end points \[[@B35]\]. The *causal confirmation*is much more complex than the empiric confirmation, even if the drug\'s previously established pharmacologic action is believed to ensure that the drug has the same intrinsic property that alters the clinical outcome -- in the RCT patients -- in a similar way and to a similar extent as in the previous patients. A number of other causes, however, may interfere in the conduct phase of an RCT. These causes include • Confounding -- a distortion of an association between an exposure and disease brought about by extraneous factors. • Interaction -- the interdependent operation of two or more factors that produce an unanticipated effect. • Transience -- an idiosyncratic property of a drug that displays its expected pharmacologic property when tested in one batch, but not in others. Many of the specific design and analysis features applied to RCTs (e.g., blinding, randomization, intention-to-treat analysis) are meant to ensure that the possibility of confounding is minimized or eliminated. The same does not apply in the case of transience or interaction, for which independent evidence is needed to eliminate those possibilities. Highly variable drugs ===================== Drugs that tend to exhibit high degrees of variability in their pharmacokinetic profiles are known to complicate the assessment of bioequivalence \[[@B22],[@B36],[@B37]\]. This is, at least in part, a function of the high intrasubject variability (previously defined as greater than 30%) of the drug or drug product \[[@B22]\]. Examples of such drugs are propafenone immediate release, verapamil, and nadolol. Narrow-therapeutic-index drugs ============================== Small changes in systemic concentration of such drugs can lead to marked changes in pharmacodynamic responses \[[@B19]-[@B22]\]. A broader term for the narrow-therapeutic-index drugs is \"critical-dose drugs.\" Characteristically, these agents require blood-level monitoring, need to be dosed on the basis of body weight or other individualized parameters, display serious clinical consequences if overdosing or underdosing occurs, and manifest a steep dose-response relationship \[[@B38]\]. A typical drug in this category is warfarin, which is widely used for its anticoagulant properties. For the most part, metabolism of S-warfarin occurs by means of the gene CYP2C9 \[[@B39]\]. Inhibition of this isoform results in several clinically important drug interactions. Fluconazole, metronidazole, miconazole, and amiodarone are a few examples of the many drugs that profoundly inhibit S-warfarin metabolism and produce marked increases in prothrombin time (PT) measurements \[[@B40]-[@B43]\]. A multitude of endogenous and exogenous factors may contribute alone or in combination to either increasing or decreasing PT ratio or the INR response \[[@B44]\]. Factors that increase PT ratio or international normalized ratio (INR) response include: Endogenous factors: 11 Exogenous factors: 117 specific drugs and 49 different classes of drugs Factors that decrease PT ratio or INR response include: Endogenous factors: 5 Exogenous factors: 42 specific drugs and 24 different classes of drugs Physician surveys have indicated that most clinicians favor more rigid bioequivalence guidelines for these types of drugs \[[@B8]\]. Others have recommended that the bioequivalence requirement for these agents be based on intrasubject variability, as well as the pharmacokinetic-pharmacodynamic relationship. Although the FDA has not modified the bioequivalence guidelines for critical-dose drugs, the Canadian regulatory authority, Health Canada, has narrowed the CI requirement for these drugs to 90% to 110%. For prescribability, the current requirements may be adequate for all drugs, including those with a narrow therapeutic index. However, some clinicians have expressed concerns about switchability \[[@B7],[@B39]\]. Several reports suggest that once a patient has been carefully titrated on a narrow-therapeutic-index drug, the formulation should not be switched \[[@B7],[@B8]\]. The same might be true of warfarin. Drug products suspected of having bioequivalence problems are listed in the Orange Book (Approved Drug Products with Therapeutic Equivalence Evaluations) \[[@B45]\]. Such drugs may exhibit a narrow therapeutic index or solubility problems, or they may be poorly absorbed or unstable in gastrointestinal fluids \[[@B33]\]. Among clinicians there seems to be widespread concern that even small changes in the bioavailability of drugs\' active ingredients might lead to significant changes in the efficacy or safety of those products \[[@B7],[@B8],[@B38]\]. The concept of risk and its application to drug development =========================================================== In light of serious concerns about risks incurred from using medical products, a variety of public and private agencies involved in health care are dedicating more attention to examining the current system of managing these risks. The main goal is to focus on the costs and value of better data concerning the incidence and causes of injuries from medical products and the roles of all stakeholders in risk management. The FDA has issued a concept paper \[[@B46]\] presenting preliminary thoughts on risk issues, including: • Important risk assessment concepts. • Generation and acquisition of safety data during product development. • Analysis and presentation of safety data in an application for approval. *Risk assessment*, defined as the process of identifying, estimating, and evaluating the nature and severity of risks associated with a product, should be continuous throughout the life cycle of any product, whereas the process of *risk management intervention*is intended to enhance the safe use of a product by reducing risk. With regard to the current trend toward *systemic risk confrontation*, the FDA, Center for Drug Evaluation and Research (CDER), and Center for Biologics Evaluation and Research (CBER) appear ready to develop and finalize, by fall 2004, guidance documents regarding risk assessment, clinical pharmacovigilance, and risk management. Table [1](#T1){ref-type="table"} highlights the components of a risk management system (RMS). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Components of a Risk Management System ::: ------------------------------------------------------------------------------------------- • Risk Assessment -- estimation & evaluation of risk • Risk Confrontation -- determination of acceptable levels of risk in societal context • Risk Intervention -- actions to control risk • Risk communication -- interactive process of exchanging risk information • Risk Management Evaluation -- measurement of effectiveness of aforementioned activities ------------------------------------------------------------------------------------------- ::: Conclusions =========== Determining the optimal initial dosage regimen (prescribability) and maintaining safety and efficacy outcomes when the regimen is changed in some way (switchability) demand careful decision making in the application of equivalence approaches. These approaches must be applied differently during the three phases of the drug development process, and the knowledge-base that is derived from this process must be transferred to and utilized by physicians and pharmacists to assist them in prescribing and dispensing medicines to patients. Consistent and appropriate management of equivalence approaches supports good assessment, management, and communication about risks associated with a therapeutic product, as expressed in product labeling, as well as in specifications and standards that control the quality of a therapeutic product in the marketplace.

PubMed Central

2024-06-05T03:55:47.294167

2004-8-16

PMC515177

Background ========== The isolation of unique DNA probes from the sub-telomeric regions of all chromosomes has opened up a field of cytogenetics research that was previously inaccessible to conventional karyotyping protocols \[[@B1]\]. Since then a number of studies have shown that cryptic structural abnormalities (deletions, duplications etc.) in the subtelomeric regions are relatively commonly found in groups of individuals with idiopathic mental retardation (UK; learning disability; LD). The biological attributes of these chromosomal regions may explain this interesting link. The frequency of meiotic recombination is at its highest at the ends of chromosomes (recently confirmed in the Icelandic microsatellite map of the human genome \[[@B2]\]). Therefore errors in this process should randomly result in a greater frequency of unbalanced chromosome rearrangement products at telomeres. There also appears to be a greater density of genes at the ends of some chromosomes, especially those with non-staining R-bands. Thus, any telomeric copy number change is likely to affect several genes; potentially resulting in clinical features typical of a contiguous gene syndrome -- dysmorphisms, developmental delay and mental retardation. A number of reports have now shown that 0.5%--23% of idiopathic mental retardation cases are associated with cryptic translocations in the vicinity of chromosome telomere (see \[[@B3]-[@B17]\] and \[[@B18]\] for a recent review). FISH, using a commercially available set of subtelomeric probes is the most commonly used screening technique \[[@B19],[@B20]\]. Variations on the theme of FISH (e.g. SKY and CGH) have also been employed. More recently, methods that rely on the detection of copy number changes at subtelomeric loci have been described. MAPH \[[@B21]-[@B24]\] is one such technique in which probes are representatively amplified by the polymerase chain reaction following hybridisation to a patient\'s genomic DNA sample to generate a quantitative profile of subtelomeric sequence copy number. Psychiatric disorders such as schizophrenia (SCZ) and bipolar affective disorder (BPAD) are relatively common in the general population and there is much evidence for a genetic component to susceptibility (for a review see \[[@B25]\]). However, it is clear from the lack of consistent findings from linkage mapping and association studies that they are likely to be complex and aetiologically heterogeneous disorders. For example, several genes might act simultaneously (oligogenic action) or interact (epistasis) to produce the clinical phenotype in any individual, and those genes might be different in different individuals (locus heterogeneity). An alternative to cohort based linkage and association approaches uses cytogenetic abnormalities as direct pointers to candidate gene loci and this has been successfully applied to patients with psychiatric disorders resulting in the identification of a number of candidate susceptibility genes including *DISC1/DISC2*\[[@B26]\], *DIBD1*\[[@B27]\] and *GRIA3*\[[@B28]\]. The chromosome abnormalities that disrupted these genes were reciprocal translocations visible by standard cytogenetic methods. The risk of schizophrenia and affective disorders in patients with idiopathic mild mental retardation is significantly raised and it is well established that schizophrenia is three times more common in this group than the general population and that there is a strong familial element \[[@B29]\]. Both bipolar illness and major depressive disorder have also been described as of increased prevalence in the population with mild mental retardation. The study also revealed a previously undetected complex re-arrangement between chromosomes 2 and 11, and a case of trisomy X, but did not address subtelomeric changes. It strongly suggested however that the co-association between mental retardation and schizophrenia is highly familial with greater rates of both schizophrenia and co-morbid schizophrenia/mental retardation occurring in the families of co-morbid probands compared to families of probands with schizophrenia alone or with mental retardation alone. Limbic system (amygdalo-hippocampal) neuropathology is especially pronounced in this group \[[@B30]\]. We have formed the hypothesis that patients who are co-morbid for severe psychiatric illness and mental retardation may be homogenous in their pathophysiology and that, in addition to large-scale structural chromosomal abnormalities, they may harbour as yet undetected cryptic telomeric changes. To test this we have screened a series of 69 patients co-morbid for mental retardation and psychiatric illness using fluorescence in situ hybridisation (FISH) and multiplex amplifiable probe hybridisation (MAPH). Methods ======= Patient Cohort -------------- Local research ethics permission was obtained for this study. The patients were initially ascertained through computerised psychiatric clinical case-registers that allowed us to identify adults with dual diagnosis of psychosis and mental retardation. A specific psychiatry service exists in Scotland to meet the needs of patients with mental retardation who also suffer from psychiatric disorders and initial clinical diagnoses were confirmed by consultation between the relevant specialist clinician involved and the research team member who is also a specialist in the psychiatry of mental retardation (WM). Confirmation that IQ fell within the mild range of mental retardation was obtained from case records. 69 patients with mild mental retardation (IQ 70 to around 50) and a referral diagnosis of co-existing schizophrenia or major affective disorder were studied. Parental samples were not available in many cases due to the age of the probands. This cohort is a subset of 74 originally ascertained subjects: 5 were removed because of aneuploidy (2 cases of 47(XXX)) or after more thorough psychiatric evaluation. One discounted subject with only mental retardation possessed a 6q subtelomeric deletion as determined by several MAPH probes (data not shown). The lifetime version of the Schedule for Affective Disorders and Schizophrenia (SADS-L \[[@B31]\]) along with extensive case record review, and interviews with key carers and relatives was used to gather the information needed to make a diagnosis of schizophrenia or affective disorder according to the Diagnostic and Statistical Manual 4^th^Edition (DSM-IV \[[@B32]\]). Diagnosis was finalised by consensus between two experienced psychiatrists (DB, WM) one of whom specialises in the psychiatry of mental retardation (WM). SADS-L has previously been successfully used in people with mild mental retardation \[[@B29]\] to establish psychiatric diagnoses. Overall 49 subjects met the DSM-IV criteria for definite schizophrenia, 3 for schizoaffective disorder, 11 for Bipolar I Disorder, 1 for recurrent Major Depressive Disorder (unipolar depression). In addition, 5 subjects were diagnosed as having a unspecified functional psychosis (DSM-IV 298.9, Psychotic disorder NOS). None had co-existing Down Syndrome or Fragile X disorder. A breakdown of the patients into their clinical categories and methodology of screening is presented in table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Subject classification and analysis Breakdown of subjects into their diagnostic categories and applied experimental methodology. MR; mental retardation, SCZ; schizophrenia, BP1; bipolar affective disorder I, SCAFF; schizoaffective disorder, UFP; unspecified functional psychosis, UPR; unipolar depression. ::: **Clinical category** **MAPH alone** **MAPH and FISH** **FISH alone** ----------------------- ---------------- ------------------- ---------------- MR/SCZ 34 7 8 MR/BP1 6 3 2 MR/SCAFF 1 2 0 MR/UFP 3 0 2 MR/UPR 1 0 0 TOTAL SCREEN 45 12 12 ::: DNA extraction -------------- DNA was extracted from venous blood samples (10 mls) of all patients by standard methods using Nucleon BACC2 kits (Nucleon Biosciences). 1 mg/ml dilutions were prepared for MAPH. MAPH ---- All 57 MAPH samples were tested in triplicate using the subtelomeric screening set described previously \[[@B23]\]. All samples were anonymised prior to MAPH analysis. Each sample was tested three times, and putative positives identified by a univariate method (standard hypothesis testing against a normal distribution) and multivariate methods employed by the software SYSTAT 8.0 (Bivariate scattergraphs and Hadi outlier analysis). Four putative positives were divided into one confident (univariate analysis, p \< 0.01, corrected for multiple observations) and three possible (univariate analysis, p \< 0.05, corrected for multiple observations, Hadi outlier distance \>4, all three results reporting a consistent change: either all \>1.0 or \<1.0) positive results. The three \"possible\" positives have since been discounted since they involved gain of the 20p telomeric probe ST18E1, which from experience with normal control subjects has shown to have unacceptably high measurement error. This probe has been replaced in more recent formulations of the subtelomeric probe set. Subtelomeric FISH ----------------- Blood samples from a subset of the patients were cultured in Peripheral Blood Medium (Sigma) for 72 hours. After colcemid treatment for one hour, lymphocytes were lysed and fixed in methanol:acetic acid (3:1). Fixed metaphase material was dropped onto microscope slides. Each slide was hybridised by three fluorescently labelled probe mixes (ToTelVysion, Vysis Inc.) under separate coverslips. All 15 mixes covering every subtelomeric region for a patient could thus be analysed on 5 slides. Images were captured on a Zeiss Axioskop2 microscope coupled to a Macintosh G4 computer running SmartCapture2.1 software (Digital Scientific). Five metaphases were scored for each probe mix. Results ======= A subtelomeric deletion identified in one subject ------------------------------------------------- Complete accord was seen in the 12 instances where both screening methodologies were used. The FISH approach did not detect any subtelomeric abnormalities (including balanced translocations, which would not be observable by MAPH). However, MAPH identified a subject with a loss of one copy of the 4q subtelomeric region (p \< 1 × 10^-3^, corrected for multiple observations). It is unlikely that this copy number change (4q^-^) defined by MAPH represents an irrelevant polymorphism; no similar change was found on analysis of 83 unrelated control individuals, giving an upper (95% confidence) limit of 1.6% for the frequency of this variant. Precise definition of 4q loss ----------------------------- Additional MAPH probes were designed to determine the extent of the 4q deletion (fig. [1](#F1){ref-type="fig"}). The results show that the proximal boundary of the subtelomeric deletion is between the *FAT*gene and the proximal end of clone 713c19 (Genbank accession number AC108073). P values for boundary probes (Ho, value = 1.00), p \< 5 × 10^-5^for deletion and p \> 0.05 for normal dosage. This 4q deletion encompasses a region of annotated genomic DNA of approximately 3 Mb. The transcript map of this region is not yet completely defined (see Table [2](#T2){ref-type="table"}) but contains at least 10 transcriptional units with varying levels of authenticity/experimental evidence attached to each one. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Subtelomeric region of chromosome 4q**Only annotated, unique chromosome sequence is shown, derived from the November 2002 version of the UCSC Human Genome assembly (subtelomeric repeats would extend to the right of the diagram). A scale bar and the gene content (see Table 2) of the region are shown. The positions of the MAPH markers are also shown which allowed the maximum and minimum extents of the deletion to be defined (black bar). Above the chromosome region is shown the result of duplicate analysis from each MAPH probe (mean +/- 95% CI), together with the 3 standard deviation threshold and the results from the other control probes (mean +/- 95% CI). :::  ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Gene content of 4q deletion Genes/putative transcriptional units within the deleted region on chromosomes 4q. ESTs a-f are represented in figure 1. An attempt to gauge the approximate expression levels of each gene was based on the number of EST clones present in the UCSC Human Genome Browser (Nov.2002/Apr.2003 releases). A brief summary of gene function and a representative accession number, where informative, is also included. TUBB4Q (4q35) is omitted from this list because it is a confirmed pseudogene. ::: **4q35 GENE** **EST exp.** **Function/comments** ------------------ -------------- ----------------------------------------------------------------------------------------------------------------- *FAT* ++++ Cadherin-related tumor suppressor homologue precursor *EST a* \+ (BE856720) Novel. *EST b* \+ (BM806339) Novel. Contains 5 1/2 copies of 34aa repeat motif *EST c* ++ (AI917275) Novel. No obvious ORF *ZFP42/FLJ32157* \+ (AK056719) Similar to transcriptional repressor protein YY1 *FLJ25801* \+ (AK098667) Protein contains SMC (chromosome segregation ATPase) domain and PRY/SPRY domains (unknown function). *EST d* +++ (BU571187) Novel. *EST e* \+ (BC033535) Novel. *EST f* \+ (BC029568) LOC256307 novel predicted gene *FRG1* ++++ Facioscapulohumeral muscular dystrophy region gene 1 *DUX4* \+ Homeobox protein, multiple copies. ::: Discussion ========== Unlike simple chromosomal translocations, large deletions associated with a certain condition can present many candidate genes for further study. In the context of psychiatric disorders velo-cardio-facial syndrome (VCFS or del22q11 syndrome) offers a model of how a cryptic deletion associated with schizophrenia has highlighted several candidate genes for future study \[[@B33]-[@B35]\]. We used two methods to screen for subtelomeric changes in our cohort of patients. The MAPH technique has proved a fast and accurate method for determining copy number changes in the human genome and represents a cost-effective route to the screening of large numbers of patients. In addition, the disorders that can be studied by this approach are limited only by the design of suitable primers so that screening for both single and contiguous gene disorders is feasible. MAPH has also been proved to be a simple method to map deletion breakpoints with greater resolution than FISH. The subtelomeric FISH approach is a more demanding approach because of the patient sample preparation, the requirement for specialised microscopy equipment, the cost of the commercial probe sets and the labour involved. Nevertheless, FISH has some advantages: the FISH approach is the only way to detect balanced chromosome rearrangements such as inversions and translocations. Other studies of subtelomeric regions in mental retardation subjects have identified such chromosomal aberrations. In addition, FISH has been a vital technique for identifying disrupted genes such as *DISC1*in psychiatric patients with (non-subtelomeric) chromosomal rearrangements. Therefore, the selection of the screening technique should be determined by the number of cases to be studied and the nature of the abnormalities expected. We have identified a subtelomeric deletion within our cohort of 69 patients using the MAPH and FISH methodologies. This 4q deletion is associated with a co-morbid phenotype of schizoaffective disorder and mild mental retardation (fullscale IQ between 60 and 70). Consent was not forthcoming to determine whether the deletion was of parental or *de novo*origin. However, psychiatric illness has not been diagnosed in other members of the family. We cannot, therefore, formally link the presence of the deletion with mental retardation and/or psychiatric illness in the patient. The annotation of transcripts at 4q35.2 is currently an active area of research (see fig. [1](#F1){ref-type="fig"} and table [2](#T2){ref-type="table"}) because of good linkage evidence (LOD score of 3.2 for microsatellite marker D4S1652) from an extended Australian kindred multiply affected with bipolar affective disorder \[[@B36]-[@B38]\]. Importantly, the principal linkage region almost exactly matches the deletion interval observed in our patient. The 4q deletion patient has been diagnosed with schizoaffective disorder (DSM-IV 295.7) -- with periods of psychotic depression but also mood incongruent hallucinations and delusions. This is in contrast to the clear bipolar affective disorder diagnosed for members of the described linkage family. However, it has been repeatedly observed that schizoaffective disorders and bipolar affective disorders overlap clinically and are indeed often difficult to separate. One postulated explanation for the now frequently reported linkage overlaps between bipolar illness, schizoaffective disorders and schizophrenia is that the inherited susceptibility is for psychosis rather than a specific disorder (reviewed in \[[@B39]\]). The subtelomeric region of 4q is also interesting because it contains a candidate gene, *FRG1*, for facioscapulohumeral muscular dystrophy. The 4q patient does not show any of the typical features of this disorder but this can be explained by the fact that copy number does not appear to be critical for the onset of the disorder \[[@B40]\]. Rather, the proximity of the gene to a variable number telomeric repeat sequence (D4Z4) seems to be the chief determinant of pathology \[[@B41]\]. Of the 69 patients with clear co-morbidity, one (1.4%) possessed a single copy subtelomeric deletion. This frequency is in line with those from studies of individuals with mental retardation alone \[[@B18]\]. As more studies examine chromosomal integrity in people with mental retardation or other conditions we hope that replication of subtelomeric abnormalities will be observed, perhaps leading to the eventual clinical definition of range of \'subtelomeric syndromes\' such as that recently described for the subtelomeric deletion of 1q \[[@B42]\]. Conclusions =========== The identification of the precisely delimited 4q deletion may contribute to the mapping of the susceptibility gene for psychiatric illness at this locus. The finding of bipolar affective disorder linkage to this region suggests that, in this case at least, the schizoaffective and mental retardation components to the co-morbid phenotype may be discrete and genetically separable in the manner of other contiguous gene disorders. If it is assumed that either component of the clinical phenotype is caused by haploinsufficiency then examining the comparative expression levels of candidate genes in normal, bipolar-linked and 4q deleted lymphoblastoid cell lines might provide a quick route to the identification of causative genes. Alternatively, gene association studies may be required to identify the candidate psychiatric illness gene at 4q35. The high rate of recombination in subtelomeric DNA means that a higher density of genetic markers will be required to establish linkage or association reliably; conversely, once detected, the high rate of recombination will allow high resolution fine mapping of significant loci. Competing interests =================== None declared. Authors\' contributions ======================= BSP participated in the design of the study and the FISH analysis and drafted the manuscript. EJH and JALA carried out all MAPH assays and associated statistical analyses. MPM carried out blood culture and FISH analysis. WJM conceived the study and was responsible for the generation of all clinical data, and DHRB and DJP participated in the study design and coordination. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2350/5/21/prepub> Acknowledgements ================ The authors wish to thank Judy Fantes, Michael Ellis and Paul Perry for advice on fluorescence microscopy, and Jane Hewitt and Dan Bolland for advice on the 4q physical map. The subtelomeric FISH work was supported by a small project grant from SHERT (RG45/01) and the MAPH work by the Wellcome Trust, grant number 060578. MPM was supported by a grant from the Scottish Executive, Chief Scientist\'s Office and Wellcome Trust Genes to Cognition programme.

PubMed Central

2024-06-05T03:55:47.296931

2004-8-13

PMC515178

Background ========== Hypertensive disorders in pregnancy complicate approximately 10--16% of pregnancies and are leading causes of maternal, fetal and neonatal morbidity and mortality worldwide \[[@B1]-[@B3]\]. Definitions, classifications, assessment and management of hypertensive disorders vary considerably in the literature and from country to country \[[@B4]\]: thus, it is difficult to compare results from different studies. Past studies have attempted to quantify the effect of hypertension on adverse perinatal outcomes. To date, the majority of study designs have included retrospective and prospective cohort studies \[[@B5]-[@B9],[@B12]-[@B20]\], as well as randomized-controlled trials that assessed the impact of antihypertensive medication on maternal and perinatal outcomes\[[@B10],[@B11]\]. For the most part, these studies have been concentrated in tertiary referral centres, and suggest that hypertension in pregnancy leads to an increased risk of small for gestational age (SGA) and preterm birth. We, therefore, carried out a population-based study to quantify the frequency of hypertensive disorders in pregnancy and also the excess risk of SGA and stillbirth that is associated with this pregnancy complication. We investigated the way in which SGA and stillbirth were modified by other factors that also have a serious influence on SGA and stillbirth, for example, whether a twin gestation modifies the effect of hypertension on SGA, or whether smoking modifies the effect of hypertension on stillbirth. Methods ======= Population ---------- The study population included all births to residents of the province of Nova Scotia, Canada between 1988 and 2000. Information on these births was obtained from the Nova Scotia Atlee Perinatal Database. The Nova Scotia Atlee Perinatal Database includes several hundred variables containing maternal and newborn information, such as demographic variables, procedures, interventions, maternal and newborn diagnoses and morbidity and mortality information for every pregnancy and birth occurring in Nova Scotia hospitals and to Nova Scotia residents since 1988. Pregnancies were excluded if delivery occurred \< 20 weeks, if birthweight was \< 500 g, if there was a higher order pregnancy (greater than twin gestation), or a known major fetal anomaly. Information in the database on the type of hypertensive disorder is directly coded from medical charts. The diagnosis of hypertensive disorders in pregnancy was made if it occurred in the antepartum or postpartum period. Mild pregnancy induced hypertension (PIH) included physician-diagnosed mild pregnancy induced hypertension if in the chart, transient hypertension or a diastolic blood pressure exceeding 90 mmHg on two or more occasions in a 24-hour period. Severe pregnancy induced hypertension included physician-diagnosed severe pregnancy induced hypertension, diastolic blood pressure ≥ 110 mm Hg on at least two occasions within a 6-hour period, if magnesium sulfate was administered for seizure prophylaxis, if there was ≥ 2+ proteinuria, low platelets (\<100,000), and/or elevated liver enzymes (ALT \> 35 u/L, AST \> 30 u/L and/or LDH \> 670 u/L). HELLP syndrome (hemolysis, elevated liver enzymes, low platelets) included physician-diagnosed HELLP. Eclampsia included physician-diagnosed eclampsia or one or more convulsions not attributable to other cerebral conditions such as epilepsy or cerebral hemorrhage in a patient with hypertension. Chronic hypertensive disease included a history of hypertensive disease when not pregnant, prior to current pregnancy or prior to 20 weeks of the current pregnancy. Pregnancy induced hypertension superimposed on chronic hypertension included physician-diagnosed pregnancy induced hypertension superimposed on chronic hypertension or if there was hemolysis, elevated liver enzymes or low platelets. We first examined the maternal characteristics of study subjects with hypertension in pregnancy. For this analysis, hypertensive disorders in pregnancy were defined according to Nova Scotia Atlee Perinatal Database definitions: mild PIH, severe PIH, HELLP, eclampsia, chronic hypertension and chronic hypertension with superimposed PIH. Maternal characteristics which were considered included age, marital status, parity, prepregnancy weight, pregnancy weight gain, administration of antenatal steroids, smoking, drug abuse, the presence of anemia (Hgb \< 10 gm%), gestational diabetes (two abnormal values on a glucose tolerance test in pregnancy or if insulin was administered for the first time in pregnancy), pre-existing diabetes, and twin gestation. For the multivariate analyses, these database definitions were grouped to more closely approximate commonly used definitions such as those proposed by the Canadian Hypertensive Society and other international organizations \[[@B4]\]. These groups were defined as gestational hypertension without proteinuria (including the database entity mild pregnancy induced hypertension), gestational hypertension with proteinuria (including the database entities severe pregnancy induced hypertension, HELLP, and eclampsia) and pre-existing hypertension (including the database entities chronic hypertension and chronic hypertension with superimposed pregnancy induced hypertension). Only live births were considered in the analysis of small for gestational age, while all births were considered for the stillbirth analyses. Small for gestational age was defined as birthweight for gestational age that was less than the sex-specific 10^th^percentile cut-off of a recently published Canadian fetal growth reference \[[@B21]\]. Stillbirth was defined as fetal death before birth, with gestational age ≥ 20 weeks and birthweight ≥ 500 grams. Ethical approval for the study was obtained from the Research Ethics Boards at Dalhousie University in Halifax, Nova Scotia, the Reproductive Care Program of Nova Scotia and the IWK Health Centre in Halifax, Nova Scotia. Statistical analysis -------------------- Exact binomial 95% confidence intervals were calculated for rates of hypertensive disorders in pregnancy. Descriptive analyses were carried out on maternal data to ascertain the association between maternal characteristics and hypertensive disorders in pregnancy. Categorical data between hypertensive and normotensive pregnancies were compared using chi-square and Fisher\'s exact tests, where appropriate. Logistic regression analyses were carried out to determine the adjusted odds ratios (OR) and 95% confidence intervals (CI) expressing the relationship between any hypertensive disorder in pregnancy and groups of hypertension (i.e., gestational hypertension without proteinuria, gestational hypertension with proteinuria, pre-existing hypertension) and the two dichotomous primary outcomes (SGA and stillbirth). Backward stepwise elimination of variables was carried out to identify all significant determinants of the outcome. Modification of the effect of hypertensive disorders on SGA or stillbirth was investigated, based on clinical understanding. Variables considered to be in the causal pathway between the determinant and the outcome (e.g., SGA in the analysis of stillbirth) were not adjusted for in the model \[[@B22]-[@B24]\]. When the outcome rate was low we assumed the odds ratio was equal to the relative risk, but not in situations when the outcome rate was high (\>10 percent). The significance level selected was *P*\< .05 and all tests were two-tailed. Statistical analyses were performed using the SAS programming package for Windows (Version 8.0). Results ======= Frequency of hypertensive disease --------------------------------- The study population included 135,466 pregnancies. Of these, 7.7% (95% CI 7.6,7.9) had mild PIH, 1.3% (95% CI 1.3,1.4) had severe PIH, 0.2% (95% CI 0.1,0.2) had HELLP, 0.02% (95% CI 0.02,0.03) had eclampsia, 0.6% (95% CI 0.5,0.6) had chronic hypertension and 0.4% (95% CI 0.3,0.4) had chronic hypertension with superimposed PIH. The overall rate of hypertensive disease in pregnancy was 10.1% (95% CI 10.0,10.3). Maternal characteristics ------------------------ Table [1](#T1){ref-type="table"} summarizes the characteristics of women with hypertensive disorders in pregnancy from 1988--2000. Women with severe PIH were less likely to be married (63.5%), while women with chronic hypertension (with or without superimposed PIH) were more likely to be married (83.2% and 77.3%, respectively) compared with normotensive women (69.6%). Women with hypertensive disorders were more likely to be nulliparous (range 42.2% to 78.2%) as compared with normotensive women (41.9%). Women with hypertensive disorders in pregnancy had higher pre-pregnancy weight (range 66.8 kg to 82.0 kg) compared with normotensive women (64.4 kg). While women with PIH (mild or severe) had on average a greater weight gain in pregnancy (16.2 kg and 16.5 kg, respectively), women with chronic hypertension (with or without superimposed PIH) had a lower weight gain (12.2 kg and 13.2 kg, respectively) compared with normotensive women (14.4 kg). A smaller proportion of women with hypertensive disorders in pregnancy smoked (range 13.0% to 22.7%) as compared with normotensive women (30.7%). Women with hypertensive disorders in pregnancy were more likely to have gestational diabetes (range 4.6% to 9.1%) and pre-existing diabetes (range 0.8% to 2.5%) compared with normotensive women (2.3% and 0.3%, respectively). Twin pregnancies were more likely to be complicated by hypertensive disorders (range 2.3% to 6.4%) as compared with normotensive pregnancies (1.0%). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Characteristics of women with and without hypertensive disorders in pregnancy, Nova Scotia, 1988--2000. ::: ------------------------------------------------------------------------------------------------------------------------------------------------------ Variable Normotensive\ Mild PIH\ Severe PIH\ HELLP\ Eclampsia\ Chronic HTN\ Chronic HTN\ n = 121,760 n = 10,460 n = 1,770 n = 202 n = 32 n = 767 +PIH n = 475 ------------------------------------- --------------- ---------------- ---------------- --------------- -------------- --------------- --------------- Mean age in years (SD) 27.8 (5.3) 27.5 (5.4)\* 26.9 (5.7)\* 28.9 (5.5)\* 25.6 (6.8)\* 31.1 (5.0)\* 30.5 (5.1)\* Married (%) 84,739 (69.6) 7,530 (72.0) 1,124 (63.5)\* 145 (71.8) 18 (56.3) 638 (83.2)\* 367 (77.3)\* Nulliparous (%) 51,057 (41.9) 6,760 (64.6)\* 1,328 (75.0)\* 158 (78.2)\* 20 (62.5)\* 324 (42.2)\* 259 (54.5)\* Mean prepregnancy weight in kg (SD) 64.4 (13.8) 70.6 (16.7)\* 67.6 (15.6)\* 66.8 (16.0)\* 68.2 (16.9) 82.0 (20.9)\* 81.3 (20.9)\* Mean weight gain in kg (SD) 14.4 (6.0) 16.2 (6.9)\* 16.5 (6.9)\* 14.8 (7.0) 16.2 (6.2) 12.2 (6.7)\* 13.2 (7.3)\* Smokes any cigarettes (%) 35,846 (30.7) 2,157 (21.5)\* 376 (22.7)\* 31 (16.0)\* 6 (21.4) 150 (20.4)\* 59 (13.0)\* Anemia (%) 3,441 (2.8) 242 (2.3)\* 78 (4.3)\* 15 (7.4)\* 0 (0) 31 (4.1) 20 (4.2) Gestational Diabetes (%) 2,749 (2.3) 485 (4.6)\* 84 (4.8)\* 3 (1.5) 2 (6.3) 70 (9.1)\* 34 (7.2)\* Preexisting Diabetes (%) 331 (0.3) 85 (0.8)\* 29 (1.6)\* 4 (2.0)\* 0 (0) 19 (2.5)\* 9 (1.9)\* Anti-hypertensive medication (%) 0 (0) 107 (1.0) 97 (5.5) 23 (11.4) 5 (15.6) 146 (19.0) 122 (25.7) Twins (%) 1,231 (1.0) 244 (2.3)\* 60 (3.4)\* 13 (6.4)\* 0 (0) 7 (0.9) 13 (2.7)\* ------------------------------------------------------------------------------------------------------------------------------------------------------ PIH denotes pregnancy induced hypertension, HELLP denotes hemolysis, elevated liver enzymes, low platelets syndrome, HTN denotes hypertension. \* Denotes hypertension categories significantly different from the normotensive category. ::: Small for gestational age ------------------------- Table [2](#T2){ref-type="table"} summarizes the crude relationships between hypertension in pregnancy and SGA. There was an increased risk of SGA among infants born to women with any hypertensive disorder (RR 1.6, 95% CI 1.5,1.6) compared with infants born to women with normotensive pregnancies. There was an increased risk of SGA among infants born to women with mild PIH (RR 1.3, 95% CI 1.3,1.4), severe PIH (RR 2.5, 95% CI 2.3,2.8), HELLP (RR 3.8, 95% CI 3.2,4.5), eclampsia (RR 3.5, 95% CI 2.2,5.7), chronic hypertension (RR 1.4, 95% CI 1.1,1.6) and chronic hypertension with PIH (RR 2.2, 95% CI 1.8,2.6) compared with infants born to women with normotensive pregnancies. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Comparison of small for gestational age (SGA, \<10^th^percentile) rates among live births to hypertensive vs. normotensive women, Nova Scotia, 1988--2000. ::: Total No. SGA Relative Risk 95% CI *P*value ------------------------------- ----------- -------- --------------- -------- ---------- --------- Normotensive women 122,394 12,032 9.8 1.0 \- \- All hypertensive women 13,940 2,131 15.3 1.6 1.5,1.6 \<0.001  Mild PIH 10,639 1,384 13.0 1.3 1.3,1.4 \<0.001  Severe PIH 1,814 453 25.0 2.5 2.3,2.8 \<0.001  HELLP 212 79 37.3 3.8 3.2,4.5 \<0.001  Eclampsia 32 11 34.4 3.5 2.2,5.7 \<0.001  Chronic Hypertension 766 102 13.3 1.4 1.1,1.6 0.002  Chronic hypertension and PIH 477 102 21.4 2.2 1.8,2.6 \<0.001 PIH denotes pregnancy induced hypertension, HELLP denotes hemolysis, elevated liver enzymes, low platelets syndrome, HTN denotes hypertension, CI denotes Confidence interval. ::: After controlling for potential confounders, women with any hypertensive disorder were 1.8 (95% CI 1.7,1.9, *P*\< .001) times more likely to have a live birth with SGA as compared with normotensive women. Women with gestational hypertension without proteinuria were 1.5 (95% CI 1.4,1.6, *P*\< .001) times more likely to have a live birth with SGA as compared with normotensive women. Similarly, women with gestational hypertension with proteinuria were 3.3 (95% CI 3.0,3.9, *P*\< .001) times more likely and women with pre-existing hypertension were 2.5 (95% CI 2.1,2.9, *P*\< .001) times more likely to have a live birth with SGA as compared with normotensive women. Stillbirth ---------- Table [3](#T3){ref-type="table"} summarizes the crude relationships between hypertension in pregnancy and stillbirth. There was an increased risk of stillbirth among women with any hypertensive disorder (RR 1.4, 95% CI 1.1,1.8) and among women with pregnancies complicated by chronic hypertension (RR 2.4, 95% CI 1.2,5.1) or chronic hypertension with superimposed PIH (RR 4.4, 95% CI 2.2,8.8), compared with normotensive pregnancies. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Comparison of stillbirth rates among all births to hypertensive vs. normotensive women, Nova Scotia, 1988--2000. ::: Total No. Stillbirths Relative Risk 95% CI *P*value ------------------------------- ----------- ------------- --------------- -------- ---------- --------- Normotensive women 122,855 461 0.4 1.0 \- \- All hypertensive women 14,013 73 0.52 1.4 1.1,1.8 0.01  Mild PIH 10,683 44 0.4 1.1 0.8,1.5 0.55  Severe PIH 1,826 12 0.7 1.8 1.0,3.1 0.50  HELLP 214 2 0.9 2.5 0.6,9.9 0.19  Eclampsia 32 0 0.0 \- 0.0,41.0 1.00  Chronic Hypertension 773 7 0.9 2.4 1.2,5.1 0.03  Chronic hypertension and PIH 485 8 1.7 4.4 2.2,8.8 \<0.001 PIH denotes pregnancy induced hypertension, HELLP denotes hemolysis, elevated liver enzymes, low platelets syndrome, HTN denotes hypertension, CI denotes Confidence interval. ::: After controlling for potential confounders, women with any hypertensive disorder were 1.4 (95% CI 1.1,1.8, *P*= .02) times more likely to have a stillbirth as compared with normotensive women (Table [4](#T4){ref-type="table"}). Women with pre-existing hypertension were 3.2 (95% CI 1.9,5.4, *P*\< .001) times more likely to have a stillbirth as compared with normotensive women. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Effect of hypertensive disorders in pregnancy on small for gestational age (\< 10^th^percentile) and stillbirth, Nova Scotia, 1988--2000. ::: Crude Adjusted ----------------------------------------------- ------- ---------- --------- ----- --------- --------- [Small for gestational age\*]{.underline} Normotensive women 1.0 \- \- 1.0 \- \- Hypertensive women (any type) 1.6 1.5,1.6 \<0.001 1.8 1.7,1.9 \<0.001  Gestational hypertension without proteinuria 1.3 1.3,1.4 \<0.001 1.5 1.4,1.6 \<0.001  Gestational hypertension withproteinuria 2.7 2.5,2.9 \<0.001 3.3 3.0,3.9 \<0.001  Pre-existing hypertension 1.7 1.5,1.9 \<0.001 2.5 2.1,2.9 \<0.001 [Stillbirth\*\*]{.underline} Normotensive women 1.0 \- \- 1.0 \- \- Hypertensive women (any type) 1.4 1.1,1.8 0.01 1.4 1.1,1.8 0.02  Gestational hypertension without proteinuria 1.1 0.8,1.5 0.55 1.1 0.8,1.5 0.60  Gestational hypertension with proteinuria 1.8 1.0,3.1 0.03 1.6 0.9,2.9 0.08  Pre-existing hypertension 3.2 1.9,5.3 \<0.001 3.2 1.9,5.4 \<0.001 CI denotes Confidence interval. \* Adjusted for smoking, maternal age, gestational diabetes, pre-existing diabetes, maternal anemia, nulliparity, marital status, drug abuse, prepregnancy weight, weight gain, antenatal steroids, twins and infant sex. \*\* Adjusted for smoking, maternal autoantibodies, maternal age, pre-existing diabetes, maternal anemia,prepregnancy weight and twins. ::: Twin pregnancy modified the effect of gestational hypertension with or without proteinuria on small for gestational age (Table [5](#T5){ref-type="table"}). A woman with a singleton pregnancy who had gestational hypertension without proteinuria had a 1.5 fold increase in risk of SGA compared with a woman with a normotensive singleton pregnancy. A woman with a twin pregnancy had a 4.7 fold increase in risk of SGA, but the woman with gestational hypertension without proteinuria and a twin pregnancy had a less than expected increase in risk of SGA (1.5 × 4.7 × 0.7 = 4.9 as opposed to 1.5 × 4.7 = 7.1) i.e., the combined effect of a twin pregnancy and gestational hypertension without proteinuria was less than what would be expected under a multiplicative (logistic) model. Similarly, a woman with gestational hypertension with proteinuria and a twin pregnancy had a less than expected increase in risk of SGA (5.2 compared to 17.4). A woman with pre-existing hypertension who smoked had a less than expected increase in risk of SGA (4.4 compared to 7.3). A woman with gestational hypertension without proteinuria and who smoked had a greater than expected increase in risk of stillbirth (2.8 compared to 1.1). ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Effect of hypertensive disorders in pregnancy on small for gestational age (\< 10^th^percentile) and stillbirth, with modeling of effect-modification by twin pregnancy and maternal smoking, Nova Scotia, 1988--2000. ::: Adjusted --------------------------------------------------------- ---------- --------- --------- [Small for gestational age]{.underline} Normotensive women 1.0 \- \- Gestational hypertension without proteinuria 1.5 1.4,1.6 \<0.001 Gestational hypertension with proteinuria 3.7 3.3,4.1 \<0.001 Pre-existing hypertension 2.5 2.1,3.0 \<0.001 Twins 4.7 4.3,5.2 \<0.001  Gestational hypertension without proteinuria × twins 0.7 0.6,0.9 0.01  Gestational hypertension with proteinuria × twins 0.3 0.2,0.4 \<0.001  Pre-existing hypertension × twins 0.7 0.3,1.4 0.29 Smoking 2.9 2.8,3.0 \<0.001  Gestational hypertension without proteinuria × smoking 1.0 0.9,1.2 0.96  Gestational hypertension with proteinuria × smoking 0.9 0.7,1.1 0.28  Pre-existing hypertension × smoking 0.6 0.4,0.8 0.002 [Stillbirth]{.underline} Normotensive women 1.0 \- \- Gestational hypertension without proteinuria 0.8 0.5,1.2 0.24 Gestational hypertension with proteinuria 1.4 0.6,2.9 0.42 Pre-existing hypertension 3.5 2.0,6.2 \<0.001 Smoking 1.4 1.2,1.7 \<0.001  Gestational hypertension without proteinuria × smoking 2.5 1.3,4.7 0.006  Gestational hypertension with proteinuria × smoking 1.8 0.4,5.3 0.33  Pre-existing hypertension × smoking 0.6 0.1,2.5 0.43 CI denotes Confidence interval.\*\* Adjusted for smoking, maternal age, gestational diabetes, pre-existing diabetes,maternal anemia, nulliparity, marital status, drug abuse, prepregnancy weight,weight gain, antenatal steroids, twins and infant sex. \*\* Adjusted for smoking, maternal autoantibodies, maternal age, pre-existing diabetes, maternal anemia, prepregnancy weight and twins. ::: Discussion ========== This large population-based cohort study examined the magnitude of the risks of small for gestational age and stillbirth in women with hypertension in pregnancy. Hypertensive disorders in pregnancy had a significant effect on rates of SGA after adjusting for potential confounders. Pre-existing hypertension had a significant effect on stillbirth rates after adjusting for potential confounders. Modification of the effect of hypertension on SGA and stillbirth was observed among women who also had a twin pregnancy or were also smokers. The rate of hypertensive disorders in pregnancy in this population (10.1%, 95% CI 10.0,10.3) was similar to that reported in the literature (10--16%). The rate of mild PIH in this population was 7.7%, compared with 6--7% in the literature. The rate of severe PIH (1.3%), HELLP (0.2%) and eclampsia (0.02%) were lower than expected (5--6%). The rate of pre-existing hypertension (0.6%) and pre-existing hypertension with superimposed PIH (0.4%) were lower than expected (3--5% and 0.8--1.3%, respectively) \[[@B25],[@B26]\]. These differences in rates from what is expected from the published literature may be explained by the fact that most previous studies were conducted in high-risk populations in referral hospitals. Differences between our study findings and those from other population-based studies \[[@B12]-[@B20]\] may be due to potential differences in population characteristics, and clinical practice factors including quality of diagnostic information. Our study also identified maternal characteristics previously known to be associated with hypertensive disorders in pregnancy, including nulliparity \[[@B25]\], older age \[[@B27]\], diabetes \[[@B28]\], twin pregnancy \[[@B29]\] and smoking \[[@B30]\]. Small for gestational age is a more appropriate measure of fetal growth than birthweight alone. Our study evaluated birthweight for gestational age and gender among live births to normotensive and hypertensive pregnancies using a recent population-based Canadian reference \[[@B21]\]. Women with any hypertensive disorder, gestational hypertension with or without proteinuria, and pre-existing hypertension were at a significantly higher risk of having a SGA infant (relative risk 1.8, 1.5, 3.3, and 2.5, respectively). The risk of SGA was higher among hypertensive women with a twin pregnancy. Similar patterns in risk for SGA have been seen in other populations \[[@B13],[@B14],[@B16]\]. Women with any hypertensive disorder and pre-existing hypertension were at significantly higher risk of stillbirth compared with women having normotensive pregnancies (RR 1.4 and 3.2, respectively). Similarities in risk for fetal mortality between women with gestational hypertension with or without proteinuria and women who were normotensive in pregnancy have been reported elsewhere \[[@B31]\]. Chance and confounding are unlikely explanations for the results of our study, because of the large study size, and adjustment for potential confounders using logistic regression. This study was not able to correct for the degree of blood pressure control, which may have an effect on fetal growth. While retrospective studies in general are limited by the reliability of data, information in the Nova Scotia Atlee Perinatal Database is of high quality. Routine data checks and edits are made at the time of data collection, and validation \[[@B32]\] and reabstraction studies attest to the quality of the data in this large clinical database. Our study was limited by the definitions for hypertensive disorders in pregnancy used by the Nova Scotia Atlee Perinatal Database, and while these definitions are not exactly the same as commonly used definitions, they approximate definitions proposed by the Canadian Hypertension Society and other organizations \[[@B4]\]. The number of comparisons carried out requires that *P*values associated with the results be interpreted with caution. Conclusions =========== This population-based cohort study demonstrates that women with hypertensive disease in pregnancy are at significantly higher risk of having pregnancies complicated by small for gestational age (\<10^th^percentile) and stillbirth in comparison with women with normotensive pregnancies. Twin gestation and smoking were important modifiers of the effect of hypertension on SGA and stillbirth. This study allowed the quantification of risks of adverse outcomes in women with pregnancies complicated by hypertension, confirming associations in the published literature and allowing appropriate counseling and monitoring in the management of these women, as well as providing baseline risks which may be used in future intervention studies. Competing interests =================== None declared. Authors\' contributions ======================= VMA proposed the study, carried out the preliminary analyses and wrote the paper. All authors discussed the analyses, contributed to the intellectual content of the paper and approved the final version. VMA and KEM provided the maternal-fetal medicine perspective, KSJ provided the general medical and epidemiologic input, LAM provided the general and obstetrical medicine perspective, and AO provided the neonatal-perinatal perspective and epidemiologic input. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2393/4/17/prepub> Acknowledgements ================ We would like to acknowledge the Reproductive Care Program of Nova Scotia for providing access to the data.

PubMed Central

2024-06-05T03:55:47.298960

2004-8-6

PMC515179

Background ========== Retroviruses ------------ Retroviruses are viruses that are found throughout the animal kingdom, including in chickens, mice, cats, sheep, goats, cattle, primates, fish and humans. The first retro viruses were identified as cell free oncogenic factors in chickens. Subsequently, many of the oncogenic retroviruses have been shown to be replication defective forms that have substituted a part of their normal viral gene complement with an oncogene sequence \[[@B1]\]. Replication competent retroviruses also cause malignant disease, as well as a range of other pathogenic states, in a broad range of species. This includes what must be the most significant transmissible disease of humans in recent times, acquired immunodeficiency syndrome (AIDS), which is caused by the retroviruses Human Immunodeficiency Virus Types 1 and 2 (HIV-1, HIV-2). However, many retroviruses cause life-long infections and appear to be relatively, if not completely benign, in their normal host species. In mice there are retroviruses that are very closely related to strongly oncogenic retroviruses but which are not themselves oncogenic, or are only very weakly oncogenic \[[@B2]-[@B5]\]. In addition, there is a whole class of retroviruses, the spumaviruses, or foamy viruses, which do not appear to be linked to any specific pathogenic state \[[@B6]\]. Even the simian equivalent of HIV-1, the causative agent of AIDS, is not pathogenic in all its hosts \[[@B7]\]. There is also a range of endogenous retroviral sequences that are not associated with specific pathologies \[[@B8]\]. Vestigial forms of retroviruses also exist; these are represented by various classes of insertional elements and can constitute a significant proportion of animal genomes \[[@B8]\]. The retroviral virion is a spherical particle of about 80--100 nm in diameter. It is enclosed by a lipid bilayer derived from the host cell plasma membrane into which one of the retroviral gene products, the envelope protein, is inserted. The virion has considerable internal structure that is mainly comprised of the products of the viral *gag*gene. In addition, the virion contains two *identical*copies of a genomic RNA molecule (the retrovirus is then genetically haploid but can also be described as pseudo-diploid), a tRNA primer for reverse transcription as well as small amounts of the products of the viral *pol*gene. The virion may also include a range of other host cell derived proteins although it is unclear whether these represent a random assortment of proteins that are coincidently incorporated into the virion or whether they play some role in the viral life cycle. Both possibilities are probably true, certainly HIV-1 is known to incorporate into its virion a number of host cell proteins that play a vital role in its life cycle \[[@B9],[@B10]\]. While the simple retroviruses have only three genes, *gag*, *pol*and *env*, the complex retroviruses encode a number of other proteins that are involved in regulating viral replication or the host cells response to the virus. For example, HIV-1 has six gene sequences in addition to the minimal retroviral complement of *gag*, *pol*and *env*. Two of these, *tat*and *rev*, encode proteins that regulate expression of the viral genome, while the other four, *vpu*, *vif*, *vpr*and *nef*, encode proteins that play multiple roles in enhancing viral replication. Retroviral life cycle --------------------- It is the unique nature of the retroviral life cycle, combined with the simplicity and advantageous arrangement of the retroviral genome, which has made retroviruses so attractive as vectors for gene therapy \[[@B11],[@B12]\]. The principal feature of the retroviral life cycle that is of interest is the ability of the retrovirus to copy its RNA genome into a double-stranded DNA form which is then efficiently and exactly integrated into the host cell genome. The integrated form is termed the provirus and it is transcribed as a normal cellular gene to produce both mRNAs encoding the various viral proteins, and the genomic RNA that is packaged into progeny virions. The genetic structure of the virus and the existence of the proviral form make it easy to manipulate retroviruses to make replication defective vectors for transfer of heterologous gene sequences. The proviral form, being DNA, can be readily isolated in standard plasmid cloning vectors and so made amenable to molecular manipulation. The genetic structure of the virus is such that the viral *cis*(sequences that are biologically active in the form of nucleic acids) and *trans*(protein coding sequences) functions (Fig. [1](#F1){ref-type="fig"}) are largely non-overlapping; indeed, as far as recombinant vectors are concerned it is possible to separate them completely, albeit at some cost in efficiency. The generation of systems capable of producing non-replication competent virus can then be achieved by placing the *cis*elements on a transfer vector construct and expressing the *trans*functions using standard recombinant plasmid expression systems (Fig. [2](#F2){ref-type="fig"}). As the genomic RNA expressed from the transfer vector construct is the only RNA molecule that carries the *cis*signals required for packaging into the virion, and for reverse transcription and integration, no viral genes are transferred to cells infected with the resulting virus. The resulting provirus, lacking all viral genes, is a replicative dead end and no further viral replication is possible. The nature of the retroviral replication process, where the U3 region of both the 5\' and 3\' LTRs of the provirus are effectively copied from the 3\' LTR of the provirus in the preceding generation, also makes possible the construction of self-inactivating (SIN) vectors. With these vectors the resulting provirus contains no active retroviral derived transcriptional promoter or enhancer elements \[[@B13],[@B14]\]. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **The *cis*and *trans*genetic functions of a retrovirus.***Cis*sequences (shown in black) are those that are directly active as nucleic acids, they include the 5\' long terminal repeat (LTR) which, in the DNA form found in the provirus acts as a transcriptional promoter, and in the RNA (genomic) form contains sequences important for reverse transcription of the genome; the primer binding site (PBS) for first strand DNA synthesis during reverse transcription; the psi (ψ) sequence which directs packaging of the genomic RNA into the virion; the polypurine tract (ppt) which is the primer binding site for second strand DNA synthesis during reverse transcription and the 3\' LTR which, in the DNA form (in the provirus) acts as a polyadenylation signal, and in the RNA (genomic) form contains sequences important for the reverse transcription process. The *trans*functions (shown in green) are the protein coding sequences, these are the *gagpol*gene, which encodes the Gag and Pol polyproteins, and the *env*gene that encodes the viral envelope protein. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Separation of the *cis*and *trans*functions of a retrovirus in a recombinant, replication defective vector system.**Replication defective retroviral vector systems are made by separating the *cis*(shown in black) and *trans*(shown in green) genetic functions of the virus into a vector construct, which contains the *cis*sequences, and helper or packaging plasmids, that encode the viral proteins (i.e. contain the *trans*sequences). To minimize overlap between the two components of the system heterologous transcriptional control elements (shown in red) are used to express the *trans*functions. Recombinant virus is made by introducing all these elements into the same cell. Only the vector transcript is incorporated into virions as this is the only RNA that contains the retroviral packaging signal (ψ). :::  ::: The use of a replication-defective retroviral vector to transfer gene sequences into target cells has been termed *transduction*, to distinguish it from the process of *infection*with replication competent viruses. It is theoretically possible that with most, if not all, recombinant vector systems, that recombination of the constituent parts of the system with each other, or with cellular sequences, can regenerate a replication competent retrovirus (RCR) \[[@B15],[@B16]\]. However, the careful engineering of these systems has led to the point where they can largely be assumed to be free of such RCR. While this does not mean that screening for RCR in preparations of vector is unimportant, as there are a number of other ways in which RCR may arise, and as quality control is obviously central to the clinical use of retroviral vectors, it does mean that in practice RCR generation should no longer be a major safety issue. This means that in terms of evaluating the safety of retroviral vectors it is the direct and indirect consequences of proviral integration that are important to consider, rather than the effects of actively replicating virus. Retroviral mediated pathogenesis -------------------------------- Retroviruses have historically been most intensively studied in animals that are either the subject of scientific experimentation (principally the laboratory mouse), or are of commercial significance (such as farmed animals such as chickens, horses, goats, cattle and fish, and pets), where they cause a number of commercially significant diseases. Indeed, the first retroviruses to be described were the oncogenic retroviruses Avian leukosis virus (ALV), and Rous sarcoma virus (RSV), which are both found in chickens. A large number of oncogenic retroviruses have now been described. These tend to cause malignant disease in a very high proportion of infected hosts. In addition, the complex retroviruses human T-cell leukemia virus (HTLV) and bovine leukemia virus (BLV) can cause leukemia in their hosts, although they do so in only a small percentage of infected individuals. The lentiviruses are also overtly pathogenic and have been shown to be the causative agent of several slow progressive diseases in animals including arthritis and encephalitis in goats, leukemia in cattle, anaemia in horses, and immunodeficiency in cats, cattle, primates and humans. The AIDS epidemic means that the lentivirus HIV-1 is now the most intensively studied retrovirus ever-incredibly, given the relative genetic simplicity of the retroviruses, there appears to be much still to learn about many aspects of HIV-1. There are also a number of viruses that cause central nervous system (CNS) pathology. For some of these, such as HIV and HTLV, CNS disease is a secondary pathology, while others are more specific in their effects. Similarly, while ALV and RSV are best known as oncogenic viruses, they are also associated with wasting syndromes. Oncogenic retroviruses ---------------------- The archetypal retroviral pathogen is the oncogenic retrovirus. Some of these are replication defective retroviruses that carry and express an oncogene sequence-indeed it was these retroviruses that largely allowed the concept of oncogenes to be first defined. These viruses induce cancers with relatively short latency periods. In addition, there are a large number of non-defective retroviruses that are oncogenic. These generally induce cancers after longer latency periods. HTLV and BLV and related viruses form a separate class of complex retroviruses that cause leukemia in a small percentage of infected individuals after very long latency periods. Retroviruses have also been associated with sarcomas in fish but these viruses have not been studied in great detail. Defective oncogenic retroviruses -------------------------------- These have been described in a number of species, but have been most extensively studied in the laboratory mouse. These are replication defective, simple retroviruses in which part of the normal viral genome has been replaced with a cDNA copy of a cellular oncogene. The viral oncogene sequence often contains mutations that make the protein it encodes act in a dominant manner. The capture of a cellular oncogene by a retrovirus is an extremely rare event, the major significance of these viruses in scientific terms is that they led to the discovery of cellular oncogenes. These viruses depend on the presence of a replication competent helper virus in order to replicate and they induce cancers with relatively short latency periods. The existence of a latency period suggests that oncogene expression is, in itself, not enough to cause malignant disease, but that additional genetic events are required. The majority of the cancers caused by these retroviruses are found in the haematopoietic system although sarcomas are also common. They are also able to transform the phenotype of cells grown in culture, principally by causing cells to lose their contact inhibition. The type of malignant event caused by any one virus is determined by the nature of the oncogene expressed by the virus and by the nature of the enhancer sequences present in the long terminal repeat which control the tissue specific expression of the oncogene. Replication defective vectors obviously also have the same potential to capture oncogenes. However, the mechanism of oncogene capture by retroviruses, and its extreme rarity, means it is probably not of major relevance when considering the risk factors associated with the use of retroviral vectors for gene therapy. Non-defective oncogenic retroviruses ------------------------------------ Non-defective, replication competent retroviruses are also associated with malignant diseases. These viruses do not carry oncogene sequences. Although first discovered in the chicken they have been most extensively studied in the laboratory mouse. These viruses induce cancer by activating cellular oncogenes *via*a number of different mechanisms. In contrast to the oncogene carrying retroviruses, these viruses are associated with much longer latency periods. This is a reflection of the relatively low probability that proviral insertion will result in activation of an oncogene, in combination with the requirement for other genetic changes before a cancer eventuates. Although proviral integration can also result in gene inactivation, inactivation of tumour suppressor genes does not appear to be a mechanism associated with any known instances of retroviral induced malignancy. The principal routes of oncogene activation are transcriptional promotion from one of the viral LTRs, and activation of endogenous cellular promoters by the strong transcriptional enhancer elements present in the viral LTRs. In the former case the provirus must obviously integrate in the sense orientation and upstream of the relevant coding sequence. Transcription can be from either LTR \[[@B17]\], and may involve splicing from either the retroviral, or cryptic, splice donor sites to a splice acceptor within the gene sequence \[[@B17]\]. If transcription is from the 3\' LTR it is usually associated with inactivating mutations in the 5\' LTR \[[@B18]\]. Transcriptional enhancement can occur with the provirus in either orientation \[[@B19]\] and over relatively large distances \[[@B20],[@B21]\]. This is by far the most common mechanism of oncogene activation. Another mechanism by which proviral integration can activate cellular oncogenes is by negation of negative regulatory elements in the oncogene or its transcript \[[@B22]\]. However, this is a rare phenomenon. If proviral integration is downstream of the oncogene translation initiation codon a dominant variant of the oncogene product may result \[[@B23]\]. Not all non-defective simple retroviruses are overtly oncogenic and the oncogenic, non-defective simple retroviruses show a spectrum of tissue specificity and oncogenic potential. Analysis of the oncogenic potential of different retroviruses has clearly shown that the major determinant of both the overall oncogenic potential of the virus, and the cell specificity of the type of cancer that results, is the viral long terminal repeat \[[@B24]-[@B27]\]. More specifically, it is the transcriptional enhancer sequences in the long terminal repeat that are the major determinant of these properties \[[@B28]-[@B33]\]. Mechanistically, this makes perfect sense. As transcriptional enhancer elements are capable of acting at a distance they will not only control transcription from the viral LTR but will also have the potential to influence transcription from promoter sequences in adjacent chromosomal genes. In contrast to oncogene activation, the oncogenic potential of some retroviruses maps to the *env*gene sequences. For example, the SU protein (p55) of the polycythemic strain of Friend virus binds to, and activates, the erythropoietin receptor resulting in massive erythroid proliferation and splenomegaly \[[@B34]\]. However, p55 does not bind to the active site of the Epo receptor and the Epo receptor is not used as the receptor for virus infection. In fact, p55 is not a functional envelope for infection and a helper virus is needed to allow the virus encoding p55 to propagate itself. In an analogous manner, the *sag*gene of Murine Mammary Tumour Viruses (MMTV) induces an immune response by interacting with the T-cell receptor \[[@B35]\]. This does not result in leukemia but facilitates the eventual induction of malignant disease in an indirect way. As the interaction between Sag and the T-cell receptor is not *via*the antigen binding site itself, a large proportion of the T-cell population (up to 10%) is stimulated. This, in turn, stimulates B-cells, the initial cellular target for infecting MMTV, allowing enhanced viral replication and the subsequent infection of mammary epithelial cells, the eventual site of tumour formation. Although Sag is a major determinant of the oncogenic potential of MMTV it should be noted that in the final analysis malignancy is due to oncogene activation. How HTLV \[[@B36]\] and BLV cause cancer is not entirely clear. Both are complex retroviruses, and in addition to the *gag*, *pol*and *env*genes common to all retroviruses, have two genes that encode regulatory proteins. HTLV causes adult T-cell leukemia, often after a very long latency period (two or three decades can pass between infection and emergence of malignant disease). Only a small percentage of infected individuals (about 1% for HTLV) develop cancer. Although the mechanism of disease induction is unclear it is certainly related to the clonal proliferation of infected cells *in vivo*. Although viral gene expression does not appear to be necessary for maintenance of the disease, evidence suggests that one of the regulatory proteins, Tax, is important in inducing the initial T cell proliferation. Given the recent development of vectors from lentiviruses, including HIV, it is worth noting that despite intense scientific scrutiny, examples of insertional mutagenesis or gene activation resulting from infection with these viruses have not been documented. However, in the case of HIV-1 the limited lifespan of most infected cells means that this observation must be interpreted with caution. In terms of replication defective retroviral vectors, the study of oncogenic retroviruses suggests that oncogene activation, *via*the provision of promoter or enhancer sequences, but especially the latter, will be the major risk factor for disease induction. In addition, selection of the retroviral envelope used for vector pseudotyping could also potentially play a role as could inadvertent transfer and expression of other retroviral proteins, at least for vectors developed from particular retroviruses, such as Friend virus. Retroviruses causing CNS disease -------------------------------- Several retroviruses cause CNS disease. Some of these, such as the murine retroviruses Cas-Br-E MLV \[[@B37]\] and FMCF98 \[[@B38]\] are specifically associated with CNS pathology. For other retroviruses, such as HTLV and HIV, CNS disease is not the defining pathology induced by the virus, even though for the latter a high proportion of infected individuals will develop CNS disease. Cas-Br-E MLV infects the brain *via*infection of the epithelial cells of the blood-brain barrier. After these become infected they release virus directly into the CNS where it infects microglial cells, resulting in a spongiform encephalopathy. The SU (env) protein has been shown to be a major determinant of the neuropathogenesis of Cas-Br-E MLV \[[@B39]\] and other neuropathogenic murine retroviruses. However, the mechanisms involved have not been elucidated although receptor activation \[[@B40]\], analogous to that caused by the SU protein of the polycythemic strain of Friend virus, has been suggested but as yet remains unproven. HTLV causes CNS disease in only a small percentage (about 1%) of infected individuals after a latent period that can be as short as two, or as long as thirty years \[[@B41]\]. The development of CNS disease is not correlated with the development of ATL. For HTLV CNS disease is characterised by a vigorous inflammatory response involving T cells that causes severe demyelination in the spinal cord. Little is known about how the virus infects the CNS and what cell types are infected, or what factors influence the induction of CNS pathology. Most individuals infected with HIV have virus within the CNS and the route of infection is thought to be transmigration of infected macrophages across the blood-brain barrier. As well as allowing opportunistic infections within the CNS there is a specific condition, AIDS dementia complex (ADC), which is a direct result of HIV infection of the CNS \[[@B42]\]. Within the CNS HIV is found in macrophages and microglia, and causes demyelination, vacuolation and gliosis. Again, the mechanism by which HIV causes CNS pathology is not well understood. The gp120 (Env) and Tat proteins have been shown to be neurotoxic *in vitro*and a number of the cytokines induced by HIV infection of monocytes and macrophages also have the capacity to damage neural tissue, either directly or indirectly \[[@B43]\]. All of the retroviruses that cause CNS disease would appear to do so as a consequence of their active replication. In the case of HIV there is direct evidence for this-treatment of patients with antiretrovirals can significantly decrease the severity of CNS disease \[[@B44]\]. However, aspects of CNS pathology remain unresolved, for example HIV encephalitis persists even during highly active anti-retroviral therapy \[[@B45]\]. Therefore, this area of retrovirus induced pathology does not appear to be of immediate relevance to replication defective retroviral vectors. However, until the mechanisms by which some aspects of CNS pathology are induced are better understood this facet of retroviral pathogenesis cannot be entirely dismissed in terms of its relevance to the design and use of retroviral vectors. Retroviruses causing immuno-deficiencies ---------------------------------------- The AIDS epidemic has brought a substantial focus to bear on the retroviruses that cause immunodeficiencies in general, and the subset of these that are lentiviruses in particular. Simple retroviruses that cause immune deficiencies in mice \[[@B46]\], cats \[[@B47]\] and primates \[[@B48],[@B49]\] have been described. Somewhat surprisingly, the pathological mechanisms in these diseases are all different. In mice, immunodeficiency is associated with proliferation of B cells (the primary target of infection), macrophages and CD4+ T-cells, all of which are non-functional. The disease is consistent with the development of anergy after antigen driven stimulation of the immune response \[[@B50]\]. Expression of a mutant *gag*gene product, Pr60 Gag, which is not processed normally \[[@B51]\], is required for induction of disease. However, the pathogenetic mechanisms involved are not understood. The defect in Gag processing makes the virus replication defective and a helper virus is required for virus spread, although not for induction of disease \[[@B52]\]. In cats the simple retroviruses that induce immunodeficiency do so *via*expression of an altered SU (Env) protein. This protein is incapable of causing resistance to superinfection \[[@B53]\]; as a consequence repeated superinfection leads directly to T cell lysis \[[@B54]\] and immunodeficiency then results due to a loss of T-cell function. The lentiviruses that have been associated with immune deficiency are FIV, SIV and HIV. All appear to share a common pathogenetic mechanism where virus infection of, and replication in, T-cells directly causes cell death, T-cell depletion and immunodeficiency \[[@B55]\]. Cell death is caused by high levels of viral replication in infected cells, although the exact mechanism is unclear. However, it is also clear that the pathogenesis of HIV-1 infection is much more complicated than this, with a complex interaction between the virus and host being played out over time \[[@B56]\]. In some non-human primates, infection with SIV is usually a chronic, but largely asymptomatic, condition \[[@B7]\]. This is thought to reflect a host/virus balance that has evolved over a long period of time. Presumably, the human AIDS epidemic reflects a recent movement of HIV into the population with a resulting imbalance between viral pathogenicity and host defences, which, after a relatively long period of infection, is resolved in favour of the pathogen. Again, the pathogenetic mechanisms involved with these retroviruses do not have major relevance to replication defective retroviral vectors. However, the pathogenetic mechanisms involved in the murine and feline immunodeficiencies caused by simple retroviruses do reiterate the point that expression of certain retroviral gene products can induce serious pathogenetic states and that this fact may have some relevance to vector design. Lentiviruses ------------ Apart from the lentiviruses mentioned above that result in immunodeficiency, there are a number of other lentiviral-associated diseases including those caused by caprine arthritis encephalitis virus (CAEV) \[[@B57]\], equine infectious anemia virus (EIAV) \[[@B58]\] and maedi/visna virus (MVV) \[[@B59]\]. For CAEV and MMV viral infection of macrophages seems to induce an inflammatory response involving macrophages and CD4+ and CD8+ T cells. It is this inflammatory response that is responsible for the different aspects of the pathology associated with infection by these viruses. EIAV causes erythrocyte lysis when high titres of cell free virus are present in the circulation. There are several mechanisms involved. Direct interaction of EIAV particles and erythrocytes results in complement mediated lysis and macrophage engulfment. This interaction is probably mediated by the Env protein. In addition, the virus appears able to suppress the differentiation of erythroid precursors. Eventually, most animals become asymptomatic carriers six to twelve months after infection. For all these viruses pathology appears to be intimately linked to viral replication. Therefore, the pathological mechanisms involved are not of direct relevance to replication defective retroviral vectors. Other retrovirus induced pathologies ------------------------------------ Retroviral infection has also been shown to be the cause of wasting and osteopetrosis in birds \[[@B60]\] and anaemia in cats \[[@B61]\]. Apart from feline anaemia, where the SU (Env) protein is a major, although not the sole determinant for the determination of pathology, the disease mechanisms involved are not well understood. However, pathology is clearly dependent on sustained viral replication meaning its significance to replication defective vectors is again limited. Pathogenic potential of retroviral vectors ------------------------------------------ From the known mechanisms of retroviral pathogenesis discussed above the most obvious pathogenic potential of retroviral vectors is (i) the production of a replication competent virus, and (ii) insertional mutagenesis, specifically oncogene activation. Clearly, the production of replication competent virus not only creates the potential of pathogenetic disease, but will also greatly increase the probability of insertional mutagenesis. In fact, in the one instance where a vector contaminated with a replication competent virus was administered to animals viral replication *per se*did not appear to have an overt pathogenetic affect, rather a T-cell lymphoma eventuated \[[@B62]\], presumably as a result of oncogene activation. Although these conclusions are obvious and widely acknowledged it is reassuring to know that there appear to be no retroviral pathogenetic mechanisms of general relevance to the safety, or otherwise, of retroviral vector systems that have been overlooked. While the inadvertent transfer of *gag, env*and other retroviral genes also has the potential of inducing a pathogenetic state this would appear to depend on the specific retroviral gene sequence in question and to not be of general significance. Even so, minimizing the inadvertent transfer of retroviral gene sequences should clearly be an objective when developing retroviral vectors, not only because of this issue but also because it will have a bearing on the likelihood of replication competent virus being produced and of an endogenous retrovirus being activated. In addition, even though oncogene capture by retroviruses is an extremely rare event, the very significant pathogenic potential of the viruses that result means that it should also be taken into consideration during the development of retroviral vector systems. For various reasons, not least of which has been the problem of achieving positive experimental outcomes, only the issue of reducing the probability of replication competent virus arising has been systematically addressed during the development of retroviral vector technology. Indeed, great care has been taken in the development of retroviral vector systems to minimise the chance of producing replication competent retroviruses \[[@B63],[@B64]\]. However, although clear means of doing so have been described \[[@B13],[@B14],[@B65]\], the need to minimize the probability of oncogene activation has often been made secondary to the issue of efficient transgene expression \[[@B66]\]. This has especially been the case with oncogenic retroviral vectors where transcriptional silencing has been a major problem \[[@B67]\]. Replication competent virus --------------------------- The generation of replication competent virus has, from the very beginning, been seen as the major safety issue for retroviral vectors and this has led to a prolonged effort to develop means of minimising the probability of it arising. There are two principal ways in which replication competent virus can be produced. The first of these is through recombination of the constituent parts of the vector system (i.e. vector and helper *trans*function plasmids), either with themselves or with endogenous viral sequences in the cell lines used for virus production \[[@B15],[@B16]\]; the second is by activation of an endogenous proviral sequence. The first of these issues has been addressed by (i) breakdown of helper functions onto different plasmids; (ii) manipulation of codon usage in helper plasmids; (iii) removal, or mutagenesis, of unnecessary *cis*sequences present in the vector; (iv) the development of SIN vectors; (v) the minimisation of homology between the separate plasmids that make up the system; and (vi) the use of cell lines that do not contain endogenous retroviral sequences with homology to the vector system \[[@B13],[@B14],[@B63]-[@B65]\]. Although for many vector systems each of these approaches requires further refinement, in principle, they clearly provide the basis for the construction of vector systems where the probability of replication competent virus being produced *via*any of these mechanisms appears to be remote. While this doesn\'t negate the need for appropriate quality control procedures, especially as there is still the remote probability of inadvertent activation of an endogenous retrovirus from the cell line used for virus production, it means that the major safety issue faced by those wishing to use retroviral vectors is that of insertional mutagenesis and oncogene activation. Insertional mutagenesis and oncogene activation ----------------------------------------------- As discussed above, oncogene activation can occur either by transcription from one of the proviral LTRs, or by activation of an endogenous promoter by provision of transcriptional enhancer elements. The transgene aside, these events would appear to depend absolutely on the presence of active transcriptional control elements in the viral LTRs as evidenced by the critical role LTR sequences play in determining the ability of most non-defective retroviruses to induce cancers, and in determining the tissue specificity of cancer induction. There is no evidence that retroviruses contain transcriptional control elements of significance in other parts of their genomes. Therefore, the main approaches to minimizing the probability of oncogene activation must be the development of vectors from non-oncogenic retroviruses, the careful development of the SIN vector principal, and careful consideration of the promoter used to drive transcription of the transgene (see below). Retroviral gene transfer ------------------------ The minimization of the inadvertent transfer of retroviral genes to target cells is clearly a worthwhile objective as some of these genes have direct pathogenic potential and they may also influence the probability of endogenous retroviral sequences in the target cell being activated. Generally, the principles applied to the design of vector systems in order to minimize the probability of RCR being produced will also minimize the probability of inadvertent retroviral gene transfer. However, as the production of RCR requires multiple recombination events more effort should be made to analyse the rate of transfer and expression of individual retroviral gene sequences by vector systems. It is clear that the rate of individual gene transfer is much higher than the rate of RCR generation and can occur at a significant frequency even in highly evolved systems where RCR cannot be detected \[[@B68]\]. This suggests that further efforts need to be made to assess and reduce the rate of transfer of retroviral genes. Oncogene capture ---------------- The mechanism of oncogene capture appears to be dependent on the generation of a chimeric retroviral-oncogene transcript (69, 70). This suggests that the risk of oncogene capture will be related to the efficiency of termination/polyadenylation of the proviral transcript and that this should be considered and assessed in the process of vector development, especially as retroviral polyadenylation sequences are often relatively inefficient, perhaps reflecting the necessity for the polyadenylation signal to be inactive in the context of the 5\' LTR. However, in transient virus production systems, where the transfected vector plasmid presumably remains either entirely, or almost entirely extrachromosomal, this mechanism would appear to preclude the probability of oncogene capture. In the case of stable producer cell lines there is clearly an argument for categorizing the integration site of the vector sequence and discarding any clones where this is in a known or suspected oncogene. Adverse events in animal experiments and clinical trials -------------------------------------------------------- The adverse events that have been observed in animal experiments and clinical trials reinforce the conclusions discussed above, that replication competent virus \[[@B62]\] and insertional mutagenesis \[[@B71],[@B72]\] are the two risk factors of significance in retroviral mediated gene therapy. The two known instances where insertional mutagenesis/oncogene activation has resulted from the administration of a replication defective retroviral vector suggest that, the design of the vector aside, there are additional risk factors that influence the probability of an adverse event, the most obvious of these being the specific transgene expressed from the vector \[[@B71],[@B73],[@B74]\] which in both cases is a gene capable of influencing cell growth (although in neither case can it be considered a classical oncogene). In terms of the influence of vector design it is interesting to note that in both of these instances the same vector, pMFG \[[@B66]\] was used \[[@B75],[@B76]\]. This vector is derived from MoMLV, a strongly oncogenic murine retrovirus, and notably uses the viral LTR to drive expression of the transgene. In both cases the vector appears to have been chosen primarily for its ability to efficiently drive transgene expression in haematopoietic lineages without consideration that this may also select for an increased risk of oncogene activation. Given the historical difficulty of obtaining good transgene expression from MoMLV derived vectors in haematopoietic lineages, and the lack of evidence to suggesting that oncogene activation was a significant safety issue with replication defective MoMLV vectors, it is not surprising that this approach was taken. Indeed, it is generally believed that, in general, the risk of insertional mutagenesis, while poorly defined, is probably substantially lower than seen in the X-SCID trial \[[@B77]\] where there appear to be a number of specific secondary risk factors \[[@B72]-[@B74]\]. In the absence of such secondary risk factors it is unclear what the real risk is; given the complexity of cellular and genetic regulatory processes and networks it is also unclear how many apparently innocuous transgenes will in fact increase the risk of adverse effects when expressed in a constitutive manner. However, no adverse events have been reported for the long running ADA-SCID trial where mature T-cells were targeted \[[@B78]\] or in PBL and PHSC targeted gene therapy for the same condition \[[@B79]\], although in both cases the number of patients who have been treated is very small. In all these protocols a non-self inactivating MoMLV derived vector was used. However, even with these unknowns it is apparent that improvements in vector technology, such as the use of SIN vectors, will greatly reduce the risk, whether or not additional risk factors are present. In terms of the vector technology used on the two occasions where oncogene activation has been observed the following comments can be made: 1\) The vector is derived from MoMLV and uses the LTR sequence to drive the transgene *via*splicing. MoMLV is a strongly oncogenic, non-defective virus that causes B-and T-cell lymphomas and leukemias in mice. As with other non-defective oncogenic retroviruses the primary determinant of its pathological properties is the long terminal repeat enhancer. MoMLV has been shown to induce oncogenesis *via*activation of any one of a number of different cellular genes (*Ahi1*, *Bla1*, *Bmi1*, Cyclin D2, *Dsi1*, *Emi1*, *Ets1*, *Evi1*, *Gfi1*, c-Ha-ras, *Lck*, *Mis2*, *Mlvi2*, *3*and *4*, c-*myb*, c-*myc*, N-myc, *Notch1*, *Pal1*, *Pim1*and *2*, prolactin receptor, *Pvt1*, *Tiam1*and *Tpl2*). 2\) The vector LTR is used to control transcription of the transgene. In the case of the X-SCID trial there is a strong selective pressure for gene corrected cells and accordingly there will clearly be an equally strong selection for transduced T-cell clones in which the LTR is active. 3\) The PHSC is notoriously difficult to transduce with oncogenic retroviral vectors and the protocol used was designed to enhance transduction by using multiple cytokines to stimulate division of PHSC. This is likely to induce many genes involved in regulating cell growth. As retroviruses preferentially integrate into active gene sequences, this would increase the number of growth regulating genes accessible as targets for provirus integration and hence promiscuous, unregulated activation. Specifically, LMO2, the oncogene activated in the X-SCID trial, is normally expressed in primitive haematopoietic cells (the target for gene transfer) but not in mature cells (80). Therefore, it will be accessible for proviral integration during the transduction process and its continuing expression in maturing T cells generated from gene corrected precursors is biologically inappropriate. The problems that occurred in this X-SCID trial, their broader relevance and possible answers, have all been reviewed from a number of aspects \[[@B72],[@B73],[@B77]\]. However, the focus has been on the biology of the system, and little attention has been paid to how technological changes in vector delivery systems and protocols might impact on the risk of insertional mutagenesis/oncogene activation. Given what is known about retroviral mediated insertional mutagenesis it is surprising that more attention has not been paid to the technology used in many of the retroviral mediated gene therapy animal studies and human trials. With hindsight, it seems that the technologies used were selected on the basis of efficacy, not safety, that is achieving adequate gene expression took preference over consideration and assessment of insertional mutagenesis. However, given the technical difficulties involved in developing a workable protocol this is not surprising, and it is a pre-occupation that was, and is, shared by all gene therapy researchers. Possible technological approaches that would appear to provide answers to these issues include: 1\) The use of self-inactivating (SIN) vectors would make a major difference in that the provirus would lack all U3 enhancer sequences, negating the ability of the LTR to activate cellular genes. The vector should also not contain active splice signals. However, given the ability of SIN vectors to be repaired at a significant rate during virus production (see below) careful selection of the retrovirus used to build the vector backbone is also important if this risk is to be minimised. Clearly the construction of vectors from non-oncogenic retroviruses and the development of more effective (i.e. less prone to LTR repair) SIN vectors is warranted. If SIN vectors are to be used the transgene must be expressed from an internal promoter which must also be presumed to have the potential for oncogene activation. Therefore it would be preferable to use a promoter without highly active enhancer elements. In addition, the wisdom of incorporating matrix/scaffold attachment regions into vectors to increase expression may also be contraindicated as these sequences have long-range enhancer like properties (81). If high levels of gene product are required, consideration should be given to other means to enhance transgene expression, such as codon-optimisation of coding sequences. 2\) Vectors should be developed from non-oncogenic retroviruses. The recent development of vectors from HIV-1 and other lentiviruses for unrelated reasons (predominantly their ability to transduce non-cycling cells) means that this has already happened. The Tat dependence of the HIV-1 LTR may also provide an extra measure of safety as long as Tat is not transferred along with the vector. However, the enhancing properties of the HIV-1 LTR in the presence and absence of Tat needs to be carefully defined in order to test the assumption that the HIV-1 LTR lacks the ability to *trans*-activate adjacent promoters. The different integration specificities of lentiviral (centrally in active gene sequences) and oncogenic (promoter adjacent in active gene sequences) retroviruses and vectors \[[@B82]\] also give reason to suppose that the former may be less likely to cause oncogene activation. However, this remains to be directly demonstrated. 3\) The incorporation of strong transcription termination/polyadenylation signals and gene isolator sequences (83) may provide another means to reduce the possibility of adjacent genes being activated. These sequences should also reduce the probability of oncogene capture in virus producer cells. However, the incorporation of insulator sequences appears to lead to a significant loss of vector titre (84). 4\) When the transgene plays a role in regulating cell growth, extra consideration should be given to using the relevant control signals from the gene in question to regulate expression of the transgene. 5\) Although the PHSC is theoretically a very attractive target for gene transfer it is extremely difficult to transduce with retroviral vectors derived from oncogenic viruses such as MoMLV. Although efficient transduction of human PHSC can now be achieved this requires exposure to multiple cytokines over a relatively long culture period. The potential of new retroviral vectors derived from lentiviruses (85) and spumaviruses (86) to transduce PHSC with shorter exposure to less cytokines needs to be fully explored. 6\) In general the limitations of vectors should be taken into account when designing gene therapy protocols. For example, in the case of X-SCID, it may be just as efficacious to target a more committed T-cell precursor that can be transduced more easily, and without biological manipulation using multiple cytokines. Alternatively, if the PHSC is to be targeted as highly enriched a PHSC population as possible should be used in order to expose the patient to the minimum number of transduction events compatible with the desired outcome. In the two X-SCID patients who developed T cell leukemia, molecular analysis of samples collected before the appearance of malignant disease showed the presence of \>50 γc transduced T cell clones. Approximately 14 to 20 million transduced CD34^+^cells were infused into these patients. Therefore, it would appear the patient is exposed to a much greater number of transduced cells than is theoretically necessary to produce the desired result. In other words, the process of generating gene corrected T cell clones by transduction of CD34^+^cells is very inefficient. SIN vectors, how good are they? ------------------------------- With hindsight SIN vectors \[[@B13],[@B87],[@B88]\] now appear likely to be one of the most important general developments in retroviral vector technology since the advent of replication defective vector systems in the 1980s. SIN vectors take advantage of the reverse transcription reaction in which the U3 region of the 3\' LTR acts as the template for the U3 region in *both*LTRs of the provirus. As the transcriptional enhancer elements in the 3\' LTR are redundant in the context of a retroviral vector construct they can theoretically be deleted without affecting vector performance. After transduction of the target cell both LTRs are deleted and are transcriptionally silent. Although this requires that an internal promoter is used to control expression of the transgene, and makes it more difficult to generate high titre stable packaging cell lines, the advantages of the approach are obvious. However, SIN vectors have not been widely used in the case of oncogenic retroviral vectors, principally because viral titres were low, because of high rates of repair of the SIN deletion \[[@B13],[@B89]\] and because of negative effects of the SIN deletion on gene transfer efficiency \[[@B90]\]. Subsequently, by the use of a heterologous promoter in the 5\' LTR an effective SIN vector based on spleen necrosis virus was developed \[[@B91]\] but this vector has not been widely utilized to date. In contrast, SIN vectors have been widely adopted in the lentiviral vector field \[[@B14],[@B65],[@B92],[@B93]\] where transient expression systems are generally used to produce virus, avoiding the difficulties of making stable cell lines associated with SIN vectors. In addition, in terms of transgene expression, lentiviral SIN vectors appear to perform as well as, if not better than, vectors with an intact 3\' LTR \[[@B93],[@B94]\]. However, even with vectors with large 3\' LTR deletions it is obvious that repair of the SIN deletion also occurs at a significant rate with lentiviral SIN vectors \[[@B14],[@B92]\]. Therefore, while the concept of SIN vectors is a powerful one, further development and rigorous testing of this technology is required before it can be confidently used to address the problems of insertional mutagenesis. Conclusion ========== The most important determinant of the safety of retroviral vectors remains ensuring they are free of replication competent retrovirus of any sort. Clearly, the technologies available for the production of vector virions would appear able to preclude the production of replication competent virus by recombination of the constituent parts of the vector system (i.e. vector and helper plasmids) with a very high degree of certainty. However, production of replication competent virus from the cell lines used for virus production remains a theoretical possibility and more work needs to be done on generic assays for replication competent retroviruses. Apart from the issue of replication competent virus, analysis of the pathologies associated with retroviruses, and the results of the X-SCID trial, demonstrate that careful attention must be paid to the ability of sequences in retroviral vectors to activate transcription of genes adjacent to proviral integration sites. Although the use of SIN vectors will greatly reduce the risk of such events, given the predilection of current SIN vectors to be repaired during virus production these vectors need to be further developed, especially for vectors derived from strongly oncogenic viruses. In addition, inadvertent transfer to, and expression in, transduced cells of *gag, env*(SU) and other retroviral gene sequences would appear to of relevance and needs to be specifically addressed in the development of vector systems. As both oncogenic (MoMLV derived) and lentiviral (HIV-1 derived) vectors have been shown to preferentially integrate into transcribed sequences it would appear logical that the likelihood of proviral integration near cellular genes involved in the positive regulation of cell growth would be increased in actively growing cell populations. This suggests that the use of transduction protocols that target non-cycling cells, or cells that are subjected to the minimum of stimulatory signals as is compatible with efficient gene transfer, would be greatly advantageous in terms of minimising the risk of malignant events after the stimulatory signals are removed. With hindsight, the observation of malignant events induced by replication defective MoMLV retroviral vectors is not surprising although the frequency of these events in the X-SCID gene therapy trail certainly was. The concern is that these events will now cause a significant backlash against the use of all retroviral vectors while the real message is that we need to make better use of the knowledge we now have in terms of designing vectors and gene therapy protocols. Clearly, the known oncogenic potential of MoMLV and its relationship to viral sequences has, for one reason or another, been largely ignored to date. Indeed, most of the retroviral vectors used in trials to date are based on MoMLV and contain an intact 5\' LTR. While the historic reasons for this are obvious we now need to evaluate and adopt more appropriate technologies as rapidly as possible. There are several obvious conclusions to be drawn from the X-SCID trial and the results of Li *et al*\[[@B71]\]. The first is that in the absence of additional risk factors the risk of malignant events resulting from exposure to a replication defective retroviral vector is low but remains to be accurately quantified. Secondly, what constitutes an additional risk factor is hard to predict, making risk assessment difficult. However, even with these unknowns, there exist technological approaches that should greatly reduce the risks associated with retroviral mediated gene therapy. These include SIN vectors and new types of retroviral vectors (namely lentiviral vectors) that may allow simpler transduction protocols that perturb the normal state of the target cell less than current approaches. This is especially true when the PHSC is the target of gene transfer; current protocols using oncogenic retroviral vectors rely heavily on manipulating the state of the target cell by exposure to multiple cytokines over relatively long periods. These protocols are also relatively inefficient; this reflects the poor match between the target cell and properties of these vectors. In contrast, vectors derived from lentiviruses and spumaviruses appear to allow more efficient transduction of PHSC with less requirement for cytokine stimulation of the target cells \[[@B85],[@B86],[@B95]-[@B97]\]. It is of note that the use of lentiviral vectors may also be preferable in other ways. Not only do they have uniquely positive properties as gene therapy vectors, there is no evidence that the viruses from which they are derived are able to induce gene activation using the same mechanisms as used by non-defective oncogenic retroviruses. Regulatory authorities also have a role to play. Clinical trials are based on extensive preclinical experimentation and animal trials that take many years to complete. Clearly, the particular vector system that has been used to develop a protocol may no longer be the best to use ten years later when clinical trials become a reality. The question is, then, how can the regulation of clinical trials be made flexible enough to allow the introduction of new and improved vector technology late in the process? In conclusion, retroviral mediated gene transfer remains an extremely attractive option for gene therapy when the stable and permanent genetic modification of the target cell is optimal. However, we must take greater care, and utilise more resources, for the pro-active, rather than reactive, refinement and testing of the basic technology that is used for gene therapy and for the adoption of improved vector systems if adverse events are to be minimised.

PubMed Central

2024-06-05T03:55:47.301884

2004-8-13

PMC515180

Background ========== Night sleep duration of Japanese children aged 10--18 y has become shorter by one hour during the last 30 years in Japan \[[@B1]\]. The so called \"24-hour society\", which is currently in progress in Japan, seems to change environmental conditions surrounding children. For example, mobile phones are used by more than 90 % of university students and more than 30% of junior high school students living in the urban area of Kochi city (33°N) \[[@B2]\]. Students can communicate with their colleagues even in the middle of the night with mobile phones. Frequent or long-time (more than 30 min) usage of the mobile phone makes university and junior high school students more evening-typed \[[@B2]\]. \"Convenience stores\" are open for 24 hours and provide several kinds of food and other goods for general civilian life. Convenience stores are now common all over Japan even in suburban areas. Illumination inside the convenience stores is very bright (2000 lux or more at the level of the eyes). Bright lighting in retail stores seems to be a merchandising technique which has been in use worldwide for at least 60 years. Unconscious use of bright light in the evening or at night inside the convenience store may promote a circadian phase delay in students exposed to the bright light during the first half of subjective night. This hypothesis is based on a \"light-pulse\" experiment in the laboratory \[[@B3]\] as follows. Adult subjects were exposed to light pulse of 4000--6000 lux for 30 min at one of several phase points of their circadian rhythm, and the direction (advance or delay) and the extent of phase shift caused by the pulse were measured at each phase point. The light pulse delayed the phase of sleep-wake cycle by the subjects when they were exposed to the light pulse in the first half (about 19:00--24:00) of subjective night. However, it advanced the phase effectively when the subjects were exposed to it within three or four hours after the minimum point of inner body temperature (about 5:00--9:00). An epidemiological study about the effect of the convenience store usage was conducted on sleep habits and diurnal rhythm by about 500 students attending junior high school aged 12--15 y in Kochi Japan \[[@B4]\]. This latter paper reported the following three points: (1) Students going to convenience stores after sunset were more evening-typed and showed shorter night sleep of 7.0 hours on average than those going to convenience stores during the daytime, who showed night sleep of 7.3 hours on average, (2) Students who went to convenience stores every day slept only 6.4 hours on average and the sleep hours were significantly shorter than the 7.5 hours shown by students who went to convenience stores only 0--1 time per week, (3) Students who stayed more than 30 min in convenience stores took shorter night sleep of 6.6 hours on average than those who stayed there less than 15 min. (7.3 sleep hours). Younger children attending kinder garden and students attending elementary school were more sensitive to \"light conditions\" in normal life than university students, according to an epidemiological study \[[@B5]\]. However, no experimental field studies on the effects of \"light conditions\" during normal life have been conducted on sleep-wake cycles of healthy children younger than 15 y. Melatonin, which is synthesized in the pineal organ and secreted to the blood, is well known as a key substance which may be effective in promoting the falling into night sleep by humans \[[@B6]\]. Blood melatonin concentration by adult human subjects is extremely low during daytime and increases rapidly at 22:00--23:00 up to as much as ten or twenty times daytime values. The high level is maintained till the early morning and then decreases again rapidly to the extremely low concentration typical of the daytime \[[@B7],[@B8]\]. The increase in melatonin concentration might occur in late evening and trigger the falling into sleep also for healthy children, although there have been no studies on melatonin concentration in salivary or blood under their normal life. In adult subjects, a single administration of 5 mg of melatonin at 13:00 was reported to induce higher subjective sleepiness during the following 2 hours and also higher EEG power density in the range of relatively low frequency of 5.25--9.0 Hz rather than that of placebo \[[@B9]\]. When adult subjects were exposed to 400 lux lights for more than 30 min or exposed to 300 lux lights for more than 2 hours, melatonin level during the night became lower than that when they stayed under dimmer lights \[[@B10],[@B11]\]. In the case of adolescents and children, the exposure to lights of 300 lux or more during the first half of subjective night in the normal life might decrease their melatonin level and prevent the falling into sleep. Currently, more than 80% of junior high school students of the third grade aged 14--15 y in Kochi go to private school in the evening. If they take a short stop at the convenience store to get some fast food and enjoy talking with their colleagues in front of the store before or after going to the private school in the evening, they suffer the double exposure to bright lights at the school and at the convenience store. Such bright lights are from fluorescent light bulbs and include blue or blue-green lights with 470--500 nm wave lengths which were reported to be powerful to suppress melatonin concentration \[[@B12]\]. Based on the epidemiological studies made in 2001--2003 on junior high school students in Kochi Prefecture (33°N), 38.8% of the students who frequented convenience stores went there after sunset, and 30.2% and 6.5% of junior high school students who used convenience stores went there and stayed there for 15--30 min and longer than 30 min, respectively. Moreover, this epidemiological study showed that 62.4% and 18% of the students who went to the evening private school studied there for 2 and 3 hours until 9 or 10 o\'clock in the evening, respectively. In total, junior high school students were estimated to be exposed to bright lights of more than 2000 lux inside private school and/or convenience store for 2--3.5 hours on average in the evening. Such exposures are expected to suppress the increase in blood melatonin level as a direct effect and also delay the phase of their circadian systems driving melatonin secretion rhythm and sleep-wake cycle. In this study, two light conditions were investigated. One was bright and high color-temperature light of more than 2000 lux, which is used in the evening and at night inside the convenience stores, at preliminary and private school for entrance examination, and at rental video shops in Japan. The other condition was a dark and low color-temperature light with less than 60 lux which is usual in the evening for traditional Japanese settings (a fireplace, candle, or a naked light bulb). Methods ======= Participants ------------ Experimental participants were ten Japanese junior high school adolescent students (4 females and 6 males) aged 14--15 y who were attending Motoyama junior high school located in the mountain area of Reihoku district (33.5°N) in Kochi Prefecture. They had enjoyed New Year holidays for 7 days before the experiment. They were instructed to keep usual diurnal rhythm (for example bed time and wake-up time) during the holidays. Before the experiment, participants were divided into the two groups of \"bright light experimental group (BLEG)\" and \"dim light control group (DLCG)\". Participants in BLEG were selected to show similar circadian typology to those in DLCG based on the scores in the morningness-eveningness (M-E) questionnaire of Torsvall and Åkerstedt \[[@B13]\] (mean ± SD: 15.00 ± 4.30 by BLEG and 14.80 ± 4.09 by DLCG). Bed time, wake-up time and sleep hours shown by BLEG for the four days just before the experiment were 23.0 ± 4.2 hours, 8.4 ± 1.9 hours and 9.1 ± 1.4 hours, respectively; corresponding values for DLCG were 23.8 ± 1.3 hours, 8.9 ± 1.3 hours, and 9.5 ± 1.5 hours. Each group consisted of two females and three males. All the ten participants sampled their own saliva using \"Salivette\" collecting tubes (SARSTEDT Aktiengesellschaft & Co., Numbrecht, Germany) at 22:30--23:00 under the 200--300 lux light from fluorescent light bulbs in their home on the day before the experimental day. Japanese civilians seem to enjoy evening time during the first half of subjective night (after sunset till bedtime) under fluorescent light bulbs based on our unpublished questionnaire study on 950 families having small children aged 0--6 yrs in Kochi. More than 85% of the 950 families enjoyed evening life under fluorescent light bulbs. We measured the illumination at the level of 1 m above floor just under a usual type of round-shaped fluorescent light bulb in a typical one-room apartment for students and it was 340 lux. Procedure --------- On the experimental day of the 5^th^January 2003, all the ten participants got together in front of Motoyama junior high school at 8:00 in the morning. A wagon car took them to the experimental place which was a Japanese style hotel located at a mountain area, Yusuhara town in Kochi Prefecture, 126 km west from Motoyama town. During the driving, illumination inside the car was 350--500 lux. The car arrived at the hotel around noon. It was snowing through the day. Behavior of all the participants was controlled during the stay in the hotel till the next morning of the experimental day. All the participants played outside exposed to the sun light with 6000--7500 lux at the eye level during 12:30--13:30 and 14:00--14:50. They were allowed to have a rest in a living room in which the floor was filled with 12 *tatami*mats and the illumination at the eye level was 250 lux from fluorescent light bulbs during the rest of the time till 16:30. Participants took bath one by one between 16:30--18:00 and had supper all together between 18:15 and 19:20 in the living room. At 19:25, the participants of BLEG moved to a Japanese style room with 8 *tatami*mats where they were exposed to the light with 2000 lux at the eye level from fluorescent light bulbs, whereas DLCG group members moved to another Japanese style room with 8 *tatami*mats where they were exposed to the light of 60 lux and relatively low color-temperature from a electronic light bulb. All the participants included in both groups were home-working or making a small wooden folk craft object that is typical in the Yusuhara district, under each light condition till 22:30. Room temperature was controlled at 15 ± 2°C with an oil heater in both groups. Then they came back to the former living room (12 *tatami*mats) and stayed there under the light of 250 lux till 23:40. Then female and male participants moved to separate rooms and went to bed just before 24:00. Salivary samples were collected in collection tubes at 21:45, 22:30, and 23:40, and these salivary samplings were preserved in a refrigerator at less than -20°C. Melatonin concentration in the samples was analyzed by a professional analyzing company (MSL Co. Ltd.) which was a specialist for several chemical and microbiological analyses. All the participants from both groups were called out to get up at 7:00 in the next morning. All the participants got up between 7:00--7:15 responding to the calling out. After taking breakfast, they left the experimental place at 9:00 back for Motoyama junior high school. Throughout the study, light exposure was measured on the eye level with a digital illumination meter. Detailed explanation of the objectives and methods of the experiment was provided before the experimental performance to the participants and their parents. The research project received full and complete agreement from all of them. Results and Discussion ====================== The results are shown in Fig. [1](#F1){ref-type="fig"}. Salivary melatonin concentration rose from 3.00 ± 3.34 (mean ± SD) pg/ml at 21:45 to 9.18 ± 7.66 pg/ml at 23:30 of the experimental day in the DLCG (t-test between values at 21:45 and 23:30: t = 3.60, df = 4, p \< 0.05), whereas it remained at less than 1.3 pg/ml till 23:30 in BLEG (t = 2.07, df = 4, p \< 0.2). There was no significant difference in the melatonin concentration between BLEG and DLCG in the day before the experimental day (Wilcoxon test: z = -1.163, p = 0.31). At 22:30 of the experimental day, melatonin concentration by BLEG tended to be lower than that on the day before the experimental day (Wilcoxon test: z = -1.604, p = 0.109), while the concentration became significantly higher in DLCG (z = -2.023, p = 0.043). On the day before the experimental day, all the participants were under Japanese standardized light condition with 200--400 lux from a fluorescent light bulb with relatively high color-temperature. On the experimental day, the bright light of 2000 lux in BLEG suppressed the expected night increase of melatonin concentration, whereas the relatively low color- temperature light with 60 lux did not. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Effects of light condition on salivary melatonin concentration.**Values shown are means (n = 5 per group) and SEM. :::  ::: In Japan, bright and high color-temperature light of more than 2000 lux is available in the evening or night inside convenience stores which are open for 24 hours and private schools for the preparation to go through the entrance examination to upper schools. Also in usual life, such exposures to bright lights in the evening private school and convenience store can suppress the night increase in blood melatonin level as a direct effect and possibly delay the circadian system that drives the melatonin secretion rhythm and sleep-wake cycle. The results of this study suggest that ancient Japanese lighting in the evening and at night, which could be supplied by a traditional Japanese hearth fire or a oil lamp or candle (20--30 lux), might be healthy for adolescents and children, because the ancient lights could allow rapid and clear increase in melatonin level leading to a smooth falling into night sleep \[[@B14]\]. Conclusions =========== Bright lights of 2000 lux and even moderate lights of 200--300 lux can inhibit, as a direct effect, nocturnal melatonin concentration in children. Ancient Japanese light conditions which could be supplied by a traditional Japanese hearth fire or a small oil lamp or candle might be healthy for children, because the ancient lights could allow rapid and clear increase in melatonin level in the evening, leading to a smooth falling into night sleep. Competing interests =================== None declared. Acknowledgement =============== I would like to thank all 10 participants attending Motoyama junior high school and their parents for well understanding of the objects of this study and nice participation with complete agreements. Thanks are also due to the staff of the Japanese style hotel which was used as the experimental place, educational staffs in Reihoku District especially Mr. Hirotaka Kageyama, Vice-principal of Motoyama junior high school and Mr. Yasuhiro Yamashita, Principal of Ohsugi junior high school, and all the staffs of Japan Broadcasting Corporation for invaluable helps to the preparation and performance of the experiment. This study was supported by Support Funds to Researches by President of Kochi University, 2003--2005, and Meiji-Yasuda Mental Health Foundation, 2004--2005 to T. Harada.

PubMed Central

2024-06-05T03:55:47.307007

2004-8-11

PMC515284

Background ========== Many *Drosophila melanogaster*homologs of the proteins required for both DNA replication and repair have been identified and in several cases purified to apparent homogeneity. These include DNA polymerase α holoenzyme \[[@B1],[@B2]\], DNA polymerase δ(pol δ) \[[@B2]-[@B4]\], replication protein A (RP-A; \[[@B5]\]), replication factor C (RF-C; e.g., see \[[@B6]-[@B9]\]) and various origin recognition complex (ORC) subunits (see e.g., \[[@B10],[@B11]\]). Moreover, complete replication of DNA containing the SV40 origin of replication has been reconstituted *in vitro*using purified SV40 T-antigen and *Drosophila*cell-free extracts \[[@B7]\]. A protein about which much information has been obtained is proliferating cell nuclear antigen (PCNA). *Drosophila*PCNA was first identified both as a highly purified protein able to substitute, albeit poorly, for human PCNA in a cell-free SV40 DNA replication system reconstituted from purified proteins \[[@B12]\] and by Yamaguchi et al. \[[@B13]\] who used an oligonucleotide probe to detect the *Drosophila*PCNA cDNA and gene, express the protein in *E. coli*and deduce its complete amino acid sequence. Further results indicated that in flies, PCNA was encoded by a single gene located at position 56F5-15 on the right arm of chromosome 2. This was subsequently identified as the *Drosophila mus*209 locus \[[@B14]\]. Recently, a second *Drosophila*PCNA gene of limited homology to the original and of unknown biological function has also been found \[[@B15]\]. Protocols have been established for purification of wild-type human PCNA from tissue culture cells \[[@B16],[@B17]\], unmodified wild-type human PCNA after regulated expression in *E. coli*\[[@B18]\] and NH~2~-terminally his-tagged but otherwise wild-type human PCNA, also engineered for bacterial expression \[[@B19]\]. All were comparably effective at stimulating mammalian pol δ. Similar protocols have been developed for *Drosophila*PCNA and strategies for site-directed mutagenesis have been devised and implemented \[[@B20]\]. Recently, Zhang et al. \[[@B21]\] (see also \[[@B22]\]) as well as others (e.g., see \[[@B23]\]) identified the interdomain connector loop of PCNA (amino acids 119-133 of human PCNA) as crucial for binding pol δ. Of note, relative to wild-type PCNA, mutations of the molecule within this region such as glutamine at position 125 changed to glutamic acid (Q125E) promoted increased pol δ-processivity \[[@B21]\]. In human PCNA, residues 123, 126, 127 and 128 were defined as being essential for interaction with pol δ \[[@B21]\]. Comparison of human with *Drosophila*PCNA sequences in this region indicated that of these four amino acids, three (residues 126, 127 and 128) are identical. The fourth, residue 123, is glutamine (Q123) in wild-type *Drosophila*PCNA. The corresponding residue in human PCNA is valine (V). To investigate the role of the interdomain connector loop of PCNA on the effects of PCNA on pol δ, we mutagenized residues within this region of *Drosophila*PCNA so that they more nearly resembled human amino acids. After bacterial expression and purification, we tested the effects of these site-specifically modified (\"humanized\") *Drosophila*PCNA molecules on purified calf thymus pol δ (two-subunit form; see \[[@B17],[@B24]\]). Calf thymus and human pol δ are highly similar in amino acid sequence \[[@B25]-[@B27]\] and can, for our purposes, be used interchangeably. \"Humanization\" of a single *Drosophila*PCNA residue, conversion of Q123 to V (Q123V), conferred upon it, enhanced ability to affect several properties of calf thymus pol δ. More extensive mutagenesis, in which the entire interdomain connector loop of *Drosophila*PCNA (amino acids 119-133) was replaced by the corresponding human residues, was still more effective at stimulation of calf thymus pol δ, than either wild-type or Q123V *Drosophila*PCNA. However, it was considerably less effective than wild-type human PCNA at altering the properties of calf thymus pol δ. These results therefore suggest that in addition to the interdomain connnector loop, other regions of PCNA are also important effectors of pol δ activity. They also provide a means to couple operationally, the considerable power of *in vivo*genetic analyses performed in *Drosophila*with the sophistication of mammalian biochemistry. Results ======= To study the role of the interdomain connector loop of PCNA (amino acids 119-133), we compared human and *Drosophila*homologs. Of the 15 interdomain connector loop residues, nine are identical between the two; identical residues are shaded (Fig. [1A](#F1){ref-type="fig"}). Overall, *Drosophila*PCNA is \>70% identical to that from mammals (e.g., humans; see \[[@B13]\]). Others showed that PCNA residues 123, 126, 127 and 128 were essential for interaction with pol δ \[[@B28]\]. Of these four, only one (residue 123) differs between flies and humans. Also shown is a model constructed from the X-ray crystallographically determined structure of PCNA indicating the locations of the sites to be mutated in *Drosophila*PCNA (Fig. [1B](#F1){ref-type="fig"}). Shown (Fig. [1B](#F1){ref-type="fig"}) is the X-ray crystal structure of human PCNA. The *Drosophila*homolog is assumed to be similar. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Structure and structural rationale for mutating *Drosophila*PCNA.**A: amino acid sequences of the interdomain loops of *Drosophila*(designated D.m.) and human (designated H.s.) PCNA. Gray boxes indicate amino acids identical for both organisms; arrows show amino acids thought essential for interaction of human PCNA with human pol δ. Amino acid 123 is the only one which is both essential and different in *Drosophila*versus human PCNA. B: the \"front\" side of the human PCNA trimer. Amino acids 119-133 of the interdomain loops are highlighted by showing their α-carbon atoms as black spheres. The α-carbon atom of Val123 is shown as a larger gray sphere. :::  ::: Purification of wild-type and site-specifically mutated PCNA ------------------------------------------------------------ Four NH~2~-terminally his-tagged PCNA variants were highly purified; purity for each is shown (Fig. [2](#F2){ref-type="fig"}). First constructs were prepared encoding 1) NH~2~-terminally his-tagged wild-type human PCNA; 2) NH~2~-terminally his-tagged wild-type *Drosophila*PCNA (dPCNA) and two dPCNA derivatives; 3) one in which amino acid 123 was mutated from glutamine to valine (Q123V dPCNA); and 4) the other, in which *Drosophila*amino acids 119-133 were replaced by the corresponding human sequence (dr119-133h dPCNA). Then all four were transformed separately into *E. coli*(strain M15 \[pREP4\]) and respective proteins were expressed. Finally bacteria were lysed and his-tagged proteins were purified using various procedures including Ni^2+^-IDA Sepharose chromatography. The purity of each was determined by SDS-PAGE and is shown as indicated (Fig. [2](#F2){ref-type="fig"}). The identity of wild-type human PCNA was confirmed using mouse monoclonal anti-mammalian PCNA antibody PC10; the identity of wild-type *Drosophila*PCNA was confirmed using affinity purified polyclonal anti-*Drosophila*PCNA antibodies prepared in rabbits \[[@B12]\] (not shown). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **SDS-PAGE analysis of his-tagged PCNA purified from *E. coli*extracts after regulated bacterial expression.**Purification and SDS-PAGE were as described (Experimental Procedures). Lane 1, 0.4 μg wild-type human PCNA was subjected to electrophoresis. Lane 2, 0.8 μg wild-type *Drosophila*PCNA was subjected to electrophoresis. Lane 3, 0.8 μg *Drosophila*PCNA containing valine substituted for glutamine at position 123 was subjected to electrophoresis. Lane 4, 0.45 μg *Drosophila*PCNA containing amino acids 119-133 substituted with the corresponding human PCNA amino acids was subjected to electrophoresis. Migration positions of molecular mass standards are indicated to the right of the figure. :::  ::: Stimulation of calf thymus pol δ activity by highly purified wild-type versus selected mutant PCNA fractions ------------------------------------------------------------------------------------------------------------ Calf thymus pol δ (apparently homogeneous two-subunit form; see \[[@B24]\]) was purified and assayed for polymerase activity in the presence of varying concentrations of both highly purified wild-type and specific mutant PCNA molecules. We showed previously that either calf thymus or human PCNA could be used interchangeably as stimulatory co-factors for calf thymus pol δ \[[@B29]\] (see also \[[@B12],[@B18],[@B19]\]). Assays were performed using poly(dA)-oligo(dT) as described (Experimental Procedures). As can be seen, human PCNA resulted in robust stimulation of calf thymus pol δ; much less stimulation was observed for wild-type *Drosophila*PCNA (Fig. [3](#F3){ref-type="fig"}). Mutation of *Drosophila*PCNA resulted in substantially increased stimulation of calf thymus pol δ; both substitution of a single amino acid (Q123V dPCNA) and replacement of the entire fly interdomain connector loop with corresponding human amino acids (dr119-133h dPCNA) had demonstrable effects. Of note, at relatively high concentrations, *Drosophila*PCNA but with the entire fly interdomain connector loop replaced by corresponding human amino acids (dr119-133h dPCNA) was similarly effective to wild-type human PCNA at stimulating the activity of calf thymus pol δ; however, it was considerably less effective at lower concentrations (Fig. [3](#F3){ref-type="fig"}). This suggests an effect on binding of PCNA to pol δ and/or on mutant PCNA multimerization. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Effect of various purified PCNA fractions on the DNA polymerase activity of calf thymus pol δ.**Calf thymus pol δ was incubated in a reaction mixture as described (see Materials and Methods) for 5 min at room temperature. Each incubation contained 10 ng of pol δ. DNA product synthesized was determined after placing 5-μl aliquots on Whatman DE-81 filters and subsequently washing with a 5% (w/v) solution of Na~2~HPO~4~•12H~2~O. Radioactivity retained on filters was then determined by liquid scintillation counter. Reaction mixtures contained increasing amounts, as indicated on the abscissa, of various PCNA samples, also as indicated. :::  ::: The effects of highly purified wild-type versus selected mutant PCNA fractions on the processivity of incorporation by calf thymus pol δ ---------------------------------------------------------------------------------------------------------------------------------------- To examine further, the stimulation of calf thymus pol δ by both wild-type and specific mutant PCNA molecules, we examined effects on processivity of nucleotide incorporation. Processivity is defined as the number of deoxyribonucleotides incorporated each time a DNA polymerase binds its template-primer. As can be seen, without PCNA (Fig. [4](#F4){ref-type="fig"} lane 1), pol δ is essentially a distributive enzyme incorporating only a few nucleotides as a result of each binding event. With increasing concentrations of wild-type human PCNA (concentrations increasing from right to left as indicated), processivity of incorporation increases dramatically (Fig. [4](#F4){ref-type="fig"} lanes 2--4). This correlates quite closely with the PCNA-mediated activity increase (see Fig. [3](#F3){ref-type="fig"}). Wild-type *Drosophila*PCNA had relatively much less effect on the processivity of calf thymus pol δ (Fig. [4](#F4){ref-type="fig"} lanes 5--7; concentrations again increasing from right to left as indicated). This is also consistent with activity data presented herein (Fig. [3](#F3){ref-type="fig"}) as well as with results reported previously \[[@B12]\]. When mutants of *Drosophila*PCNA were tested, both Q123V dPCNA (Fig. [4](#F4){ref-type="fig"} lanes lanes 8--10; concentrations again increasing from right to left as indicated) and dr119-133h dPCNA (Fig. [4](#F4){ref-type="fig"} lanes lanes 11--13; concentrations again increasing from right to left as indicated), promoted increased pol δ processivities, again consistent with increased activities (Fig. [3](#F3){ref-type="fig"}). Increases were concentration-dependent, also as expected. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Effect of various purified PCNA fractions on the processivity of nucleotide incorporation by calf thymus pol δ.**Incorporation of \[α^32^P\]dTMP by calf thymus pol δ was monitored by standard denaturing PAGE. The substrates used were (dA)~\~500~-(dT)~12--18~as template-primer and \[α-^32^P\]dTTP. Concentrations of PCNA, both wild-type and mutant proteins, are as indicated. h, human; dr, *Drosophila melanogaster.*NH~2~-terminally his-tagged-PCNA fractions are as indicated; wt, wild-type; Q123V, recombinant *Drosophila*PCNA containing a single amino acid, glutamine at position 123, changed to valine; dr119-133h, recombinant *Drosophila*PCNA containing the entire interdomain connector loop (amino acids 119-133) replaced with the corresponding human PCNA amino acids. :::  ::: Stable complex formation among pol δ, ^32^P-labeled oligonucleotide template-primer and highly purified wild-type versus selected mutant PCNA fractions ------------------------------------------------------------------------------------------------------------------------------------------------------- PAGE band mobility shift assays were used to evaluate, in an essentially qualitative manner, the stability of complex formation among calf thymus pol δ, labeled template-primer and highly purified wild-type versus selected mutant PCNA molecules. As can be seen, wild-type *Drosophila*PCNA promoted almost no pol δ•PCNA•template-primer complex formation (Fig. [5](#F5){ref-type="fig"}). In contrast, complex-formation with both *Drosophila*PCNA mutants (Q123V dPCNA and dr119-133h dPCNA) was readily detectable but neither gave results as robust as those seen with wild-type human PCNA (Fig. [5](#F5){ref-type="fig"}). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Effect of various purified PCNA fractions on calf thymus pol δ•PCNA•^32^P-labeled oligonucleotide template-primer complex formation.**Complex formation among pol δ, various purified PCNA fractions and ^32^P-labeled synthetic oligonucleotide template-primers (30-21-mers) was monitored by standard non-denaturing PAGE-band-mobility-shift assays \[32\]. Each incubation contained 10 ng of pol δ, 70 ng of PCNA and 0.1 pmol/reaction (useable 3\'-OH) of annealed template-primer. NH~2~-terminally his-tagged-PCNA fractions are as indicated; wt, wild-type; Q123V, recombinant *Drosophila*PCNA containing a single amino acid, glutamine at position 123, changed to valine; dr119-133h, recombinant *Drosophila*PCNA containing the entire interdomain connector loop (amino acids 119-133) replaced with the corresponding human PCNA amino acids. :::  ::: DNA synthesis beyond chemically defined template base lesions promoted by highly purified wild-type versus selected mutant PCNA fractions ----------------------------------------------------------------------------------------------------------------------------------------- As a final test, we examined the abilities of various PCNA fractions to promote pol δ-dependent DNA synthesis beyond template base lesions (TLS). PCNA-dependent TLS by pol δ was first reported by O\'Day et al. \[[@B30]\] and subsequently analyzed in detail biochemically \[[@B29]\]. The structure of the synthetic oligonucleotide used for evaluation is shown in Fig. [6A](#F6){ref-type="fig"}. For the data shown (Fig. [6B](#F6){ref-type="fig"}), *X*represents the model abasic site (hereafter termed the abasic site \[[@B31]\]) used previously for many of our studies (e.g., see \[[@B29]\]). The mobility of the labeled 21-mer primer, PAGE-purified but without any subsequent enzymatic incubation is shown (Fig. [6B](#F6){ref-type="fig"} lane 1). When calf thymus pol δ alone was added, primer extension opposite the template abasic site was detected but there was no discernible elongation of the resulting 22-mer primer and no full-length product (30-mer) was observed; some degradation of the 21-mer primer, presumably resulting from the activity of the intrinsic pol δ 3\'-5\' exonuclease, was seen (Fig. [6B](#F6){ref-type="fig"} lane 2). Addition to incubations of wild-type *Drosophila*PCNA resulted in slight but readily detectable DNA synthesis beyond the template abasic site; this included some full-length 30-mer (Fig. [6B](#F6){ref-type="fig"} lane 3). Relatively more full-length 30-mer was seen when Q123V mutant *Drosophila*PCNA was included in addition to calf thymus pol δ (Fig. [6B](#F6){ref-type="fig"} lane 4) and still more full-length 30-mer was seen when dr119-133h *Drosophila*PCNA was added (Fig. [6B](#F6){ref-type="fig"} lane 5). Clearly, the greatest amount of full-length 30-mer product was seen when wild-type human PCNA was incubated with calf thymus pol δ (Fig. [6B](#F6){ref-type="fig"} lane 6). Of note, wild-type human PCNA also promotes the tightest complex formation between calf thymus pol δ and ^32^P-labeled template-primer DNA (see Fig. [5](#F5){ref-type="fig"}). ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Effect of various purified PCNA fractions to promote nucleotide incorporation by calf thymus pol δ beyond chemically defined template base lesions.**A: the structure of the 5\'-^32^P-labeled 30-21-mer template-primer; only the primer (21-mer) was radiolabeled and X indicates the position of a modified tetrahydrofuran moiety (model abasic site) on the 30-mer template. B: lane 1, gel-purified primer alone was subjected to electrophoresis; lanes 2--6, incubations were formulated as indicated with the template-primer shown in A followed by standard denaturing PAGE. h, human; dr, *Drosophila melanogaster.*For lanes 2--6, each incubation contained 0.5 pmol of labeled primer (3\'-OH ends) annealed to 0.5 pmol of template (3\'-OH ends), 10 ng pol δ and 70 ng PCNA as indicated. NH~2~-terminally his-tagged-PCNA fractions are as indicated; wt, wild-type; Q123V, recombinant *Drosophila*PCNA containing a single amino acid, glutamine at position 123, changed to valine; dr119-133h, recombinant *Drosophila*PCNA containing the entire interdomain connector loop (amino acids 119-133) replaced with the corresponding human PCNA amino acids. :::  ::: Discussion ========== Although human PCNA and *Drosophila*PCNA are more than 70% identical at the level of primary amino acid sequence, wild-type *Drosophila*PCNA is only a very poor substitute for human PCNA in cell-free reactions with calf thymus pol δ. This is documented both in this report and previously \[[@B12],[@B32]\]. However, mutating only a single *Drosophila*PCNA amino acid, glutamine at position 123 (Q123) to valine (V), leads to a dramatic enhancement in the abilities of *Drosophila*PCNA to stimulate calf thymus pol δ. Effects were shown on total activity (Fig. [3](#F3){ref-type="fig"}), processivity (Fig. [4](#F4){ref-type="fig"}), pol δ•PCNA•template-primer complex formation (Fig. [5](#F5){ref-type="fig"}) and extended DNA synthesis beyond a template abasic site (Fig. [6](#F6){ref-type="fig"}). Replacing the entire interdomain connector loop of *Drosophila*PCNA (amino acids 119-133) with the corresponding residues from human PCNA resulted in additional enhancement (Figs. [3](#F3){ref-type="fig"},[4](#F4){ref-type="fig"},[5](#F5){ref-type="fig"},[6](#F6){ref-type="fig"}), but in neither case were the mutants of *Drosophila*PCNA (Q123V dPCNA or dr119-133h dPCNA) equivalent to wild-type human PCNA in the stimulation of calf thymus pol δ. Our data indicate that although a single *Drosophila*PCNA amino acid at position 123 (in addition to conserved residues 126--128) is very important for pol δ-stimulation, the further enhancement of stimulation seen when the entire interdomain connector loop of *Drosophila*PCNA (amino acids 119-133) was replaced with the corresponding residues from human PCNA suggests that other residues in this loop are also involved directly in binding pol δ. Alternatively, it is possible that loop residues other than 123 and 126--128 play a secondary or indirect (e.g., conformational) role in positioning crucial amino acids so as to optimize their direct binding to pol δ. In this context, we would like to call attention to the fact that at relatively low concentrations, dr119-133h dPCNA is considerably less effective than wild-type human PCNA in stimulating the activity of calf thymus pol δ; at higher concentrations, dr119-133h dPCNA and wild-type human PCNA stimulate calf thymus pol δ similarly. This implies complex protein-protein interactions between PCNA and pol δ such that biochemical properties recorded in dilute solutions *in vitro*may not accurately predict properties manifest at much different and generally much higher intranuclear concentrations present *in vivo*. Alternatively, PCNA must be present as a trimer (three-subunit ring) in order to function. Since the equilibrium among monomer, dimer and trimer was shown to depend on PCNA protein concentration \[[@B33]\], it is certainly possible that the difference observed between dr119-133h dPCNA and wild-type human PCNA actually reflects differences in the K~eq~for PCNA multimerization. These two possibilities, concerning both complicated pol δ•PCNA interactions and PCNA multimerization, are not mutually exclusive. Similarly, the fact that replacement of the entire interdomain connector loop of *Drosophila*PCNA (amino acids 119-133) with the corresponding residues from human PCNA did not result in a molecule as effective in stimulating calf thymus pol δ as human PCNA suggests that regions other than the interdomain connector loop are important for pol δ-stimulation. Our data do not address the question of whether these putative \"other regions\" affect pol δ directly (e.g., like the interdomain loop) or indirectly (e.g., through conformational effects on other regions of the molecule that do bind pol δ directly). Additional mutagenesis studies may shed light on this issue. For example, based on experiments of others, it seems likely that the extreme C-terminus of PCNA also interacts directly with pol δ (see \[[@B23],[@B34]-[@B36]\]). Hence it may be of interest to perform similar mutagenesis experiments to those reported here, focusing instead on the C-terminal region of *Drosophila*PCNA, rather than the interdomain connector loop. We think it should also be noted that both Oku et al. \[[@B35]\] and Ola et al. \[[@B36]\] prepared hybrid proteins between human and *S. cerevisiae*PCNA. As in our studies, Ola et al. \[[@B36]\] found that regions other than the interdomain connector loop of PCNA were important for interaction with pol δ. These authors suggested that additional interacting regions were likely to exist both in the PCNA C-terminus and N-terminus. It may also be of interest to prepare double-mutants, first in the interdomain connector loop of *Drosophila*PCNA, thereby allowing efficient *in vitro*function with purified calf thymus pol δ, and then elsewhere in the PCNA molecule corresponding to interesting sites defined phenotypically by *in vivo*genetic studies of others. For example, it might be possible to determine if particular *mus*209 mutations leading to enhanced mutagen sensitivity among affected organisms (see \[[@B37]\] and references therein) alter any functional interactions between PCNA and pol δ *in vitro*. Results of such studies could lead to novel biochemical insights regarding the mechanism(s) by which point mutations in the *Drosophila*PCNA gene lead to enhanced mutagen sensitivity among animals bearing these mutations. The strategy taken here will presumably allow study of interactions between PCNA and other proteins with which it interacts. In this context, we think it important to note that partial effects on pol δ-stimulation have been recorded. This suggests that our methodology will also allow detection of partial rather than complete effects on the binding of other proteins. Interactions between PCNA and many of the molecules with which it interacts have recently been mapped \[[@B23]\] and for example, one might immediately compare interactions between several mammalian proteins (e.g., human RF-C, DNA ligase I, FEN I and/or p21) and both various wild-type and mutant PCNA molecules described in this paper. Functional (e.g., effects on pol δ activity) as well as direct binding measurements may be made. As with PCNA•pol δ interactions, it may ultimately be feasible to correlate interesting PCNA molecules defined phenotypically using genetic analyses performed in living animals and biochemical studies of specific PCNA•protein binding. For example, do mutagen sensitive *mus*209 animals bear mutations in a region of PCNA responsible for MSH binding? Both MSH3 and MSH6 were reported to possess a consensus motif for binding to the interdomain connector loop of PCNA \[[@B38]\]. Finally, we think it important to note that pol δ has most recently been reported to contain at least four subunits (see e.g., \[[@B39],[@B40]\]) yet all experiments performed here were with the two-subunit form of the enzyme purified from calf thymus. We and others have shown that the larger subunit, p125, is catalytic while the smaller, p50, does not seem to contact the DNA closely (see e.g, \[[@B41]\]), but instead, is required for processivity-stimulation by PCNA (e.g., see \[[@B42]\]) to which it apparently binds. It is also clear that PCNA binds to what has been termed, the third pol δ subunit, p68 or p66 in mammalian systems \[[@B39],[@B43],[@B44]\], Cdc27p in *S. pombe*\[[@B40]\] and Pol32p in *S. cerevisiae*\[[@B45],[@B46]\]. Clearly the physiologically important interaction between PCNA (either mutant or wild-type) and this third pol δ subunit was omitted from our analyses, but could markedly affect any or all of the responses of polymerase to PCNA that we reported here. Conclusions =========== Through our experiments, we showed that *Drosophila*PCNA could be \"humanized\" and that \"humanization\" (mutation of key *Drosophila*residues to human ones) increased effects on mammalian pol δ. The highly purified two-subunit form of pol δ was used for all of our studies. It is possible, though we think it unlikely, that different conclusions would be reached if a different form of pol δ (three-or four-subunit) was used. Nevertheless two of the effects we observed could be considered beneficial. They were enhancement of polymerase activity and processivity. A third effect seems likely to be detrimental, at least over the long term, that is increased synthesis opposite and beyond a chemically defined template base lesion (TLS). Our data suggest that all three of these effects result from enhancement of PCNA-dependent stability of the pol δ•PCNA•template-primer complex. In other words, in the range that we have studied, the more tightly pol δ binds to DNA, the greater its activity, the greater its processivity, but also the more likely it is to catalyze TLS. Our results provide an explicit approach to correlate *in vivo*genetic studies with rigorous *in vitro*biochemistry. Methods ======= Materials --------- Unlabeled deoxyribonucleoside triphosphates (dNTPs) were from Boehringer-Mannheim; \[α-^32^P\]ATP and \[α-^32^P\]dTTP were from Amersham Corp. *E. coli*DNA polymerase I Klenow fragment without 3\'-5\' exonuclease activity (exo-), was expressed and purified according to standard protocols \[[@B47]\]. Terminal deoxynucleotidyl transferase (TdT) was from Sigma. Micrococcal nuclease was from Boehringer-Mannheim. Pfu DNA polymerase was from Stratagene. Ni^2+^-IDA Sepharose was from Pharmacia (Piscataway, NJ). Acrylamide and methylene bis-acrylamide were from Eastman Organic Chemicals and for protein SDS-PAGE, were further purified by adsorption of impurities to activated charcoal. For PAGE of nucleic acids, they were purified by adsorption to an ion exchange resin. All other materials were of reagent grade and were used without additional purification. Proteins -------- PCNA was purified to apparent homogeneity from calf thymus \[[@B17]\] as was pol δ \[[@B24],[@B48]\]. Human PCNA cDNA was cloned into a bacterial expression vector and human PCNA was purified from an *E. coli*extract, also to apparent homogeneity \[[@B18]\]. *D. melanogaster*PCNA was purified to apparent homogeneity identically after bacterial expression \[[@B13]\]. A his-tag was added to the NH~2~-termini of both human and *Drosophila*PCNA by cDNA insertion into pQE30 (Qiagen, Valencia, CA) using *Bam*H1 and *Hind*III restriction endonuclease sites. Nucleic acids ------------- Templates and primers, all of defined sequence, were synthesized conventionally by Dr. F. Johnson and colleagues (Stony Brook). Before use, they were purified by standard denaturing PAGE \[[@B49]\]. All other DNA manipulations were performed according to standard techniques \[[@B49]\]. Methods ------- Much of the methodology was described in detail previously \[[@B12],[@B19],[@B20],[@B24],[@B29],[@B32],[@B41],[@B50],[@B51]\]. SDS-PAGE was according to Laemmli \[[@B52]\] as modified \[[@B53]\] on minigels or as reported previously \[[@B54]\]. For immunoblots, proteins were transferred electrophoretically to nitrocellulose \[[@B55]\] and resulting replicas were probed with antibodies. Reactivity was visualized colorimetrically \[[@B56]\] with alkaline phosphatase-conjugated goat anti-IgG antibodies \[[@B57],[@B58]\] and a one-solution phosphatase substrate (Kirkegaard and Perry, Gaithersburg, MD). Immunologic detection of human PCNA was with mouse monoclonal antibody (mAb) PC10 (Oncogene Sciences, Uniondale, NY). Detection of *Drosophila*PCNA was with affinity purified polyclonal rabbit anti-*Drosophila*PCNA antibodies \[[@B12]\]. Restriction endonucleases were from Boehringer (Indianapolis, IN) and were used according to the vendor\'s instructions. DNA sequencing performed in both directions was according to Sanger et al. \[[@B59]\] using a fluorescence-based method and an ABI 373 (Applied Biosystems, Foster City, CA) automated DNA sequencer. Site-directed mutagenesis of *Drosophila* PCNA ---------------------------------------------- Site-directed mutagenesis of NH~2~-terminally his-tagged *Drosophila*PCNA was performed exactly as described \[[@B20]\] to generate either the Q123V protein or chimeric molecules containing the entire *Drosophila*PCNA sequence except for amino acids 119-133 which were replaced by the corresponding residues from human PCNA. Purification of his-tagged PCNA ------------------------------- Purification of his-tagged PCNA to apparent homogeneity was performed exactly as previously described \[[@B20]\]. Characterization was by SDS-PAGE (Fig. [2](#F2){ref-type="fig"}) and immunoblot analysis. DNA polymerase δ incubations ---------------------------- Assays of pol δ on synthetic oligonucleotide template-primers were performed essentially as previously described \[[@B24]\]. Primers were 5\' end-labeled with T4 polynucleotide kinase in the presence of \[γ-^32^P\]ATP. Afterward, labeled primer was annealed to an unlabeled template. The standard reaction mixture for pol δ contained 40 mM Bis-Tris, pH 6.7, 6 mM MgCl~2~, 1 mM dithiothreitol, 10% glycerol and 40 μg/ml bovine serum albumin. Additional details are provided in the figure legends. Incubations were terminated by addition of standard stop solution and aliquots were subjected to 12% PAGE in the presence of 7 M urea and 15% formamide. After electrophoresis, gels were subjected to autoradiography and/or Molecular Dynamics 445 SI PhosphorImager analyses. Pol δ processivity ------------------ Processivity was evaluated qualitatively using (dA)~\~500~annealed to (dT)~12--18~(both from Pharmacia) in a final volume of 5 μl containing 6 nmol poly(dA) (nucleotide), 0.2 nmol (dT)~12--18~(nucleotide), 10 μM dTTP, 100 μCi \[α-^32^P\]dTTP, 40 mM Bis-Tris, pH 6.7, 6 mM MgCl, 1 mM dithiothreitol, 10% glycerol, 40 μg/ml bovine serum albumin, 10 ng of highly purified pol δ and various quantities of different PCNA samples as indicated. Assays were for 5 min at room temperature and were stopped by addition of standard PAGE stop solution and PAGE in the presence of 7 M urea. After electrophoresis, gels were subjected to autoradiography and/or Molecular Dynamics 445 SI PhosphorImager analyses. Nondenaturing PAGE band mobility shift assays --------------------------------------------- Nondenaturing PAGE band mobility shift assays were performed essentially as previously described \[[@B32]\] but without MgCl~2~and otherwise as detailed in the figure legend. EDTA was included in each incubation and in the gel electrophoresis buffer at a final concentration of 3 mM. Authors\' contributions ======================= DJuM performed all enzymologic and mobility shift assays with DNA polymerase δ in combination with both wild-type and various mutant PCNA molecules. He also designed, engineered and characterized all recombinant PCNA molecules. DJuM expressed several recombinant proteins in bacteria and purified them. Finally, he participated in DNA polymerase purification and drafted the original manuscript. MM expressed some recombinant proteins in bacteria and purified them. She also purified and characterized most DNA polymerase substrates. HM participated in DNA polymerase purification and manuscript preparation. PAF advised DJuM on execution and interpretation of experiments and assisted both in figure design and all other aspects of manuscript preparation. All authors read and approved the final manuscript. Acknowledgements ================ These studies were supported by NIH Research Grant ES04068.

PubMed Central

2024-06-05T03:55:47.308741

2004-8-13

PMC515295

Background ========== Guanine nucleotide exchange factors (GNEFs) are components of signaling pathways that link transmembrane receptors to intracellular GTPase family members regulating a wide variety of cellular functions such as proliferation, differentiation, adhesion and apoptosis. C3G (RapGEF1) is an ubiquitously expressed GNEF for Ras family proteins that particularly targets Rap1, Rap2 and R-Ras \[[@B1]-[@B4]\]. It has been shown to mediate signals received from B and T cell receptor activation, growth factors, cytokines, G protein coupled receptors and also adhesion \[[@B5]-[@B15]\]. C3G is present in the cytoplasm in a complex with members of the Crk family of small adapter molecules. In response to stimuli, this complex is recruited to the cell membrane involving association of Crk with phosphotyrosine containing proteins like receptor tyrosine kinases, p130 Cas, IRS-1 and paxillin \[[@B16]-[@B18]\]. Following translocation from cytosol to cell membrane, C3G activates downstream signaling. Its activation has been shown to lead to an activation of mitogen activated protein kinase and Jun N-terminal kinase \[[@B9],[@B12],[@B19]-[@B21]\]. Studies involving overexpression of membrane targeted C3G or dominant negative forms have shown that C3G is involved in both growth suppression as well as transformation \[[@B22]-[@B24]\]. C3G appears to play an important role in mammalian development because C3G-/- mice die before embryonic day 7.5. These studies have shown that C3G is required for vascular myogenesis and for cell adhesion and spreading \[[@B25],[@B26]\]. The C-terminus of C3G, which shows homology to CDC25, harbors the catalytic domain. The central region of C3G, which spans about 300 residues, has polyproline tracts with the ability to bind to SH3 domains of various proteins like Crk, p130 Cas, Grb2 and Hck \[[@B1],[@B2],[@B9],[@B18],[@B27]\]. No function has particularly been attributed to the N-terminal sequences, which do not show homology to any defined protein sequences. The non-catalytic domain of C3G has been shown to negatively regulate its catalytic activity. Deletion of the N-terminal sequences or its association through its proline sequences to Crk leads to its activation \[[@B16]\]. Integrin mediated cell adhesion causes tyrosine phosphorylation of C3G \[[@B28]\]. It has been shown that overexpression of c-Crk1 or stimulation of cells with growth hormone leads to specific phosphorylation of Y504 \[[@B21],[@B29]\]. This modification results in an increase in C3G catalytic activity towards Rap1. Src and JAK have been implicated in Y504 phosphorylation of C3G. More recently we have used site -- specific antibodies to show that the activation of Src family kinase Hck, leads to C3G phosphorylation on Y504 suggesting that Src family kinases can directly regulate C3G activity and function \[[@B27]\]. The effectiveness and precision of intracellular signal transduction depends on protein-protein interactions that regulate enzyme activity as well as subcellular localization. Cell surface receptor activation leads to assembly of adaptor protein complexes at the plasma membrane, which serve to localize guanine nucleotide exchange proteins. Earlier, both endogenous as well as exogenously expressed C3G has been shown to localize to the cytoplasm and not to associate with plasma membrane \[[@B22],[@B30]\]. Since activation of C3G occurs primarily through phosphorylation at Tyr 504 and membrane recruitment, we undertook a detailed study of the subcellular localization of both exogenously expressed and endogenous Y504 phosphorylated C3G (pY504 C3G). Expression of Src family kinases or pervanadate treatment of cells, which mimics stimulation by growth factors, resulted in marked tyrosine phosphorylation of C3G at Y504. Conventional as well as optical sectioning microscopy revealed that pY504 C3G was predominantly located at the Golgi complex and the subcortical actin cytoskeleton unlike non-phosphorylated C3G, which was largely cytosolic. Results ======= Colocalization of C3G with Hck ------------------------------ We have recently shown that Hck interacts with and phosphorylates C3G in vivo and we wished to determine if their interaction leads to changes in the subcellular distribution of C3G and whether pY504 C3G locates to specific subcellular domains. Cos-1 cells were transfected with C3G in the presence or absence of Hck and immunostained using anti C3G antibodies. As shown in Fig. [1A](#F1){ref-type="fig"} in a majority of cells, exogenously expressed C3G showed diffuse cytoplasmic staining that extended up to the plasma membrane. Variation in the level of C3G was observed among the transfected cells with weakly expressing cells showing a more prominent juxtanuclear staining. When cotransfected with Hck, most cells showed prominent staining of C3G at the plasma membrane and a juxtanuclear organelle in addition to the diffuse cytoplasmic staining. This pattern appeared similar to that seen for exogenously expressed Hck and therefore we performed colocalization studies to confirm their distribution. When coexpressed, Hck and C3G are targeted predominantly to the plasma membrane and other intracellular membranous structures (Fig. [1B](#F1){ref-type="fig"}). Merged images show that these two proteins colocalize in the subcellular context. Similar patterns of colocalization were observed when C3G and Hck were expressed in HeLa cells (data not shown). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Subcellular localization of C3G.**(A) Cos-1 cells grown on coverslip were either transfected with C3G or cotransfected with Hck and indirect immunofluorescence staining performed using anti-C3G antibodies and Cy3 conjugated anti rabbit secondaries. (B) Cells transfected with Hck and C3G were stained for both the antigens as described in Materials and Methods. Hck was visualized using FITC conjugated secondaries and C3G by Cy3 conjugated secondaries. The dual panel shows the merged image of an optical section taken using the confocal microscope where the yellow signal generated shows colocalization of the two proteins. :::  ::: Src family kinases phosphorylate C3G and phospho-C3G localizes to the Golgi and plasma membrane ----------------------------------------------------------------------------------------------- To determine the subcellular distribution of phospho-C3G, which is known to be the activated form, specificity of a rabbit polyclonal phosphorylation site-specific antibody (pY504-C3G) was verified by examining its reactivity using cell lysates expressing C3G or Y504F-C3G alone or with Hck. As shown in Fig. [2A](#F2){ref-type="fig"}, pY504-C3G antibody recognizes only C3G when coexpressed with Hck. Neither the C3G protein expressed in itself nor the Y504F mutant coexpressed with Hck show any reactivity with this antibody suggesting that it reacts only with Y504 phosphorylated C3G. Phosphotyrosine blotting showed that a large number of cellular polypeptides are phosphorylated on tyrosine upon Hck expression, (Fig. [2A](#F2){ref-type="fig"}, right panel) but except for C3G none of the others show any reactivity with pY504 antibody. Y504F mutant of C3G, which shows low level of phosphorylation on other tyrosine residues is not detected by this antibody indicating its specificity towards Y504 phosphorylated C3G. Unlike Hck, whose expression is restricted to a subclass of hematopoietic cells, C3G is ubiquitously expressed and we wished to determine if other Src family kinases could phosphorylate C3G. We coexpressed C3G with an expression construct for the fusion protein c-Src-GFP and western blotting was performed using pY504-C3G antibody. As shown in Fig. [2B](#F2){ref-type="fig"}, in vivo, c-Src was also able to induce Y504 phosphorylation of C3G, but not that of the Y504F C3G mutant. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Specificity of phosphospecific antibody, and phosphorylation of C3G on Y504 upon coexpression with Src family kinases.**Cos-1 cells were transfected with the expression constructs for Hck (A) or c-Src (B) along with C3G as indicated and western blotting of whole cell lysates was performed using the phosphospecific antibody pY504. The blots were reprobed with C3G, Hck and anti pTyr (Panel A) or Src (Panel B) to show their expression in the lysates. pY504 C3G and pTyr was detected by ECL and C3G, Hck and Src by alkaline phosphatase dependent color development. UT indicates untransfected cell lysates. Y504F is a mutant of C3G in which tyrosine 504 is replaced by phenylalanine. :::  ::: We wished to determine whether C3G that colocalizes with Hck was the phosphorylated component and therefore used pY504-C3G antibodies to determine the localization of pY504 C3G in cells expressing Hck and C3G. As shown in Fig. [3A](#F3){ref-type="fig"}, pY504 C3G showed a staining pattern that exactly matched that of Hck with prominent staining of the plasma membrane, a juxtanuclear organelle and other intracellular membranes (3A). The prominent staining appeared to correspond with the Golgi structure and Hck has earlier been shown to localize to the Golgi \[[@B31]\]. In cells transfected with C3G and Hck, the pattern of pY504 C3G staining was also compared with that of total C3G as detected by the Flag tag antibody. As shown in Fig. [3B](#F3){ref-type="fig"}, it was observed by confocal analysis that the tag antibody detects the presence of C3G spread throughout the cytoplasm with some prominence in the juxtanuclear region and plasma membrane. The pattern suggests that the majority of the protein is cytosolic. In contrast staining for phospho-C3G was non-uniform and was particularly prominent at the Golgi and cell membrane. Colocalization of C3G with that of phospho-C3G is seen at the plasma membrane and in the juxtanuclear region. This also suggests that only a proportion of the expressed C3G is phosphorylated at Tyr504. To confirm the presence of pY504 C3G in the Golgi, we coexpressed the viral protein, VSVG-GFP known to localize to the Golgi with Hck and C3G and observed the staining pattern of pY504 C3G and that of GFP. VSVG-GFP locates predominantly at the Golgi, trans-Golgi network and also the endoplasmic reticulum and plasma membrane in a temperature-dependent manner \[[@B32]\]. As shown in Fig. [3C](#F3){ref-type="fig"}, the yellow signal generated in the dual image showed colocalization of pY504 C3G with VSVG-GFP suggesting that pY504 C3G was predominantly targeted to the Golgi complex. Unlike C3G, pY504 C3G appeared to be restricted to the plasma membrane and other intracellular membranes with particular concentration in the Golgi. When overexpressed, a large amount of C3G was present in the cytosol and we wished to determine whether cytosolic C3G does not get phosphorylated upon Hck coexpression or whether pY504 C3G in the cytosol is transient due to the action of tyrosine phosphatases. Cos-1 and HeLa cells transfected with Hck and C3G were either left untreated, or, subjected to pervanadate treatment for 10 minutes prior to fixation and stained for pY504 C3G. Pervanadate is a strong inhibitor of tyrosine phosphatases; therefore treatment of cells with pervanadate results in dramatic augmentation of tyrosine phosphorylation on cellular proteins \[[@B33]\]. As shown in Fig. [3D](#F3){ref-type="fig"}, pervanadate-treated cells showed an increase in the pY504 C3G staining in the cytoplasm suggesting that it was dephosphorylated by cytosolic tyrosine phosphatases. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **pY504-C3G colocalizes with Hck and shows predominant Golgi and membrane localization.**(A) pY504 C3G colocalizes with Hck. Cos-1 cells transfected with Hck and C3G were stained for pY504 C3G (Cy3) and Hck (FITC) and examined using a confocal microscope. Figure shows an optical section for the individual stains as well as that of the merged (Dual) image. (B) Cos-1 cells transfected with Hck and C3G were dual labeled to detect phospho-C3G (Cy3 staining) and C3G using the Flag tag antibody (FITC staining). Panels show optical sections taken using the confocal microscope. (C) pY504 C3G is localized to the Golgi apparatus. Cos-1 cells were transfected with Hck, C3G and VSVG-GFP expression constructs and stained using pY504 primary antibody and Cy3 conjugated secondary. An optical section taken using the apotome is represented. (D) HeLa or Cos-1 cells transfected with Hck and C3G were left untreated (control) or treated with pervanadate (PV) prior to fixation and stained for pY504. Counter staining with Dapi shows cell nuclei. :::  ::: Phosphorylation of endogenous C3G and its localization to the Golgi and subcortical actin cytoskeleton ------------------------------------------------------------------------------------------------------ In the above experiments phosphorylation and localization of C3G was studied using exogenously expressed protein and we wished to determine whether endogenous C3G could be phosphorylated and similarly targeted. Towards this end we checked the phosphorylation of endogenous C3G under conditions of Src and Hck overexpression or upon activation of cellular tyrosine kinases by pervanadate treatment. C3G protein is expressed as a doublet of about 140--150 kDa, which are products of two differentially spliced mRNAs \[[@B34]\]. Whole cell lysates were prepared from Cos-1 cells and those transfected with Hck or Src and western blotting performed using pY504 antibodies. As shown in Fig. [4A](#F4){ref-type="fig"}, overexpression of Src or Hck induces tyrosine 504 phosphorylation of endogenous C3G. The same blot was reprobed with C3G, Src and Hck antibodies to show their presence in the lysates. We examined the localization of endogenous phosphorylated C3G after c-Src expression and found that similar to the phosphorylated form of the exogenously expressed C3G, endogenous pY504 C3G was present predominantly at sites of c-Src localization. Intense staining of the Golgi and cell membranes was evident and merged images show colocalization of the two proteins (4B). Cells that did not express Src, did not show any phosphorylated C3G. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Phosphorylation of endogenous C3G upon overexpression of Hck and its localization to the Golgi.**(A) Cos-1 cells were transfected with expression constructs as indicated and whole cell lysates used in western blotting for pY504-C3G. ECL was used for detection. Blots were reprobed to show expression of C3G and the kinases. (B) Endogenous pY504 C3G colocalizes with c-Src. Cos-1 cells were transfected with the c-Src GFP fusion protein vector and cells stained for pY504-C3G expression (Cy3). c-Src expression was visualized as GFP fluorescence. Images shown are optical sections taken using the apotome. (C) Endogenous pY504-C3G localizes to the Golgi. Cos-1 cells were transfected with Hck along with VSVG-GFP and stained for pY504 C3G. Cells were left untreated (control) or treated with nocodazole (Noc) prior to fixation as described in Methods. Panels show optical section for pY504 by Cy3 and the VSVG-GFP by GFP fluorescence. :::  ::: The localization of endogenous pY504 C3G to the Golgi was examined in Cos-1 cells transfected with Hck and VSVG-GFP. Immunostaining for pY504 C3G was seen predominantly at the cell periphery and the Golgi (Fig. [4C](#F4){ref-type="fig"}), which was confirmed by colocalization with VSVG-GFP. The effect of Golgi perturbing drugs on the localization of pY504 C3G was examined by treatment of cells with nocadazole for depolymerization of microtubules and concomitant Golgi fragmentation. Under these conditions pY504 C3G was detected as dispersed vesicles scattered in the cytoplasm and remained colocalized with VSVG-GFP (4C) confirming that endogenous pY504 C3G localized to the Golgi complex. The localization of pY504 C3G formed by the activation of endogenous tyrosine kinases was determined in cells treated with pervanadate, which is known to activate Src family kinases, in addition to inhibiting tyrosine phosphatases \[[@B35]-[@B37]\]. Pervanadate treatment results in the dramatic augmentation of phosphorylation of a large number of cellular proteins on tyrosine and therefore mimics activation of signaling pathways by growth factors \[[@B38]\]. While normal HeLa cells do not show any pY504 C3G, cells treated with pervanadate showed distinct presence of pY504 C3G in whole cell lysates as seen by western blotting (5A). The large number of other cellular proteins phosphorylated on tyrosine (seen upon blotting with antiphosphotyrosine antibodies), as a consequence of pervanadate treatment, do not show reactivity with pY504 antibody. In order to confirm that the signal observed upon pervanadate treatment was specific to phospho Y504-C3G, Cos-1 cells were transfected with either C3G or Y504F mutant of C3G. They were left untreated, or treated for 10 minutes with pervanadate and indirect immunofluorescence performed to observe expression of the wild type or mutant proteins as well as that of phospho-C3G. C3G, and Y504FC3G expression was monitored by staining for Flag and His tags respectively. As observed in Fig. [5B](#F5){ref-type="fig"} only cells expressing C3G showed intense staining for phosphoC3G while Y504F expressing cells showed no enhanced signal above that of the other non-expressing cells in the field. These results reaffirmed the specificity of the phospho-C3G antibody in detecting only Y504 phosphorylated C3G. Phosphorylated endogenous C3G staining was seen weakly in the non-expressing cells upon pervanadate treatment. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Phosphorylation of endogenous C3G upon activation of endogenous tyrosine kinases.**(A) Cells were either left untreated (UT) or treated with pervanadate (PV) and western blotting was performed using pY504 antibody. The same blot was reprobed with C3G to show the presence of endogenous C3G in these cells. (B) Cos-1 cells on coverslips were transfected with either C3G or Y504F mutant of C3G and fixed without any treatment (cont.) or after pervanadate treatment (PV). Dual labeling was performed using the tag antibodies (stained with FITC) and pY504 antibody (stained with Cy3). Panels show optical sections obtained by confocal microscopy. (C) Cos-1 and HeLa cells grown on coverslips and transfected with VSVG-GFP were fixed without any treatment (control) or after treatment with pervanadate and stained for pY504 expression. GFP fluorescence was used to visualize the staining pattern of VSVG-GFP protein. Optical sections taken using the apotome are shown. Areas of colocalization are seen from the yellow color generated in the merged images. :::  ::: Indirect immunoflourescence was performed to determine the localization of endogenous pY504 C3G formed by the activation of intracellular tyrosine kinases. Cos-1 and HeLa cells were stimulated by pervanadate and as seen in Fig. [5C](#F5){ref-type="fig"}, pY504 C3G staining, which is evident only in the treated cells, localized at the cell periphery and the Golgi. Colocalization with VSVG confirmed its localization to the Golgi complex. The staining at the cell periphery appeared to match that of the subcortical actin cytoskeleton. In order to confirm this, we dual stained the cells treated with pervanadate for F-actin and found that the pY504 C3G seen at the cell periphery colocalizes with F-actin suggesting that pY504 C3G is targeted to the subcortical actin cytoskeleton upon activation of endogenous tyrosine kinases by pervanadate (Fig. [6A](#F6){ref-type="fig"}). It was also observed that pY504 C3G staining at the cell periphery was particularly prominent in confluent cells compared to cells that were sparsely growing in isolation suggesting that pY504 C3G is particularly enriched along cell-cell junctions. Phospho C3G also shows partial colocalization with filamentous actin known to be associated with the Golgi complex \[[@B39]\]. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **pY504C3G localizes to the subcortical actin cytoskeleton**. (A) HeLa cells grown on coverslips were left untreated or treated with pervanadate and stained for pY504 expression using Cy3 secondaries. The coverslips were then stained with Oregon-green phalloidin to detect F-actin. (B) C3G phosphorylation requires the activity of Src family kinases and the presence of an intact cytoskeleton. HeLa cells were pretreated with PP2 or cytochalasin D as described in methods prior to pervanadate treatment. Images show the localization of pY504 C3G labeled with Cy3 and F-actin stained with oregon green. Images shown are a single optical section visualized using the apotome. :::  ::: We have observed that pervanadate treatment increased tyrosine phosphorylation of endogenous C3G. Since overexpression of Src as well as Hck results in phosphorylation of C3G, it was of interest to determine whether pervanadate induced tyrosine phosphorylation of C3G was mediated by Src family kinases within the cell. Cells were treated with PP2, a specific SFK inhibitor prior to PV treatment \[[@B40]\]. As shown in Fig. [6B](#F6){ref-type="fig"}, pY504 C3G staining of cells stimulated with pervanadate was considerably reduced, but not totally abolished when they were pretreated with PP2 suggesting the possibility of other tyrosine kinase family members activated by pervanadate contributing to C3G phosphorylation. To determine whether the increase in pY504 C3G staining was dependent on the presence of an intact cytoskeleton, we observed its localization in cells treated with cytochalasin D, a reagent that effectively disrupts the actin cytoskeletal network. Under these conditions there is a collapse of cell morphology and F-actin staining shows an irregular distribution at the cell cortex. As observed in Fig. [6B](#F6){ref-type="fig"}, the staining for pY504 C3G in the subcortical cytoskeleton was largely absent under conditions of moderate disruption of actin organization. Under these conditions, pY504 C3G staining at the Golgi complex, which shows a more dispersed morphology appeared not to be affected. Discussion ========== C3G is involved in a variety of signaling pathways and therefore its dynamic localization under normal and activated situations may be physiologically relevant. In this study we demonstrate the limited subcellular distribution of Y504 phosphorylated C3G, which is predominantly targeted to the Golgi apparatus and the subcortical actin cytoskeleton. This localization has been substantiated by colocalization with a Golgi marker protein and F-actin respectively. Rap1, the substrate of C3G has been localized to the Golgi, lysosomal vesicles and cortical actin cytoskeleton \[[@B41]\]. But, of the at least eight known exchange factors for Rap1, C3G is the only one that has been linked definitively to the tyrosine kinase signaling pathway. Src family members like Src and Hck have been shown to localize to the plasma membrane and other intracellular membranes with particular concentration in the Golgi \[[@B31],[@B42]\]. When endogenous C3G was phosphorylated by overexpressed Hck or Src, the localization of pY504 C3G matched that of the kinases suggesting that they may be part of the same molecular complexes. This is also evident from the staining pattern of pY504 C3G when C3G is expressed along with Hck, which is distinctly seen in the Golgi and plasma membrane. Since exogenously expressed C3G is predominantly cytosolic, it implies that, at any given time, only a small fraction of it is phosphorylated at Y504 at the plasma membrane and the Golgi. We observed more exogenously expressed pY504 C3G in the cytoplasmic compartment under conditions of inhibition of tyrosine phosphatases suggesting that pY504 C3G may be targeted by cytosolic tyrosine phosphatases. This regulation may help in restricting the activity of C3G to specific compartments. We observe very little endogenous pY504 C3G in the cytosol when HeLa or Cos-1 cells are treated with pervanadate, which not only inactivates tyrosine phosphatases, but also activates tyrosine kinases. It is possible that upon PV treatment endogenous C3G present in the cells is phosphorylated at the sites of location of the activated kinases. Pervanadate treatment has been shown to increase phosphotyrosine staining at the cell periphery indicating activation of kinases present in this subcellular domain \[[@B43]\]. Recently c-Src and Jak2 have been implicated in the phosphorylation of C3G in response to growth hormone stimulation of NIH 3T3 cells because dominant negative mutants of these kinases inhibit C3G phosphorylation \[[@B21]\]. It was suggested that this phosphorylation of endogenous C3G by c-Src occurs at Y504 because exogenously expressed Y504F mutant of C3G was not phosphorylated. Using a phosphospecific antibody we have directly shown the phosphorylation of endogenous C3G at Y504 upon overexpression of Hck \[[@B27]\] and c-Src (this report). c-Src is present in Cos-1 and HeLa cells which lack Hck. c-Src localizes to the cell membrane, focal adhesions and also to the Golgi \[[@B42],[@B44]\]. Since pervanadate is a good activator of Src (36,37), it is possible that pY504 C3G seen in pervanadate treated cells is because of C3G being a Src substrate. Adhesion dependent Src activation leads to Rap-1 activation mediated by Crk and C3G \[[@B14]\]. Fibroblasts lacking C3G are essentially compromised in adhesion-mediated responses \[[@B25],[@B26]\]. The localization of endogenous pY504 C3G at the subcortical actin cytoskeleton therefore suggests that this may be the site of action of C3G in mediating responses to cell adhesion. Modification of C3G by phosphorylation at defined subcellular domains may be important for restricted activation of C3G mediated signaling functions in the cells. Close structural and functional relationship is known to exist between the structural elements at the cell periphery and the signal transduction machinery. Several tyrosine kinases are known to be located in adherence junctions and the kinetics of phosphorylation and dephosphorylation appears to be controlled by structural molecules at the junctions. Our observation that disruption of actin cytoskeleton results in a loss of pY504 C3G staining at the cell periphery, but not at the Golgi complex reveals an important role for cytoskeletal network in the regulation of C3G. Conclusions =========== The activity of guanine nucleotide exchange factor C3G is known to be regulated by tyrosine phosphorylation and membrane targeting. Using phospho-specific antibodies, we directly demonstrate that expression of Src family kinases or pervanadate treatment of cells induces phosphorylation of C3G on Y504. Unlike C3G, which is mostly cytosolic, pY504C3G locates to the Golgi and subcortical actin cytoskeleton. Demonstration of the localization of the active component of C3G to the Golgi and subcortical cytoskeleton provides evidence for a possible function for C3G at these cellular compartments. Methods ======= Cell culture and treatment of cells ----------------------------------- HeLa and Cos-1 cells were cultured in DMEM supplemented with 10% FCS. Transfections were performed on cells grown as a monolayer in either 35 mm dishes or glass coverslips using the cationic lipid DHDEAB as described \[[@B45]\]. Briefly, 1 μl lipid diluted in 50 μl serum free DMEM was mixed with 1 μg DNA in 50 μl serum free DMEM. The mix was kept at room temperature for 30 min to allow complex formation before adding to the cell monolayer. Cells were fed with serum 5 hrs later and harvested 24--30 hrs after transfection. Cells were subjected to pervanadate treatment by the addition of a freshly prepared solution of pervanadate at 50 μM conc. for 10 min prior to harvesting. Pervanadate stock solution (50 mM) was prepared by mixing equal volumes of 100 mM solution of H~2~O~2~, with 100 mM solution of sodium orthovanadate. It was added to the cells within 5 mins of preparation. Golgi disruption was performed by treating the cells with 5 μg/ml of nocadazole for 30 min prior to fixation. To disrupt actin cytoskeleton, cells were treated with 1 μg/ml cytochalasin D for 20 mins. PP2 was added to cells 2 hrs before pervanadate treatment at a concentration of 10 μM to inhibit Src family kinases. Expression constructs --------------------- Full length human C3G cloned in pcDNA3-FLAG was kindly provided by Dr S Tanaka. Y504F mutant of C3G in which tyrosine 504 is mutated to phenylalanine cloned in a His tagged expression vector was provided by Dr M Matsuda. The wild type rat p59 Hck cDNA was cloned in the pCI plasmid (Promega) and has been described earlier \[[@B27]\]. Expression plasmid for vesicular stomatitis virus glycoprotein as a GFP fusion protein (VSVG-GFP) was a kind gift from Jennifer Lippincott-Schwartz \[[@B32]\]. c-Src-GFP expression vector expressing wild type c-Src fused to GFP at C-terminal was from Dr D L Anders \[[@B46]\]. Wild type human Hck cDNA cloned into pCDNA6 expression vector was a kind gift of Dr Todd Miller \[[@B47]\]. Western blotting ---------------- Whole cell lysates were prepared by lysing cells directly in Laemli\'s sample buffer and subjected to SDS-polyacrylamide gel electrophoresis. After transfer onto nitrocellulose membranes, they were processed for western blotting using the required primary antibodies. Detection was based on either color development using alkaline phosphatase conjugated secondary antibodies or on chemiluminescence using horse radish peroxidase conjugated secondaries. Indirect immunoflourescence and microscopy ------------------------------------------ Cells were processed for immunoflourescence staining as described earlier \[[@B27]\]. The primary antibodies used were rabbit polyclonal anti-C3G (Santa Cruz Biotechnology), rabbit polyclonal anti pY504-C3G (SC-12926 R from Santa Cruz) and anti-Hck (3E9 monoclonal) made in our laboratory \[[@B48]\]. Dual labeling for Hck and C3G was performed by incubating the cells serially with C3G antibody, anti rabbit Cy3, monoclonal anti-Hck, and anti-mouse FITC. Cells were incubated with Oregon-green phalloidin after staining for pY504 with Cy3 to visualize F-actin. Cells transfected with vectors encoding GFP fusion proteins (GFP-Src or VSVG) were observed directly by fluorescent microscopy. Dual labeling for the C3G constructs and phospho-C3G was performed using the corresponding monoclonal tag antibodies (detected by FITC) and pY504 antibody (detected by Cy3). C3G was detected using Flag tag antibody (from Sigma) and Y504FC3G by His tag antibody (from Qiagen). Cells were examined using an Olympus microscope equipped with a cool SNAP color CCD camera. Images were captured using Image Pro Plus software. Immunoflurescence staining and colocalization was also observed using a Zeiss Axioplan 2 microscope fitted with an Apotome. The apotome (from Carl Zeiss Microimaging) is a new 3D imaging system for contrast enhancement in fluorescence microscopy. It uses structured illumination to reject signals belonging to regions of the sample that are outside the best focus position of the microscope. Images were captured using the Axiocam (Zeiss) CCD camera and processed using the Axiovision 4 software. Colocalization was also determined by observing the staining patterns using the LSM 510 Meta confocal microscope from Carl Zeiss. Abbreviations ============= GNEF -- guanine nucleotide exchange factor SFK -- Src family kinase PV -- Pervanadate pY504 C3G -- Tyrosine 504 phosphorylated C3G VSVG -- Vesicular stomatitis virus glycoprotein DMEM -- Dulbecco\'s modified Eagle\'s medium FITC -- Fluorescein isothiocyanate Authors contributions ===================== VR designed and carried out the experiments, analysed the data and drafted the manuscript. GS helped with designing the experiments, analyzing the data and writing the manuscript. AR provided technical help for western blotting and indirect immunoflourescence experiments. Acknowledgement =============== We wish to acknowledge Ms Nandini R for help with confocal microscopy.

PubMed Central

2024-06-05T03:55:47.311471

2004-8-20

PMC515296

Background ========== The shorebirds and allies (Aves: Charadriiformes; \[[@B1]\]) present an exceptional group for studying numerous evolutionary hypotheses. Their remarkable diversity of social mating system, parental care, sexual dimorphism, ecology and life-history make them an ideal group for unravelling the mechanisms of, for example, sexual selection and sexual conflict. Previous comparative studies have made significant contributions to our understanding of the evolution of mating systems \[[@B2]\], parental care \[[@B3],[@B4]\], sexual size dimorphism \[[@B5]-[@B7]\], locomotion and morphology \[[@B8]\], migratory behaviour \[[@B9]\], egg size \[[@B10]\], and plumage colouration \[[@B11]\]. The importance of phylogeny in cross-species comparative studies is well documented \[[@B12]-[@B14]\]. Large and well-resolved phylogenies that incorporate divergence times provide powerful tests of a wide range of hypotheses whilst accounting for the effects of shared evolutionary history \[[@B13],[@B15]\]. However, the shorebird studies listed above were limited by the lack of a complete phylogeny for the group. Most of these studies are based on derivations of the seminal work of Sibley and Ahlquist \[[@B16]\], yet this study included less than a quarter of extant and recently extinct shorebird species. Recently extinct taxa (according to Monroe and Sibley \[[@B1]\]) are: the Tahitian sandpiper *Prosobonia leucoptera*, the Canary Islands oystercatcher *Haematopus maedewaldoi*, and the Great auk *Pinguinus impennis*. Recent molecular studies covering a wide range of shorebird families have drawn attention to conflict in the reconstruction of the deep basal nodes of shorebird phylogeny (figure [1](#F1){ref-type="fig"}; reviewed by van Tuinen *et al.*\[[@B17]\]). For example, morphological data \[[@B18],[@B19]\] places Alcinae (auks, puffins, murres) at the base of the shorebird tree whilst sequence \[[@B20]-[@B22]\] and DNA-DNA hybridisation \[[@B16]\] data suggests that they are a highly derived sister group to Stercorariini (skuas and jaegers), Larini (gulls), Sternini (terns), and Rynchopini (skimmers). It is important to note that taxon coverage differs between these studies and this may be an important factor in determining the tree topology. Specific phylogenies have been derived, for example, for sandpipers \[[@B23]\], the genus *Charadrius*\[[@B24]\], and jacanas \[[@B25]\] using DNA sequence data. In contrast, morphological evidence provided the basis for Chu\'s \[[@B26]\] study of gull phylogeny. Strauch \[[@B18]\] presented the most complete data set of 227 Charadriiformes species. However, despite the plethora of cladograms for particular shorebird groups (see reviews by Sibley and Ahlquist \[[@B16]\]; Thomas *et al.*\[[@B22]\]), those that address relationships across the whole clade use either sparse taxon sampling \[[@B16],[@B27]\], or are based on reassessments of Strauch\'s \[[@B18]\] data \[[@B19],[@B28]-[@B30]\]. Note that Dove \[[@B30]\] included a feather microstructural analysis in addition to her reanalysis of Strauch\'s \[[@B18]\] data. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Previous hypotheses shorebird phylogeny.**Family and subfamily level relationships of shorebirds based on: a) Morphological data \[19\]; b) DNA-DNA hybridisation \[16\]; c) Sequence analysis of RAG-1 \[20, 21\], cytochrome-*b*\[22\] and myoglobin intron II \[21\]. :::  ::: Combining phylogenetic data --------------------------- Numerous methods and types of data can be used to infer phylogeny. Frequently, as in Charadriiformes, a single analysis incorporating all taxa of interest is absent. Under the principle of total evidence \[[@B31]\], all sources of phylogenetic information should be combined to maximize their explanatory power. Eernisse and Kluge \[[@B32]\] define total evidence as a method for seeking the best fitting phylogenetic hypothesis for an unpartitioned set of synapomorphies (shared derived characters) using character congruence (characters combined in a supermatrix). Hence, this method combines the primary data (molecular, morphological and behavioural characters) into a single analysis. The approach is powerful because weak signals in the partitioned data sets may be enhanced when combined, and previously obscured relationships may be revealed \[[@B33]\]. The total evidence approach has both practical and theoretical problems. First, only certain types of data can be combined. For example, nucleotide sequences and morphological traits can be readily assessed together as characters, but it is not generally possible to include nucleotide sequences and genetic distance data in a single analysis \[[@B34]\]. We acknowledge that Lapointe *et al.*\[[@B35]\] suggest a distance based approach to combine otherwise incompatible data in a total evidence analysis, although this method has not been tested beyond a single application. The consequence is that it is rarely possible to combine all sources of data in practice and the lack of overlap in combinable data sets may result in a reduction of the number of taxa included. Second, Miyamoto and Fitch \[[@B36]\] contend that combining data sets is rarely justified because partitions of phylogenetic data are real and unequivocal. They argue that several partitions producing similar topologies provide multiple lines of independent evidence supporting that topology. Theoretical arguments over the benefits of total evidence will undoubtedly continue, but perhaps the major barriers to its use are the often very high computational demands of large matrices, and the *a priori*exclusion of certain data types. This is particularly true of Charadriiformes phylogeny, where one of the most significant contributions to the field -- DNA-DNA hybridisation -- cannot be included. An alternative set of techniques, collectively termed supertrees (e.g., Matrix Representation with Parsimony, MRP; \[[@B37],[@B38]\]), enables combination of trees (rather than raw data) from otherwise incompatible sources. MRP methods code source phylogenies based on the presence and absence of taxa at each node of the tree \[[@B37]-[@B39]\] and are thus one step removed from the primary data. It is important to recognise that supertrees should not be regarded as a replacement for exhaustive phylogenetic studies of the primary data and there are drawbacks to the methods (see Discussion). However, they do enable very large phylogenies to be constructed rapidly \[[@B15]\]. Supertrees have been constructed successfully for a wide variety of taxa including carnivores \[[@B15]\], primates \[[@B39]\], seabirds \[[@B40]\], dinosaurs \[[@B41]\], and grasses \[[@B42]\]. Shorebirds are particularly well suited for supertree treatment, since there are numerous incomplete phylogenies available and a broader phylogeny is desirable to facilitate powerful analyses of numerous evolutionary hypotheses (see above). Here, we present the first complete composite phylogeny of extant and recently extinct \[[@B1]\] shorebirds using the MRP approach. We are therefore combining data on tree topologies, and not conducting a simultaneous analysis on the original data. We also use fossil and molecular data to estimate divergence times (see Methods). The combination of complete taxonomic coverage and the inclusion of branch lengths provide the basis for future comparative analyses of Charadriiformes evolution. In addition, conflicting and unresolved areas of Charadriiformes phylogeny are revealed. Results and Discussion ====================== Supertree resolution and topology --------------------------------- We found 1469 equally short trees of length 1847 steps using the parsimony ratchet approach (see Methods). This compares favourably to a standard heuristic search that yielded shortest trees of 1853 steps. All subsequent results and discussion refer to the parsimony ratchet analyses. Figure [2](#F2){ref-type="fig"} shows the family and subfamily level relationships of shorebirds based on the strict and 50 % majority-rule consensus tree (see [additional file 1](#S1){ref-type="supplementary-material"} for branch length estimates). Figures [3](#F3){ref-type="fig"},[4](#F4){ref-type="fig"},[5](#F5){ref-type="fig"},[6](#F6){ref-type="fig"},[7](#F7){ref-type="fig"},[8](#F8){ref-type="fig"},[9](#F9){ref-type="fig"} show the species level phylogeny. The full 50% majority rule consensus and the strict consensus trees are available as [additional file 2](#S2){ref-type="supplementary-material"} and [3](#S3){ref-type="supplementary-material"} respectively. The 50% majority-rule consensus tree is well resolved (73.1%; 255 nodes out of a possible 349 in a fully bifurcating tree), although the strict consensus tree is only 49.6% resolved (173 from 349 possible nodes). The majority rule tree includes nine novel clades (numbers 20, 29, 57, 85, 89, 108, 122, 139, 140) that do not appear in any of the source trees; all of these occur towards the tips of the tree. This is a general problem in supertree construction and such clades should be collapsed as they have no support \[[@B41]\]. To demonstrate where the MRP method has performed badly we have included the novel clades in all figures and list details in the figure legends. In addition, 58 nodes are supported by only one character (see [additional file 1](#S1){ref-type="supplementary-material"}). Each of these nodes is left over from a single source tree. Assessing the support for such nodes is problematic because this may simply reflect a lack of research directed at the taxa in question. A major challenge for supertree construction is to develop measures of support that reflect the robustness of nodes in the source trees. We list the number of characters supporting each node ([additional file 1](#S1){ref-type="supplementary-material"}) but stress that these are not measures of tree robustness and may not be directly comparable even within the same tree. This is because the taxon coverage across source trees is highly variable so some nodes have more potential support than others. Furthermore, because measures of support used in the source trees differ between studies (some source trees include no measures of support), it is impractical and of dubious value to use these measures to assess the robustness of the supertree. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Summary of shorebird supertree.**Family and subfamily level relationships of shorebirds based on 50% majority rule tree. Numbers on nodes refer to age estimates in additional file 1. Boxed node numbers indicate that node collapses to its immediate ancestor in the strict consensus tree (see also additional files 2 and 3 for the full 50% majority rule and strict consensus trees respectively). :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Phylogeny of Larini.**50% majority rule supertree showing the relationships of the Larini. Numbers on nodes refer to age estimates in additional file 1. Boxed node numbers indicate that node collapses to its immediate ancestor in the strict consensus tree (see also additional files 2 and 3 for the full 50% majority rule and strict consensus trees respectively). :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Phylogeny of Sternini.**50% majority rule supertree showing the relationships of the Sternini. Numbers on nodes refer to age estimates in additional file 1. Boxed node numbers indicate that node collapses to its immediate ancestor in the strict consensus tree (see also additional files 2 and 3 for the full 50% majority rule and strict consensus trees respectively). :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Phylogeny of Rynchopini, Stercorariini, *Dromas*, Alcinae, and Glareolidae**50% majority rule supertree showing the relationships of the Rynchopini, Stercorariini, *Dromas*, Alcinae, and Glareolidae. Numbers on nodes refer to age estimates in additional file 1. Boxed node numbers indicate that node collapses to its immediate ancestor in the strict consensus tree (see also additional files 2 and 3 for the full 50% majority rule and strict consensus trees respectively). Node numbers 139 and 140 have no support from any source tree and are novel clades. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Phylogeny of Jacanidae, Rostratulidae, Thinocoridae, Pedionomidae and Scolopacidae**50% majority rule supertree showing the relationships of the Jacanidae, Rostratulidae, Thinocoridae, Pedionomidae and Scolopacidae. Numbers on nodes refer to age estimates in additional file 1. Boxed node numbers indicate that node collapses to its immediate ancestor in the strict consensus tree (see also additional files 2 and 3 for the full 50% majority rule and strict consensus trees respectively). Node numbers 85 and 89 have no support from any source tree and are novel clades. :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Phylogeny of Scolopacidae**50% majority rule supertree showing the relationships of the Scolopacidae. Numbers on nodes refer to age estimates in additional file 1. Boxed node numbers indicate that node collapses to its immediate ancestor in the strict consensus tree (see also additional files 2 and 3 for the full 50% majority rule and strict consensus trees respectively). Node numbers 108 and 122 have no support from any source tree and are novel clades. :::  ::: ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **Phylogeny of Pluvianellidae, Chionidae, Burhinidae, Haematopodini and Recurvirostrini**50% majority rule supertree showing the relationships of the Pluvianellidae, Chionidae, Burhinidae, Haematopodini and Recurvirostrini. Numbers on nodes refer to age estimates in additional file 1. Boxed node numbers indicate that node collapses to its immediate ancestor in the strict consensus tree (see also additional files 2 and 3 for the full 50% majority rule and strict consensus trees respectively). Node numbers 20 and 29 have no support from any source tree and are novel clades. :::  ::: ::: {#F9 .fig} Figure 9 ::: {.caption} ###### **Phylogeny Charadriinae**50% majority rule supertree showing the relationships of the Charadriinae. Numbers on nodes refer to age estimates in additional file 1. Boxed node numbers indicate that node collapses to its immediate ancestor in the strict consensus tree (see also additional files 2 and 3 for the full 50% majority rule and strict consensus trees respectively). Node number 57 have no support from any source tree and are novel clades. :::  ::: The majority of unresolved nodes in the shorebird supertree are located towards the tips of the phylogeny. For example, the genus *Gallinago*forms a monophyletic clade but only two pairs of species are resolved from 14 species (*G. megala*and *G. negripennis*; *G. macrodactyla*and *G. media*) in the majority-rule tree. Only the latter relationship remains in the strict consensus tree. In addition, clades including the genera *Charadrius*and *Vanellus*, *Calidris*and *Tringa*, *Sterna*, and *Scolopax*are poorly resolved. This may reflect a bias in phylogenetic studies of shorebirds. For instance, we found six source trees for Alcinae \[[@B43]-[@B48]\] but none devoted to *Scolopax*or *Gallinago*. Thomas *et al.*\[[@B49]\] indicate that this may be a problem for shorebird studies in general and reported a strong skew favouring research on northern hemisphere species. In contrast to the within genera relationships, the generic and family levels are generally well resolved. The supertree indicates three monophyletic Charadriiformes lineages (figure [2](#F2){ref-type="fig"}). Family and subfamily resolution within each lineage is high, however the relative position of each group is unresolved in the strict consensus tree. This is an important point because the deepest relationships of shorebird phylogeny are contentious \[[@B22]\]. The 50% majority-rule consensus tree indicates that the gulls and allies (Larini, Sternini, Rynchopini, Stercorariini, *Dromas*, Alcinae, and Glareolidae) are sister to the sandpipers and allies (Scolopacidae, Jacanidae, Rostratulidae, Thinocoridae, Pedionomidae). The most basal lineage includes the plovers and allies (Charadriinae, Pluvianellidae, Chionidae, Burhinidae, Haematopodini and Recurvirostrini). The gulls and allies clade is most consistent with DNA-DNA hybridisation \[[@B16]\], indicating that Larini are sister to Sternini and that Rynchopini are sister to this group. This conflicts with morphology-based topologies where Stercorariini are sister to Larini and Sternini with Rynchopini basal to both. Indeed, the position of Stercorariini remains controversial and most recently they were placed as sister to Alcinae \[[@B20]-[@B22]\]. In contrast, morphological evidence \[[@B18],[@B19]\] places Alcinae at the base of the whole Charadriiformes tree with Stercorariini sister to Larini. Thus, the position of Alcinae is uncertain and appears to be dependent on the type of data, with fundamental differences between molecular based analyses and morphological analyses. The taxon sampling of previous morphological and molecular studies varies considerably and it may be this, rather than genuine differences in the phylogenetic signal of different data types, that is the cause of conflict in resolving the phylogenetic position of Alcinae. However, it is encouraging that van Tuinen *et al.*\[[@B17]\] suggested that new unpublished osteological data are consistent with the more derived position indicated by molecular data. The supertree resolves Glareolidae outside the Larini, Sternini, Rynchopini, Stercorariini, *Dromas*, Alcinae clade. This is also the case with recent molecular and previous DNA-DNA hybridisation studies. Morphological studies have failed to resolve the position of Glareolidae, placing the family in a large polytomy with all other major groups except Alcinae and the sandpipers and allies (fig. [1](#F1){ref-type="fig"}). A novel development in shorebird phylogeny is the placement of the black-rumped buttonquail *Turnix hottentotta*as a sister to the gulls and allies (Larini, Sternini, Rynchopini, Stercorariini, *Dromas*, Alcinae, and Glareolidae) based on the nuclear RAG-1 gene \[[@B20]\]. We did not include this species in the supertree because to date Paton et al. \[[@B20]\] remains the only study to reveal an apparently robust relationship. More diverse sampling of the buttonquails (Turnicidae) is essential to corroborate the general affinities of this family. The relationships within the plover clade appear to be reasonably stable. Morphological, molecular, and DNA-DNA hybridisation all place Charadriinae as sister to Haematopodini and Recurvirostrini; our supertree is consistent with these relationships. However, it is not clear whether Burhinidae and Chionidae are sister to each other \[[@B20]-[@B22]\] or whether Chionidae are sister to a Charadriinae, Haematopodini, Recurvirostrini, and Burhinidae clade \[[@B16]\]. Our supertree also included Pluvianellidae, a family consisting of only one species (magellanic plover *Pluvianellus socialis*) and places this as sister to Chionidae. If Pluvianellidae are excluded, the supertree is consistent with the sister group relationship of Burhinidae and Chionidae. The sister group relationship of Jacanidae to Rostratulidae is well established \[[@B16],[@B18]-[@B22]\] and is found in our supertree. The supertree resolves the Thinocoridae and Pedionomidae as sister taxa and this group is sister to the Jacanidae and Rostratulidae. The large Scolopacidae clade is at the base of the sandpiper clade consistent with recent molecular studies \[[@B20]-[@B22]\] and the DNA-DNA hybridisation tapestry \[[@B16]\]. Taken together, it is evident that the supertree is generally more consistent with molecular data (both recent sequence studies and DNA-DNA hybridisation) than with analyses based on morphology. However, it is of course possible that this reflects the greater number of molecular source trees available rather than indicating that molecular data is actually better at resolving shorebird phylogeny. We included several large morphological phylogenies \[e.g \[[@B18],[@B19],[@B26],[@B30],[@B43]\]\] but the majority of source trees (29 out of 51) were based on molecular evidence (see [additional file 5](#S5){ref-type="supplementary-material"}). Node dates ---------- The higher resolution of the majority-rule tree means it is more likely to be of use in comparative studies. We therefore estimated node ages for this topology only (see [additional file 1](#S1){ref-type="supplementary-material"} and [2](#S2){ref-type="supplementary-material"}). We stress that our estimates of node dates are a first attempt at dating the whole tree and have several limitations. First, the fossils used to calibrate seven nodes in the tree are unlikely to be the earliest members of their respective families thus these dates will be underestimates. Second, we assumed that the fossils are grouped with the extant members of the family but this requires formal testing in a phylogenetic framework. Third, the pure birth model assumes that no extinction occurs but this may be unrealistic and it is likely that extinction processes have reduced the representation of older lineages \[[@B15]\]. Furthermore, this model is derived from the topological structure of the tree so errors in tree reconstruction will likely lead to errors in branch length estimation. However, this approach has been employed previously in supertrees of primates \[[@B39]\] and carnivores \[[@B15]\] explicitly to facilitate comparative analyses. Despite these caveats, simulation studies have demonstrated that comparative methods such as independent contrasts are robust to errors in branch length \[[@B50]\] and no viable alternative for dating supertrees has been proposed. Nonetheless, we urge that alternative branch length assumptions are explored if the shorebird supertree is used in future comparative studies. At present, the calibrated RAG-1 tree of Paton et al. \[[@B20]\] remains arguably the most thorough and reliable measure of divergence times for Charadriiformes. A fuller understanding of the phylogenetic affinities of fossil shorebirds will probably improve estimates of node ages for the group. For example, the extinct form Graculavidae, is represented by fossils from the Maastrichtian of New Jersey \[[@B51]\] and Cretaceous of Wyoming \[[@B52]\] but its position within the shorebird clade is unclear. Feduccia \[[@B53]\] suggests that it may be basal and a formal corroboration of this would support proposals for a late Cretaceous origin of shorebirds. The difficulties in dating the shorebird tree are further illustrated by fossil representatives of Recurvirostrini and Burhinidae which are much older than current estimates suggests. The earliest record of the Recurvirostrini is estimated to be over 50 million years old \[[@B54]\] whilst recent discoveries of a possible member of the Burhinidae are dated to around 70 mya \[[@B55],[@B56]\]. There is clearly a need for an integrated phylogenetic study including both extinct and extant shorebirds. Supertree bias -------------- Supertrees are still at an early stage of development and many aspects of MRP, and supertree methods in general, are not yet clearly understood. Steps can be taken to ensure that the supertree includes the most appropriate sets of sources trees, such as only using trees from explicitly phylogenetic studies. This is not always straightforward and could result in the exclusion of important information. For instance, in our shorebird supertree, we included Sibley and Ahlquist\'s DNA-DNA hybridisation tapestry \[[@B16]\] although this is based on distance measures rather than more rigorous phylogenetic methods. Even if very strict tree selection criteria are applied, there are still likely to be biases in the data set. For example, not all source trees are equally well supported, yet in most supertree analyses each tree is treated equally \[[@B57]\]. This is a problem for supertree construction because whilst it is theoretically possible, and indeed beneficial, to weight source trees based on support values \[[@B57]\] it is rarely possible in practice. Many source trees do not have support values and those that do may use different methods, (e.g, bootstrapping or decay indices) which cannot be directly compared with each other. An additional problem that has not been fully resolved relates to correlations between source trees \[[@B58]\]. Several source trees based on the same data set may unduly increase the influence of that data set on the supertree analysis. However, there is no formal way of determining how much overlap to allow and the choice of source trees that go into supertree construction inevitably involves some degree of subjective reasoning. For the shorebird supertree we used strict Reduced Cladistic Consensus trees to summarise potential source trees that were from the same data set but based on different methods. For example, Thomas *et al.*\[[@B22]\] based their phylogeny on cytochrome-*b*but used a range of methods including parsimony and Bayesian analyses. We therefore combined these trees to minimise bias. In contrast, Ericson *et al.*\[[@B21]\] used two types of data: sequences from the nuclear RAG 1 gene and sequences from the myoglobin intron II. They carried out three analyses: each gene separately and then the two combined in a single analysis. In this case, we used three source trees. It could be argued that the combined analysis of Ericson *et al.*\[[@B21]\] should be excluded because of the possible overlap with the individual analyses. However, under the principle of total evidence, the combined data set may result in novel relationships being revealed \[[@B31],[@B33]\] and therefore could contribute important information to the supertree. Simulation and empirical studies are required to fully understand these and other possible biases in supertree construction (e.g., the influence of source tree size and shape) and formal protocols for the selection of source trees are desirable. For transparency, we include a summary of the source trees used, data type, and the main taxa included in the study ([additional file 5](#S5){ref-type="supplementary-material"}). Our shorebird supertree is highly consistent with recent advances in the molecular phylogenetics Charadriiformes. However, we urge caution when using the tree in comparative analyses and encourage the additional use of alternative phylogenies and branch length assumptions. It is particularly important to note that the position of some groups such as the Alcinae remains controversial and that although the majority rule tree is consistent with recent molecular studies, the strict consensus tree fails to resolve the deepest nodes. Conclusions =========== The supertree presented here is, to our knowledge, the first attempt to reconstruct the phylogeny of the entire order Charadriiformes. Overall, the supertree is highly consistent with recent molecular hypotheses of shorebird phylogeny. However, it is apparent that fresh attempts to resolve both the phylogeny and estimates of age will be dependent on further gene sequencing and new fossil discoveries. The affinities of the Alcinae and the relationships between the three major shorebird clades require further corroboration, and studies of several genera such as *Gallinago*and *Vanellus*are desirable. Furthermore, additional work is required to establish the true affinities of the Turnicidae. Nonetheless, it appears that shorebird phylogeny is gradually approaching a consensus view. The broad taxonomic scope and consistency of the supertree mean that is of potentially great value to future comparative studies (accepting the caveats discussed above) of the behaviour, life-history, ecology and conservation of this diverse group. Methods ======= Supertree construction ---------------------- Possible source trees were identified from online searches of Web of Science <http://wos.mimas.ac.uk> covering the years 1981 to 2004. We used the single key strings phylogen\*, cladistic\*, clado\*, classif\*, systematic\*, and taxonom\* (where the asterisks allow variations such as \"phylogeny\" or \"phylogenetics\") in the topic field, in conjunction with a major Charadriiformes taxon name (scientific or common). As supertree methods have been criticized for being biased towards historical trends, we preferred those studies that explicitly set out to derive a phylogenetic hypothesis and so exclude purely (and typically older) descriptive taxonomic works. The Sibley and Ahlquist \[[@B16]\] DNA-DNA hybridisation tapestry may be viewed as non-cladistic, but it was clearly the authors\' intention to reconstruct the phylogeny of birds. Furthermore, it provided a vital catalyst for subsequent studies of avian (including shorebird) phylogeny. We therefore included the DNA-DNA hybridisation hypothesis as a source tree in our analyses. Simulation studies have demonstrated that the performance of supertree methods is improved by including at least one taxonomically complete (or near complete) source tree \[[@B57]\]. We therefore make an exception to our self-imposed rule, and in addition use the taxonomic hierarchy of Monroe and Sibley \[[@B1]\] as a source tree as this includes all extant Charadriiformes species. We acknowledge that this taxonomy is based largely on Sibley and Ahlquist\'s \[[@B16]\] DNA-DNA hybridisation tapestry. The initial search identified 78 source trees from 44 publications. Each source tree was typed as a text file in Nexus format \[[@B59]\]. We coded trees to the species level with species names taken from Monroe and Sibley \[[@B1]\], but note that *contra*Monroe and Sibley \[[@B1]\], we use Charadriiformes not Charadrii to refer to the whole group. Several studies included the gull *Larus thayeri*\[[@B26],[@B60]-[@B63]\] either as a subspecies of *Larus glaucoides*(*Larus glaucoides thayeri*in Monroe and Sibley \[[@B1]\]) or a species in its own right. In recognition of this, we included *Larus glaucoides thayeri*as the only subspecies in our data set thus increasing the total taxa to 366. Monroe and Sibley \[[@B1]\] include 16 species of the family Pteroclidae within the Charadriiformes. However, the relationship of this family to the Charadriiformes is uncertain and they have recently been placed in their own order \[[@B64]\]. We include the Pteroclidae in our analyses only as a means of rooting the tree. Where there were multiple most parsimonious trees (MPTs), or where source trees had been derived from predominantly overlapping data (e.g., from the same data but using alternative methods), we used RadCon \[[@B65]\] to produce strict Reduced Cladistic Consensus trees (RCC \[[@B66],[@B67]\]). The output is in the form of a reduced consensus profile and from this we selected the tree with the highest Cladistic Information Content (CIC) \[[@B65],[@B68]\]. This resulted in a total of 51 source trees from which our supertree is derived and these are summarised in [additional file 5](#S5){ref-type="supplementary-material"}. We produced an MRP matrix of the 51 Nexus \[[@B59]\] source trees in RadCon \[[@B65]\] (see [additional file 6](#S6){ref-type="supplementary-material"} for the MRP file). We used the original MRP coding method of Baum \[[@B37]\] and Ragan \[[@B38]\]. Weighting source trees based on node support such as bootstrapping improves the accuracy of MRP supertrees \[[@B57]\]. However, this is only possible if all source trees can be weighted on the same criteria \[[@B57]\]. The absence of branch support measures in many of the shorebird source trees precludes this approach from the present study; hence, subsequent analyses were conducted using equally weighted parsimony. The tendency of large data sets to produce many sub-optimal trees that are close in length and topology to the shortest tree is a serious problem in phylogenetics. Standard heuristic searches frequently are trapped searching within globally sub-optimal \"islands\" and the tree search is often aborted before completion. Nixon \[[@B69]\] proposed a new method to avoid this problem. The \"Parsimony Ratchet\" reweights a random set of characters from the data set. This may result in the tree island no longer representing a local optimum and the heuristic search continues until a new optimum is reached. The algorithm then reverts to the original weighting and the search continues. Nixon \[[@B69]\] demonstrated the efficacy of the method on a 500-taxon data set, where the ratchet-based search found a tree two steps shorter than standard heuristic searches. We used PAUPRat \[[@B70]\] to implement a parsimony ratchet in PAUP\* \[[@B59]\]. The default settings of 200 iterations and 15% perturbation of characters for reweighting were used and we carried out 20 replicates. Equally parsimonious trees were summarized using both strict and 50% majority-rule consensus methods. We did not calculate any measures of branch support for two reasons. First, their validity and meaning is questionable in MRP supertrees \[[@B41]\]. Second, the number of taxa included in our data set is too large to allow practical calculation of any branch support indices (e.g., decay indices \[[@B71]\]) on a desktop computer. Dating the supertree -------------------- Following Purvis \[[@B39]\] and Bininda-Emonds *et al.*\[[@B15]\] we dated the supertree using both absolute and relative dates. We used data from the Fossil Record 2 \[[@B54]\] as the source of fossil-based absolute dates. This yielded estimates for Jacanidae (*Nupharanassa tolutaria*, Rupellian), *Phalaropus*(*Phalaropus elenorae*, Middle Pliocene), Burhinidae (*Burhinus lucorum*, Lower Miocene), Glareolidae (*Paractiornis perpusillus*, Lower Miocene), Alcinae (*Petralca austrica*, Rupellian), Stercoariini (*Stercorarius*sp., Middle Miocene), and Larini (undetermined, Rupellian). We took the midpoint of the range from the Fossil Record 2 \[[@B54]\] as our date estimate. More recent publications of fossil Charadriiformes were not included because they either represent specimens that are younger or have not been assigned to families that are represented amongst the extant Charadriiformes (such as Turnipacidae \[[@B72]\]). We assumed that fossil dates represent the earliest occurrence for each group which inevitably resulted in underestimates of clade age. The fossil record of Charadriiformes is amongst the best of the modern bird groups \[[@B17]\] in terms of the numbers of taxa, but many specimens are fragmentary and reliable estimates of divergence dates are dependent on a limited number of exceptional specimens \[[@B73]\]. The phylogenetic affinities of the fossil shorebirds in relation to their extant relatives have not yet been fully established, hence have implicitly assumed that fossil representatives of extant groups would be resolved amongst their living relatives. Source trees may include estimates of relative branch lengths (e.g., genetic distances). This allows further dating of the supertree but is problematic because different relative estimates are not comparable and cannot be applied directly to the supertree \[[@B39]\]. However, where a source trees shares a node that has an absolute date in the supertree (a node dated from fossil evidence), the relative branch lengths can easily be converted to estimates of age. All taxa in our supertree are either extant, or very recently extinct; hence, the tips of the calibrated supertree should be equidistant from the root of the tree. In source trees where the relative branch lengths are not equidistant from the root, we followed the protocol of Purvis \[\[[@B39]\]; p.407--8\]. We estimated relative dates using the local molecular clock logic \[[@B74]\] as implemented by Purvis \[[@B39]\] and Bininda-Emonds *et al.*\[[@B15]\]. For example, consider three taxa *A, B,*and *C*where *A*and *B*are sister taxa and *C*is sister to *A*and *B.*The root is dated to 10 million years (myr) from fossil evidence, and independent molecular data provides estimates of divergence based on the number of substitutions per site. The molecular estimates of branch lengths are as follows: *A*, 6 substitutions; *B*, 8 substitutions; *C*, 20 substitutions; *A*and *B*are 11 substitutions from the root. *A*and *B*are therefore separated from their common node by a mean of 7 substitutions. The total length from *A*and *B*to the root is thus 18 substitutions compared to 20 for *C*(a mean of 19). This can be converted to date estimates such that 19 substitutions are equivalent to 10 myr. The dates of the tree are then: ((*A*: 3.68, *B*: 3.68), *C*: 10)). There were no cases where multiple source trees with molecular divergence dates were able to provide estimates for the same node. We estimated relative dates from multiple nodes rather than a single dated node to minimise correlative errors in estimates. To provide date estimates for all nodes in the tree we employed a pure birth model to date nodes for which absolute and relative dates could not be attained \[[@B39]\]. Pure birth models infer that a clade\'s age is proportional to the logarithm of the number of species within the clade: date of daughter = date of ancestor \*(log daughter clade size/log parent clade size) For example, the age of a daughter node that subtends 12 taxa, estimated from its immediate ancestor dated to 20 myr and which subtends 19 taxa is: 20\*(log(12)/log(19)) = 16.879 We applied this approach to estimate the ages of daughter nodes based on dates (absolute or calibrated) of ancestral nodes. We had no ancestral node on which to base estimates of the most basal clade. In this case, we rearranged the pure birth formula and calculated the age of the ancestral node from its two daughter nodes, taking the mean as our \"best estimate\". Finally, to estimate the ages of nodes between daughter and ancestor nodes of known age we spaced the nodes evenly along the branches length \[[@B75]\]. Authors\' contributions ======================= GHT assisted in the design of the study, carried out the phylogenetic analyses and node dating, and drafted the manuscript in partial fulfillment of a doctoral degree at the University of Bath. MAW assisted in the design of the study and with editing and revision of the manuscript. TS assisted in the design of the study, collection of source trees, and editing and revision of the manuscript. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 **Estimates of node ages and node support (branch lengths.xls)**Node numbers correspond to figures 2-9. Five types of estimate were used: a) absolute dates from the fossil record; b) absolute dates from molecular point estimates; c) relative dates based on branch length estimates from molecular studies; d) estimates based on a pure birth model (see text for details); and e) even spacing of nodes along branches with daughters and ancestors of known age. The numbers of characters supporting each node are provided (column D), this is equivalent to the number of source trees that share the equivalent node (see text for details). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 **Shorebird supertree (50% majority-rule consensus; majrulesupertree.tiff)**Shorebird supertree based on 50% majority-rule consensus of 1496 shortest trees with calibrated branch lengths. Scale bar indicates time from the present in millions of years. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 3 **Shorebird supertree (strict consensus; strictsupertree.tif)**Shorebird supertree based on 50% majority-rule consensus of 1496 shortest trees. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 5 **Source trees (source trees.xls)**A summary of each tree used is given including the data type and main taxa studied. This is a brief summary and the original papers should be consulted for full details. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 6 **MRP matrix (shorebirdMRP.txt)**The MRP matrix used in the shorebird supertree analysis. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 4 **Calibrated supertree (shorebirdsupertree.txt)**The supertree in nexus format including branch length estimates. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We thank for Davide Pisani for stimulating discussion of supertree methods during the early planning of this work, Gareth Dyke for information on recent shorebird fossil discoveries, and three anonymous reviewers for helpful comments on an earlier version of the manuscript. Gavin Thomas was supported by a University of Bath Research Studentship.

PubMed Central

2024-06-05T03:55:47.313864

2004-8-24

PMC515297

Background ========== The Indian subcontinent is regarded as a natural genetic laboratory, owing to the co-existence and interaction of socio-culturally, linguistically, ethnically and genetically diversified endogamous populations in a geographical terrain. It is believed that the earliest humans leaving Africa for Eurasia might have taken a coastal route across Saudi Arabia, through Iraq, Iran, to Pakistan and finally entered India along the coastlines \[[@B1]\]. A second wave of migration (\~10,000 years ago) brought in Proto-Dravidian Neolithic farmers from Afghanistan, who were later displaced southwards by a large influx of Indo-European speakers \~3500 years ago in to the subcontinent \[[@B2],[@B3]\]. The origin and settlement of the Indian people still remains intriguing, fascinating scientists to explore the impact of these past and modern migrations on the genetic diversity and structure of contemporary populations \[[@B4]-[@B6]\]. Anthropologically, southern and northern populations are distinct and these differences are further substantiated by (i) the presence of Neolithic sites in this region suggests that Neolithic people of southern India came from north by land and the west-coast by sea \[[@B7]\], (ii) the southern megaliths resemble closely with those of the Mediterranean and western-Europe, while those from northern India are similar to megaliths found in Iran and Baluchisthan \[[@B8]\], and (iii) the predominance of Dravidian language in this region as opposed to their secluded occurrence in central Asia and other parts of India, suggests that the Dravidian languages might have originated within India \[[@B9]\]. It is, thus, of considerable genetic interest to understand the genetic structuring and relationships of southern populations. The present study was carried out on one of the largest southern states, Karnataka, positioned on the southwest coast of India, with a dwelling of about 50 million people. This expanse has been a rich source of prehistoric discoveries dating back to the Paleolithic era that are akin to those seen in Europe \[[@B7]\]. Karnataka has received continuous gene flow from different caste and linguistic groups residing in the adjoining areas of Maharashtra, Andhra Pradesh and Tamil Nadu \[[@B10]\], resulting in the congregation of a large number of diverse endogamous groups within this region. Its large coastline of about 400 Km also attracted the Portugese, Dutch and French traders, who were seeking more profitable ventures on the southern coast at large \[[@B2]\]. Southwest India is, thus, one of the most disparate terrains, with extensive colonization in the past and justifies an in-depth genetic study. A few studies utilizing classical markers have been carried out on southern populations \[[@B5],[@B11],[@B12]\], including few communities of Karnataka \[[@B13],[@B14]\]. However, sound inferences relating to their genetic structuring and diversity could not be drawn due to low discriminatory power of these markers. Recently, microsatellite markers have gained immense popularity in precisely defining population structure, diversity, affinities, gene flow and other crucial aspects associated with population genetics \[[@B15]-[@B21]\] because of the relative expediency, with which a large number of loci and alleles can be typed, facilitating the accumulation of vast data sets that can be readily analyzed with an extensive array of statistical tools \[[@B22],[@B23]\]. These markers also demonstrate high heterozygosity \[[@B24]\], rendering them highly suitable for carrying out the present study. Among the different caste and tribal groups inhabiting the southwest coast of India, we have selected four predominant Dravidian-speaking communities from Karnataka: Iyengar Brahmin, Lyngayat, Gowda and Muslim, they not only belong to dissimilar groups of the Indian caste hierarchy but also have varied migration histories, conferring them uniqueness and significance from a genetic perspective. The present microsatellite study primarily attempts to understand the genetic structure of the four selected populations and to determine their genetic relationship with other linguistically and ethnically similar groups of southern India and Brahmin groups of northern India. It has been suggested that that despite the linguistic homogeneity in southern India, these populations have remained genetically diversified \[[@B25]\]. Hence, we sought to determine the role played by geographical location and linguistic affiliation in genetically differentiating Indian populations. Also, as mentioned earlier, the western coast has witnessed colonization from different world populations, we aim to divulge the impact of these past migrations on the gene pool of the present southern populations by discerning their relationship with historically acclaimed and established migrant groups, ethnically represented by European, Hispanic, East Asian and African populations. Results ======= Allele frequency at 15 STR was used to compute the heterozygosity (observed) for the four studied populations, which varied for each locus, and population but reflected similar values, ranging between 0.724 and 0.797 (Table [1](#T1){ref-type="table"}). An average G~ST~value of 0.009 elucidates the low degree of genetic differentiation in them. However, the G~ST~value for the pooled Indian and global populations demonstrated a high value at 2.3% (data not shown). Genetic relationship of studied populations with other similar southern groups; Vanniyar, Gounder, Pallar and Tanjore Kallar \[[@B26],[@B27]\], northern Brahmins belonging to Orissa \[[@B28]\] and Bihar \[[@B29]\], and four relevant global ethnic groups: European, Hispanic, African \[[@B30]\] and East Asian \[[@B31]\] was divulged by computing DA distances (Table [2](#T2){ref-type="table"}) and represented using NJ tree (Fig. [1](#F1){ref-type="fig"}). Among the four studied populations, Iyengar, Gowda and Muslim formed a distinct cluster. Although NJ tree clearly depicts the clustering of southern populations, D~A~distances indicate that among these groups, Iyengar, Lyngayat and Vanniyar are more similar to the northern Brahmins (0.030). Furthermore, genetic distances emphasize the affinity of Lyngayat with Tanjore Kallar (0.029), Iyengar (0.026) and Vanniyar (0.028). Estimation of relatedness between the southern and global populations shows that all the southern communities formed a separate cluster, nevertheless, genetic distances disclose the affinity of upper caste Indian communities; Iyengar, Lyngayat, Vanniyar, Bihar and Oriya Brahmin with Europeans and East Asians. The Indian populations were most distant to Africans. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Average heterozygosity and G~ST~values for 15 loci in the four studied populations. ::: --------- ------------------------- ---------- ------- -------- ------- OBSERVED HETEROZYGOSITY LOCUS G~ST~ BRAHMIN LINGAYAT GOWDA MUSLIM TPOX 0.707 0.581 0.542 0.555 0.010 D3S1358 0.661 0.793 0.559 0.488 0.036 THO1 0.815 0.785 0.678 0.688 0.005 D21S11 0.876 0.857 0.779 0.733 0.005 D18S51 0.907 0.938 0.779 0.888 0.006 PENTA E 0.921 0.876 0.864 0.800 0.014 D5S818 0.692 0.724 0.525 0.733 0.007 D13S317 0.753 0.714 0.745 0.733 0.007 D7S820 0.723 0.734 0.754 0.800 0.005 D16S539 0.861 0.846 0.830 0.777 0.010 CSF1PO 0.723 0.734 0.745 0.733 0.009 PENTA D 0.815 0.755 0.741 0.933 0.006 vWA 0.784 0.734 0.779 0.688 0.002 D8S1179 0.861 0.822 0.745 0.755 0.007 FGA 0.861 0.894 0.803 0.911 0.017 Average 0.797 0.785 0.724 0.747 0.009 --------- ------------------------- ---------- ------- -------- ------- ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### D~A~distance matrix between ten Indian and four global groups based on allele frequency at 15 microsatellites. ::: **Pop** **HS** **NE** **MO** **CA** **OB** **IB** **LY** **GO** **MU** **BB** **PL** **VN** **TK** **GD** --------- -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- -------- **HS** **NE** 0.079 **MO** 0.049 0.123 **CA** 0.029 0.086 0.07 **OB** 0.044 0.092 0.052 0.04 **IB** 0.044 0.091 0.041 0.04 0.03 **LY** 0.047 0.096 0.047 0.04 0.03 0.026 **GO** 0.066 0.122 0.072 0.07 0.055 0.036 0.047 **MU** 0.076 0.118 0.078 0.07 0.066 0.051 0.056 0.054 **BB** 0.043 0.101 0.052 0.05 0.038 0.031 0.037 0.054 0.068 **PL** 0.061 0.11 0.067 0.07 0.064 0.05 0.056 0.075 0.077 0.063 **VN** 0.045 0.105 0.042 0.04 0.034 0.023 0.028 0.039 0.053 0.037 0.049 **TK** 0.047 0.096 0.053 0.05 0.044 0.028 0.029 0.052 0.062 0.044 0.052 0.032 **GD** 0.064 0.112 0.064 0.06 0.051 0.036 0.043 0.057 0.073 0.054 0.059 0.032 0.043 **Abbreviations used in Table-**Hispanic -- HS, African -- AF, Asian -- AS, European -- EU, OriyaBrahmin -- OB, Iyengar Brahmin -- IB, Lyngayat -- LY, Gowda -- GO, Muslim -- MU, Bihar Brahmin -- BB, Pallar -- PL, Vanniyar -- VN, Tanjore Kallar -- TK, Goundar -- GD. ::: ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Neighbor-joining tree depicting the genetic relationship of Karnataka populations with related Indian and global ethnic groups based on 15 STR markers. :::  ::: The regression model (Fig. [2](#F2){ref-type="fig"}), of mean per locus heterozygosity against distance from centroid assumes that when a population experiences same amount of gene flow from a homogenous source, a linear relationship exists between the expected and observed heterozygosity. A change in gene flow directly affects this linear relationship. The R-matrix when applied to the Indian populations assists in understanding the influence of external gene flow and admixture among populations. The higher observed than expected heterozygosity of Iyengar and Lyngayat, placed above the theoretical regression line helps infer that these populations have received more than average external gene flow, which was also observed in Vanniyar, Pallar and Oriya Brahmin. The Gowda and Muslim groups exhibit lower than expected heterozygosity values and fall below the regression line, suggesting lesser admixture in them. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Regression plot demonstrating the relatively higher gene flow levels in high-ranking populations of India. **Abbreviations used in figure:**OB-Oriya Brahmin, PL-Pallar, IB-Iyngar Brahmin, VN-Vanniyar, LY-Lyngayat, TK-Tanjorekallar, GD-Goundar, Mu-Muslim, BB-Bihar Brahmin, GO-Gowda. :::  ::: The microsatellite diversity computed using AMOVA revealed that the genetic variation observed in Indian populations was mainly confined to variation amongst individuals (\~98%), irrespective of their geographic or linguistic grouping (Table [3](#T3){ref-type="table"}). The geographical clustering of populations into three regions: north, southwest (Karnataka) and southeast (Tamil Nadu) demonstrated a low variance of 0.29%, p = 0.010 (Table [3a](#T3){ref-type="table"}). As compared to geographical grouping, the linguistic clustering (Indo-Caucasian and Dravidian) exhibited a noticeable increase in the molecular variance between the two groups, 0.65% (p = 0.06, Table [3b](#T3){ref-type="table"}). The genetic diversity among populations within each group remained almost similar at both levels of analysis. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Genetic differentiation of Indian populations based on AMOVA ::: **(a) Geographical grouping** ------------------------------------------------ ----------------------------- ---------------------- Groups in set 1 Source of Variation Percentage Variation Group 1 -- North: Bihar and Orissa populations Among groups 0.29 Group 2 -- South-west: Karnataka populations Among populations in groups 0.97 Group 3 -- South-east: Tamil Nadu Populations Within populations 98.74 **(b) Linguistic grouping** Groups in set 2 Source of Variation Percentage Variation Group 1 -- Indo-European: Orissa and Bihar Among groups 0.69 Group 2 -- Dravidian: Southern populations Among populations in groups 0.94 Within populations 98.40 ::: Discussion ========== In recent years, population genetics has witnessed extensive use of microsatellite markers to understand and evolutionary histories of contemporary human populations \[[@B17],[@B32]-[@B34]\]. Though, the populations inhabiting south India have played a major role in formation of the Indian gene pool, however, very few genetic studies have been carried out on them. The present study utilizes 15 STRs to provide comprehensive genetic information on four predominant communities inhabiting the southwest coast of India, which may significantly help in understanding the genetic composition of southern populations. Genetic structure of Karnataka populations ------------------------------------------ The most distinctive feature revealed by the fifteen microsatellites was the considerable genetic homogeneity amongst the four diverse caste groups residing in southwest India. The presence of an almost similar allele frequency pattern \[[@B34]\], suggests that these populations might have a common ancestry or probably experienced very high gene flow during the period of their coexistence. The above finding is further supported by the low genetic differentiation of 1.0% among the studied groups irrespective of their caste and migration histories. The high heterozygosity and rii values in Lyngayat reflect the admixture and stochastic processes experienced by it. The genetic affinity of Lyngayat with other related southern caste populations, like, Iyengar, Vanniyar and Tanjore Kallar reiterates its heterogeneous past. It is noteworthy that although the southern populations exhibited higher affinity amongst each other, the high-ranking populations, like, Iyengar, Lyngayat and Vanniyar also displayed some genetic similarity to Brahmins from Bihar and Orissa, indicating that the gene pool of Iyengar and Lyngayat probably consists of genetic inputs from both southern and northern groups. However, strong conclusions cannot be drawn due to low genetic differentiation among the studied populations. Though the Gowda is known to have moved in to Karnataka from the adjoining area of Tamil Nadu, our study reveals that Gowda cluster with the studied populations and not with Tamil groups. The low hetetozygosity and high rii values of Gowda implies that it might have differentiated as a result of stochastic processes. Furthermore, the relatively lower heterozygosity and admixture levels of Gowda and Muslim might be attributed to the socio-cultural practice of consanguineous marriages in them. The Muslim group was found to be genetically similar to local populations. Regional conversions from diverse castes that occurred during the period of Islamic dominance might elucidate the more or less identical genetic relationship between Muslims and other studied groups. The microsatellite study emphasizes the genetic similarity among the Karnataka populations, with the lack of any strong caste or religious bias in them. Analysis of genetic variance ---------------------------- AMOVA test strongly suggests that genetic diversity among the southern populations was mainly confined to intra-population variation, further emphasizing the genetic homogeneity in them. Analysis using different genetic markers corroborate with our finding that the genetic diversity in human populations can be mainly attributed to variation within populations \[[@B4],[@B17],[@B19],[@B34],[@B36],[@B37]\]. An exploration of the genetic differentiation based on geographical grouping of populations discloses the genetic similarity among populations residing in a region. Nevertheless, the geographic affinity was comparably lesser to that observed within the two linguistic families, viz., Dravidian and Indo-European. Our finding provides evidence to the strong linguistic affinity prevailing amongst the Dravidian speaking populations and imparts them genetic distinctness from the Indo-European linguistic group. Even though prior studies have indicated that genetic clusters often correspond closely to predefined regional and linguistic groups \[[@B34]\], AMOVA suggests that caste system along with geographical contiguity are not ideal platforms for differentiating the analyzed Indian populations. It must, however, be acknowledged that use of less number of polymorphisms in this study might plausibly have led to the greater influence of linguistic affiliation on these populations rather than geographical proximity. Genetic affinity with global populations ---------------------------------------- The genetic differentiation of the studied populations with relevant global migrant groups was estimated to be 2.3%, relatively lower than the 9% observed in another similar study \[[@B16]\], which had used a different set of microsatellite markers. Sampling from a confined area, as well as the use of lesser number of loci might have contributed to this apparent difference in the results. The southern populations formed a separate cluster from the world populations. Molecular studies on Indian populations using diverse markers (nuclear, mtDNA and Y-chromosome) have demonstrated that the upper caste populations have higher semblance with Europeans than Asians \[[@B26]\]. Intriguingly, in the present study, communities belonging to the upper strata of the Hindu caste hierarchy, i.e., Iyengar, Lyngayat, Vanniyar and northern Brahmins, displayed almost identical genetic affinity with both Europeans and East Asians. Therefore, all though it is believed that south India remained isolated and cushioned from the foreign invasions, the southern populations, especially, the high-ranking groups might have genetically admixed with migrant groups that entered via the west coast and north. Further exploration of their relationship is essential before drawing concrete conclusions. A more comprehensive picture would emerge on analysis of mtDNA and Y chromosome markers. Methods ======= The populations --------------- The populations selected in this study comprise of three major Hindu castes-Iyengar, Lyngayat, Gowda and a Muslim community, inhabiting the southwest coastal terrain of Karnataka (11.3 -- 18.45°N latitudes and 74.12 -- 78.40°E longitudes). All the populations belong to the Dravidian linguistic family and are speakers of the local dialect, Kannada, but differ in caste hierarchy and socio-religious practices. Consanguineous marriages have been reported in Karnataka, with inbreeding levels of the order 0.020 to 0.033, in general \[[@B38]\]. Iyengar hold a high position in the Indian caste hierarchy and sporadic accounts on Brahmin, suggests that they primarily migrated from the upper Gangetic plains to southern India. Nonetheless, few bioanthropological studies have revealed that morphologically Brahmins of a geographical region are similar to the local groups. Lyngayat community was initially formed, as a religious cult by the amalgamation of people from different castes and geographical regions but later developed into a distinct community practicing strict marriage endogamy with social sub-divisions such as clans, sub-castes and sects \[[@B10]\]. Gowda is a low ranking agriculturist caste group that typically exhibits the Dravidian socio-cultural characteristic of consanguineous marriage. It is believed to have moved in from the adjoining area of Tamil Nadu. Muslim is a linguistically heterogeneous, complex religio-ethnic group, \[[@B10]\]. It is believed that the invasion of Turks, Afghans (A.D 998--1030) and Moghals during the 15^th^century, introduced new genes only in northern India, suggesting that Muslims from Southern India are mainly local converts \[[@B3]\]. Micosatellite loci studied -------------------------- The 15 STR marker set analyzed in this study consists of thirteen tetra nucleotide repeat loci: D3S1358, THO1, D21S11, D18S51, D5S818, D13S317, D7S820, D16S539, CSF1PO, vWA, D8S179, TPOX, FGA and two penta nucleotide repeat loci: Penta D, Penta E. Their repeat size makes them less prone to slippage of polymerase during enzymatic amplification compared to the dinucleotide repeats, allowing unambiguous typing \[[@B20]\]. The 15 selected loci are situated on 13 different chromosomes, with D5S818 and CSF1PO being present on chromosome 5 and Penta D and D21S11, located on chromosome 21. The alleles across the loci are substantially unlinked, making them suitable for analyzing inter and intra-population genetic diversity. STR Typing ---------- The blood samples were collected from unrelated individuals belonging to -- Iyengar (65), Lyngayat (98), Gowda (59) and Muslim (45) communities, residing in different districts of Karnataka. DNA was extracted from blood by the phenol-chloroform method \[[@B40]\], followed by quantitation using the QuantiBlot™ kit (Perkin-Elmer, Foster City, CA, USA). Two nanogram of the isolated DNA was used as template for the PCR amplification of the 15 STRs using the PowerPlex™16 kit (Promega Corp., Wisconsin Madison, USA). Raw data were collected with the GeneScan™ software, Ver. 3.2.1 (Applied Biosystems, Foster City, CA, USA) and typed using the PowerTyper™ 16 Macro (Promega Corp., Wisconsin Madison, USA). Statistical Analysis -------------------- Allele frequencies of the 15 STR loci were calculated using the gene counting method \[[@B40]\]. The genetic diversity (G~ST~), observed heterozygosity and pairwise genetic distances (DA) were computed using allele frequencies \[[@B42]\]. The DA distance is least affected by sample size and can precisely obtain correct phylogenetic trees under various evolutionary conditions \[[@B43]\]. Neighbor-joining trees were constructed using DA distances \[[@B44]\], and its robustness was established by bootstrap resampling procedures. Analysis of molecular variance (AMOVA) was performed using the Arlequin Ver. 2.00 package \[[@B45]\]. Two levels of analysis were performed to explore the microsatellite diversity among the four studied populations along with six other socio-culturally similar groups inhabiting different regions of India. At the first level, three geographical groups were constructed: (1) north (2) southwest: Karnataka and, (3) southeast: Tamil Nadu, to estimate the genetic variance among populations from diverse geographical regions. The second set of analysis was aimed at investigating the genetic diversity between the Dravidian and Indo-European linguistic family. To assess the gene flow experienced by these populations, the rii value, i.e., the genetic distance of a population from the centroid was calculated using the regression model \[[@B46]\]. This model utilizes the heterozygosity of each population and the distance from the centroid as the arithmetic mean of allele frequencies:  where, r~ii~is the distance from the centroid, p~i~is the frequency of the allele in i^th^population and  is the mean allelic frequency. List of abbreviations ===================== STR -- Short Tandem Repeat AMOVA -- Analysis of Molecular Variance NJ tree -- Neighbor-Joining tree Authors\' contributions ======================= RR carried out the molecular studies, analyzed the genetic data and drafted the manuscript. VKK participated in the design, conceiving and preparation of manuscript. Both authors read and approved the final manuscript. Acknowledgements ================ This work was supported by a research grant under the IX Five Year Plan to CFSL, Kolkata and a research fellowship from the Ministry of Home Affairs to the first 1 author. The technical assistance of Dr. R Trivedi is highly appreciated. This work would not have been possible without the co-operation of volunteers of blood samples used for genotyping in the study. The comments of two anonymous reviewers\' were extremely helpful in improving the text of the paper.

PubMed Central

2024-06-05T03:55:47.317756

2004-8-19

PMC515298

Background ========== The type 1 diabetes, previously called insulin-dependent diabetes mellitus (IDDM), is the consequence of progressive and selective destruction of pancreatic β cells by an immune-mediated process \[[@B1]\], resulting in an absolute lack of insulin \[[@B2]\]. It is well established that this self-destruction is primarily provoked by the activation of the autoreactive T lymphocytes by the production of T-helper 1 cytokines \[[@B3],[@B4]\]. In addition, the difference of the epidemiological data from one region to another could largely explain why the release of the autoimmunity is stimulated under the influence of one or more environmental factors \[[@B5],[@B6]\], in genetically predisposed subjects. It is currently obvious that the strongest genetic susceptibility of predisposition is allotted to the IDDM1 alleles located in the HLA locus of the chromosome 6p21 \[[@B7]-[@B9]\], and the non-HLA alleles, particularly the IDDM2 polymorph gene located in the region 5\' of the insulin gene (*INS*) promoter situated on the chromosome 11p15 \[[@B10],[@B11]\]. Other regions of the genome were identified as IDDM3, coding for the IGF1 receptor, IDDM4, located near the fibroblast growth factor 3 gene, IDDM5, near the estrogens receptor gene \[[@B7]\] etc. The majority of these regions have no possible statistical criteria allowing them to be clearly linked to the disease \[[@B11]\]. In all cases, it is mainly HLA alleles that present a high risk of contracting the disease compared to non-HLA alleles \[[@B6],[@B12]\]. Due to this, they are qualified as high genetic predictive markers of developing type 1 diabetes in families having a type 1 diabetic member. Interestingly, the disease prediction offered with success the possibility of clinical trials to delay, or even prevent the appearance of type 1 diabetes. However, the different HLA types associated with diabetes depend also on the population. The purpose of our study is to measure HLA DR3 and/or DR4 antigen frequencies and their association in diabetic and nondiabetic subjects originating from the west-Algerian region of Tlemcen. Using a case-control retrospective study design, we attempt to determine which is the greatest HLA DR susceptibility contributing to developing type 1 diabetes. Ninety-one (91) eligible subjects (thirty-nine (39) type 1 diabetics and fifty-two (52) nondiabetics with relatives with type 1 diabetes as controls), were recruited at the Internal Medicine Board of the Medical Centre University of Tlemcen. Results ======= Table [1](#T1){ref-type="table"} summarizes HLA DR3, DR4 and DR3DR4 antigen frequencies in type 1 diabetics and their nondiabetic relatives. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### HLA DR phenotypic frequencies according to *p*-values and Odds ratio in type 1 diabetic and nondiabetic subjects. ::: **Type 1 diabetics n = 39: 15 M/24 F** **Nondiabetic controls n = 52: 21 M/31F** -------------------- ---------------------------------------- ------------------------------------------- ----------- ------------ --------------- ------------------- **DR3** 3 (7.69) 7 (17.9) 5 (9.62) 9 (17.31) 0.2646^a^ 0.95 (0.48--1.87) 10 (25.64) 14 (26.92) 1.000 **DR4** 4 (10.26) 9 (23.08) 6 (11.54) 4 (7.69) 0.0085 ^a,^\* 2.10 (1.04--4.24) 13 (33.33) 10 (19.23) 0.0361\* **DR3DR4** 4 (10.26) 4 (10.26) 2 (3.85) 7 (13.46) 0.0699^a^ 1.30 (0.60--2.80) 8 (20.51) 9 (17.31) 0.5887 **DR3 and/or DR4** 11 (28.21) 20 (51.28) 13 (25.0) 20 (38.46) 0.5791^a^ 2.21 (1.13--4.36) 31 (79.49) 33 (63.46) 0.0194\* **X/X** 4 (10.26) 4 (10.26) 8 (15.38) 11 (21.15) 0.0111 ^a,^\* \- 8 (20.51) 19 (36.54) 0.0194\* ^a^*p*-Values related to the sex ratio comparison (male = M, female = F). \* *p*\< 0.05. X = non-DR3, non-DR4. ::: DR4 and DR3DR4 antigens showed an association with susceptibility to type 1 diabetes (DR4; OR = 2.10, DR3DR4; OR = 1.30, that is respectively OR confidence interval 1.04--4.24 and 0.60--2.80, 95% CI), in contrast, DR3 antigens showed no association with the disease (OR = 0.95, OR confidence interval 0.48--1.87, 95% CI). It is important to notice that the strongest association is found in DR4 phenotype as indicated in Figure [1](#F1){ref-type="fig"}. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **HLA DR antigen associations with type 1 diabetes.**The block boxes represent the odds ratios (OR). The horizontal lines represent the lower and upper confidence limit of OR (confidence interval, 95 % CI). :::  ::: In addition, the phenotypic frequency of DR4 or DR3DR4 molecules is higher in diabetic group than in the control one, although the difference did not reach significance level in DR3DR4 frequencies (*p*= 0.0361 and *p*= 0.5887 respectively). On the contrary, the DR3 molecules frequency is slightly decreased in type 1 diabetic patients compared to the controls and presents no statistically significant difference (*p*= 1.000). Furthermore, no incidence was related to the sex criteria for the frequencies of DR3 and DR3DR4 molecules (*p*\> 0.05). However, significant differences in HLA DR4 frequencies are linked to the female sex and present a value definitely higher in type 1 diabetic patients compared to those of nondiabetics for the same sex (*p*\< 0.05) (Figure [2](#F2){ref-type="fig"}). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **HLA DR antigen frequencies compared to the sex ratio.**\* *p*\< 0.05, Type 1 diabetics vs. control subjects. :::  ::: Discussion ========== Type 1 diabetes is a polygenic disease which results from the interaction between environmental (viral, toxic, nutritional, socioeconomic \[[@B6]\]) and genetic factors. It is the form of diabetes which occurs mainly in children and young adults \[[@B13]\]. Fortunately, molecular epidemiology offers the hope of the possibility of preventing the disease in the future, by evaluating the potential factors of risk of developing this pathology \[[@B6]\]. Although almost 90 % of new cases of type 1 diabetes occur sporadically, studies of individuals with a diabetic relative in their family are essential \[[@B4]\]. It is well established that associations between type 1 diabetes and certain HLA antigens largely facilitate the identification of the subjects having a potential risk to develop the disease. Among the putative HLA molecules known to confer a high susceptibility are DQA1 (\*0301)-DQB1 (\*0302) (= DQ8), DQA1 (\*0501)-DQB1 (\*0201) (= DQ2), DRA-DRB1 (\*0401) (= DR4) and DRA-DRB1 (\*0301) (= DR3) \[[@B14]\]. However, epidemiological studies showed that the different HLA types associated with the diabetes depend also on the various populations. For instance, the risk of developing type 1 diabetes in Caucasians is greater if they are carrying the HLA A8 and B15, while DQ6 alleles are protectors \[[@B11],[@B14],[@B15]\]. Among Japanese, it is the association with the HLA B54 that confers a higher susceptibility to develop the disease, while the strongest association was found with HLA DR and HLA DQ locus \[[@B11]\]. In Algeria, and especially in its western region (Tlemcen), which is known for its history of consanguineous marriage, there is a high rate of consanguinity. This fact could increase the risk of developing type 1 diabetes by favouring the transmission of HLA haplotypes and recessive genes of susceptibility except for HLA antigens that are common in both parents. For these reasons, we were interested in checking whether phenotypically DR3DR4 antigens would involve less risk to develop the disease, in comparison with DR3 or DR4 antigens, knowing the fact that the presence of a probable consanguinity could reduce the frequency of DR3DR4 polymorph phenotype compared to that of DR3 or DR4 phenotype. On a purely comparative basis, similarities seem to be found between our results and those of other investigators \[[@B5],[@B13],[@B15]\] with regards to HLA DR4 or DR3DR4 frequencies, which are higher in the type 1 diabetic than the nondiabetic population. On the contrary, the HLA DR3 antigens showed comparable frequencies in both groups of our sample. Consequently, these observations associating DR3 phenotype to a protector effect against type 1 diabetes in our studied population are thus do not conform with those reported in the literature \[[@B15]-[@B18]\]. However, DR4 and DR3DR4 antigens are obviously associated with susceptibility of developing the disease. It should also be noted that DR3DR4 antigens might represent a weaker predictive value of disease risk. We concluded from this that the non-excess of DR3DR4 antigens, or comparable frequencies of DR3 molecules between type 1 diabetic patients and ND relative controls, might be a strong indices of consanguinity in our sample. A recent study carried out in Sweden showed that HLA DR3 is associated to the development of type 1 diabetes and the incompatibility of blood group ABo \[[@B19]\]. One can thus note that the role of the HLA DR3 antigens in conferring risk for type 1 diabetes can be masked in the homogeneous populations. Due to the restricted size of our sample, which may influence our interpretation, we should not consider the odds ratio of HLA DR3 antigens or the eventual consanguinity of our studied population. Moreover, many evidences \[[@B20]-[@B23]\] incriminate DR3 and/or DR4 antigens in the susceptibility to type 1 diabetes and their association with the detected autoantibodies in this disease. Indeed, the marks of autoimmunity are much more observed in the diabetic patients carrying DR3 or DR4 antigens than in the diabetic patients where DR3 or DR4 antigens were absents \[[@B13]\]. Furthermore, recent research indicates that a subject with HLA DR4 or DR3 alleles has three or four times more chance of developing type 1 diabetes compared to the general population; DR3DR4 is associated with the highest risk (20 to 40 times more) \[[@B11]\]. In Algeria, very few investigations have been undertaken to study the impact of the genetic background on the risk to develop type 1 diabetes in its population, where an annual average incidence of 4.7 per 100000 has already been listed \[[@B24]\]. In a similar study carried out on Algerian unrelated type 1 diabetics (n = 50) and nondiabetics controls (n = 46), the presence of DR3-DQ2 (linkage disequilibrium) in 45 % of patients and in 13 % of controls was detected by molecular genotyping method using PCR (polymerase chain reaction) and SSO (sequence specific oligonucleotide). DR4-DQ8 was found in 37 % of diabetic cases and in 4 % of control groups \[[@B25]\]. Finally, association with type 1 diabetes attributed to DR-B1\*0405 (alleles of DR4 antigens expression) susceptibility showed a match with our results; however, the DR4 antigens were found to be linked to the female sex. According to the results showed in Table [1](#T1){ref-type="table"}, there is no difference between men and women patients with type 1 diabetes carrying DR4 antigens, but a significant difference was noticed in female sex, either women were contracted or not with diabetic disease. It is certain that screening of HLA class II sub-types and determination of DNA coding sequences allows more precise characterization of ethnic groups, especially, because the same coded molecules can differ by the position of few amino acids. For example, it is well established that DQA1 and DQB1 alleles (sub-types of HLA DQ) code respectively for the alpha and beta chain of the DQ molecule \[[@B6]\]. Thus, the combination of DQ alpha with Arg in position 52 (Arg-52) and DQ beta in position 57 without Asp (non-Asp 57) is called a diabetogenic heterodimer which is the biggest risk factor of type 1 diabetes in Caucasian. Nevertheless, among the Japanese population, type 1 diabetes is particularly associated with HLA DQ alpha Arg-52, but not with HLA DQ beta non-Asp 57 \[[@B5]\]. Moreover, it is true that the search of an association with a candidate gene allows a better characterization for most of the frequent multifactorial diseases, because the candidate genes are directly implied in the pathological processes. Today, it is reported that the CTLA-4 (cytotoxic T-lymphocyte antigen-4) allele, localised on the chromosome 2p33, former IDDM12, is largely associated with the susceptibility of numerous common complex diseases, such as the common autoimmune disorders Graves\' disease, the autoimmune hypothyroidism and type 1 diabetes \[[@B26]\]. Combined, these pertinent observations still open new perspectives for debating this crucial subject concerning the public health. Conclusion ========== Type 1 diabetes, or youth diabetes, is a multifactorial disease occurring on a genetic ground of predisposition and starts under certain environmental conditions. The most effective preventive strategy must be designed at the pre-diabetes stage, since immune and/or genetic markers can easily indicate subjects with high risk before the clinical symptoms. Thus, the genetic analysis of HLA class II associations has allowed a screening of the contributing molecules to the type 1 diabetes development and the selection of subjects, which are likely to contract the disease. In this study, we were confronted with difficulties in interpretation of our results which are mainly due to the presence of some indices of consanguinity in our studied population, showing a non excess of DR3DR4 phenotype compared to the DR3 or DR4 phenotype, and a no link of DR3 phenotype to the disease. This result could be due to an ethnic characteristic of Tlemcen population, a history of consanguineous marriages. Nevertheless, these preliminary results made it possible to answer the asked questions. Thus, the strongest type 1 diabetes association is statistically revealed with phenotypic expression of DR4 followed by DR3DR4 phenotype. Association studies between the disease and genetic polymorphisms should be carried out on the population having more limited consanguinity to reduce confusions in result interpretations. Methods ======= Subjects -------- The study was conducted on ninety-one first-degree related subjects (brothers, sisters and siblings), randomly recruited at the Internal Medicine Board of the Medical Centre University of Tlemcen (west-Algeria). The sample included thirty-nine patients with type 1 diabetes (15 males, 24 females), and fifty-two healthy subjects (21 males, 31 females) selected from diabetes relatives as controls. Prior medical histories and personal characteristics were obtained from participants *via*a questionnaire. The patients\' mean (± Standard Deviation) age at clinical onset was 12.28 ± 5.97 years with range of 5 to 22 years and median of 11 years. Subjects who were not first degree related and who were not originating from Tlemcen region were excluded. The use of first-degree relatives eliminates exposures of environmental factors such as food items and viruses since first-degree relatives usually share the same milieu. For execution of the protocol, the informed consent was obtained from all the participating subjects to the designated study. HLA phenotyping (standard complement-dependent assay) ----------------------------------------------------- The applied serologic technique lies on the aptitude of the antibodies to recognize allotropic determinants of HLA molecules on cellular surface \[[@B27]\]. This method is sensitized by a reaction of microlymphocytotoxicity \[[@B28],[@B29]\], which uses specific anti-HLA DR antisera and rabbit complement of commercial typing tray (Biotest, Germany). Initially, the peripheral blood lymphocytes (PBL) were separated from the other illustrated elements of venous blood (collected in EDTA-containing tubes) by density gradient centrifugation on Ficoll-Hypaque \[[@B30]\]. B lymphocytes were isolated by using nylon-wool-separated columns \[[@B31]\]. Statistical analysis -------------------- The comparison of phenotypic frequencies was obtained by using chi-square analysis with Yates\' correction or by Fisher\'s Exact test, whenever appropriate. The application of the observed χ^2^vs. expected χ^2^was employed to show significance of frequency differences with the sex ratio. A *p*value of less than 0.05 was considered statistically significant (two-by-two table: degrees of freedom (df) = 1, chi-square ≥ 3.84) \[[@B32]\]. The association between HLA antigens and type 1 diabetes was performed by determination of odds ratio (OR) \[[@B33],[@B34]\] (confidence interval, 95 % CI). All statistical analyses were performed using the Epi Info 2000 Version 1.0 for Windows 95, 98, NT, and 2000 computers (Epi Info, Atlanta, Georgia, USA) and STATISTICA Version 5.0, \'97 (STATISTICA, StatSoft, Paris, France). Author\'s contributions ======================= AM drafted the manuscript, performed statistical analyses and carried out the bibliography research. MS participated by coordinating and orienting the designated study. BA carried out HLA phenotyping and participated in the study design. KM recruited the eligible subjects. All authors read and approved the final version of the manuscript. Acknowledgments =============== The authors would like to express their deep recognition and greatest thanks to Professor Kaoual Meguenni, Head of the Epidemiology Board (Medical Centre University of Tlemcen) for his precious assistance and using data processing materials of his Board. They are also indebted to Dr Malika Bendahmane for critically reading the manuscript. They are grateful to the staff of Immunology and Hematology Boards of the Medical Centre University staff of Tlemcen for their collaboration and technical assistance, and to the Scientific and Technical Research Centre (SERIST) staff of Tlemcen for their documentary assistance in this study.

PubMed Central

2024-06-05T03:55:47.320287

2004-8-24

PMC515299

Background ========== Defensins constitute a large family of small, cysteine-rich, cationic peptides that are capable of killing a broad spectrum of pathogens, including various bacteria, fungi, and certain enveloped viruses \[[@B1]-[@B5]\]. These peptides play a critical role in host defense and disease resistance by protecting the hosts against infections. Transgenic mice expressing human enteric defensin *HD5*are fully protected against the doses of *Salmonella typhimurium*that are otherwise lethal to the wide-type mice \[[@B6]\]. Conversely, mice deficient in the matrilysin gene, which is responsible for activating enteric defensins, become more susceptible to oral infection with *S. typhimurium*\[[@B7]\]. Defensins have been identified in species ranging from plants, insects to animals and humans \[[@B1]-[@B5]\]. Characterized by the presence of 6--8 cysteine residues in relatively defined positions, all defensins are structurally related in that they form 3--4 intramolecular disulfide bonds and 2--3 antiparallel β-sheets with or without an α-helix. Based on the spacing pattern of cysteines, these peptides are broadly divided into five groups; namely plant, invertebrate, α-, β-, and θ-defensins \[[@B1]-[@B5]\]. Alignment of all known defensin sequences revealed the consensus defensin motif of each group as follows: plant defensin: C-X~8--11~-C-X~3--5~-C-X~3~-C-X~9--12~-C-X~4--11~-C-X~1~-C-X~3~-C; invertebrate defensin: C-X~5--16~-C-X~3~-C-X~9--10~-C-X~4--7~-C-X~1~-C; α-defensin: C-X~1~-C-X~3--4~-C-X~9~-C-X~6--10~-C-C; and β-defensin: C-X~4--8~-C-X~3--5~-C-X~9--13~-C-X~4--7~-C-C. The α- and β-defensins are unique to vertebrate animals with α-defensins only being found in rodents and primates, while β-defensins are present in all mammalian species investigated \[[@B1]-[@B3]\]. On the other hand, θ-defensins have only been found in certain primates as a result of posttranslational ligation of two α-defensin-like sequences \[[@B8]-[@B10]\]. A pseudogene for θ-defensin is also present in humans \[[@B11]\]. Analysis of human and mouse genomes indicated that β-defensins form 4--5 distinct clusters on different chromosomes with each cluster consisting of multiple defensin genes \[[@B12]\]. Interestingly, the single mammalian α-defensin locus is located on a β-defensin cluster with θ-defensins residing in the center of α-defensins \[[@B12]\]. Studies with mammalian defensins suggested a rapid duplication followed by positive selection and diversification within each group \[[@B13]-[@B18]\]. However, the evolutionary relationships among three groups of mammalian defensins and among plant, invertebrate, and mammalian defensins remain controversial. Similarity in spatial structure and biological functions favors the notion that all mammalian defensins are evolutionarily related \[[@B19]\], although a phylogenetic analysis suggested a closer relationship between β- and insect defensins than between α- and β-defensins \[[@B16]\]. Existence of a large number of expressed sequence tag (EST) sequences and recent completion of chicken genome sequencing at a 6.6× coverage \[[@B20]\] provided a timely opportunity to discover a complete repertoire of defensin-related sequences in birds for studying the evolutionary relationship between invertebrate and mammalian defensins. Here we report identification of a single β-defensin cluster that is composed of 13 genes located on the chicken chromosome 3q3.5-q3.7. Evolutionary and comparative analyses of these chicken β-defensins with mammalian homologues strongly suggested that all mammalian defensins have evolved from a common β-defensin-like ancestor, which has undergone rapid duplication, positive diversifying selection, and chromosomal translocations, thereby giving rising to multiple gene clusters on different chromosomal regions. Results and Discussion ====================== Discovery of novel chicken defensins ------------------------------------ To identify novel defensin genes in the chicken, all five groups of known defensin-like peptide sequences from plants, invertebrates, and vertebrates were first queried individually against the translated chicken nonredundant (NR), EST, high throughput genomic sequence (HTGS), and whole-genome shortgun sequence (WGS) databases in the GenBank by using the TBLASTN program\[[@B21]\]. All potential hits were then examined manually for the presence of the characteristic cysteine motifs. For every novel defensin identified, additional iterative BLAST searches were performed until no more novel sequences could be found. In addition to three known chicken β-defensins (*Gal 1--3*) \[[@B22],[@B23]\], nine novel putative sequences, namely *Gal 4--12*, have been found in the EST database with at least two hits for each, and such sequences have also been confirmed in genomic sequences (Table [1](#T1){ref-type="table"}). Because of the fact that mammalian defensins tend to form clusters \[[@B12],[@B14],[@B15],[@B18]\], all chicken HTGS and WGS sequences containing defensin sequences were also retrieved from GenBank, translated into six open reading frames, and manually curated. As a result, an additional putative β-defensin, *Gal13*, was identified in several genomic clones (Table [1](#T1){ref-type="table"}). The open reading frame of *Gal13*was predicted by GENSCAN \[[@B24]\] and confirmed by directly sequencing of RT-PCR product amplified from chicken kidney. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Identification of chicken β-defensins ::: -------------------------------------------------------------------------------------------------------- Gene EST^1,2^ HTGS WGS Gene Size (bp)^3^ ------- ---------- ---------- --------------- ------------------- ------ ----- ------ ----- ------ ----- Gal1 BX260462 AC110874 AADN01058097 70 972 88 482 127 496 217 Gal2 BX540940 AC110874 AADN01058097 66 1113 143 183 121 674 204 Gal3 AC110874 AADN01058097\ 53 980 109 1180 215 AADN01058096 Gal4 BU451960 AADN01058096 136 461 127 117 141 Gal5 BU389548 AADN01058096 290 445 127 355 187 Gal6 CF251501 AC110874 AADN01058097\ 52 704 86 705 130 249 234 AADN01058098 Gal7 CF251115 AC110874 AADN01058098 50 656 86 201 130 234 248 Gal8 BU242665 AC110874 AADN01058098 71 915 91 259 134 706 494 Gal9 BX270804 AC110874 AADN01058098 220 1592 130 781 343 Gal10 AW198592 AC110874 AADN01058099 118 268 133 1719 381 Gal11 BM440069 AC110874 AADN01058101 63 966 129 1001 460 Gal12 BX257296 AC110874 AADN01058102 84 396 420 Gal13 AC110874 AADN01058102 61 1016 118 3322 91 -------------------------------------------------------------------------------------------------------- ^1^Abbreviations: *EST*, expressed sequence tag; *HTGS*, high throughput genomic sequence; *WGS*, whole-genome shortgun sequence; *E*, exon;*I*, intron. ^2^One EST sequence entry is given only for the exemplary purpose. In each case, more than two independent EST sequences have been found, except for *Gal3*and *Gal13*, both of which have no EST sequences. *Gal3*was found through homology cloning \[23\], and *Gal13*was predicted by us from the genomic sequence. ^3^All *Gal*genes are predicted to consist of four exons separated by three introns, except for *Gal12*, whose last two exons are fused together. The absence of additional sequence information at the 5\'-untranslated regions of the cDNA sequences prevented prediction of the sizes of first exon and intron for *Gal 3--5*and *Gal 9--13*genes. ::: No other sequence containing β-defensin-like six-cysteine motif has been found in NR, EST or genomic databases, suggesting that 13 *Gal*genes constitute the entire repertoire of the β-defensin family encoded in the chicken genome. Although it is highly unlikely, we could not rule out the possibility that additional defensin-related genes with distant homology might be uncovered in the chicken by different computational search methods such as the use of Hidden Markov models \[[@B12],[@B15]\]. It is noted that none of other groups of defensins have been discovered in the chicken, indicating that plant, invertebrate, α-, and θ-defensins are absent in the chicken lineage. Similar to Gal 1--3, 10 novel β-defensins, deduced from either EST or genomic sequences, vary from 63 to 104 amino acid residues in length. Alignment of these peptides revealed a conservation of the signal sequence at the N-terminus and the characteristic six-cysteine defensin motif at the C-terminus (Figure [1](#F1){ref-type="fig"}). Consistent with the fact that all β-defensins are a group of secreted molecules in response to infections, the signal sequences of all chicken defensins are hydrophobic and rich in leucines. In addition, the mature C-terminal sequences are all positively charged due to the presence of excess arginines and lysines. Interestingly, Gal11 contains two tandem, but highly divergent, copies of the six-cysteine motif at the C-terminus, and is the only defensin having such sequences. Functional significance for existence of such two defensin motifs remains to be studied. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Multiple sequence alignment of chicken β-defensins.**The intervening region between signal and mature peptide sequence is the short propiece. The conserved residues are shaded. Also shown is the length of each peptide. Notice the six-cysteine defensin motif is highly conserved. The six cysteines in the second tandem copy of the defensin motif in Gal11 are boxed. :::  ::: Evolutionary analysis of vertebrate β-defensins ----------------------------------------------- Phylogenetic analysis of vertebrate β-defensins showed that chicken defensins clustered with various different groups of mammalian β-defensins (Figure [2](#F2){ref-type="fig"}). However, the bootstrap support for these patterns was very weak (less than 50% in all cases). The clustering of certain chicken β-defensins with mammalian homologues suggests that major subfamilies of β-defensins arose before the last common ancestor of birds and mammals, estimated to have occurred about 310 million years ago \[[@B25]\]. This in turn implies that some duplication of β-defensin genes must have taken place before the divergence of birds and mammals. The apparent lack of α-defensins in the chicken and other non-mammalian species (G. Zhang, unpublished data) suggests that α-defensins may have evolved after mammals diverged from other vertebrates. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Phylogenetic relationship of vertebrate β-defensins.**The tree was constructed by the neighbor-joining method and the reliability of each branch was assessed by using 1000 bootstrap replications. Numbers on the branches indicate the percentage of 1000 bootstrap samples supporting the branch. Only branches supported by a bootstrap value of at least 50% are indicated. Chicken β-defensins are highlighted in yellow. Abbreviations: *BNBD*, bovine neutrophil β-defensin; *LAP*, lingual antimicrobial peptide; *EBD*, enteric β-defensin; *TAP*, tracheal antimicrobial peptide; *PBD*, porcine β-defensin; *DEFB/Defb*, β-defensin; *Gal*, Gallinacin; *GAPDH*, glyceraldehyde-3-phosphate dehydrogenase. :::  ::: Comparison of the numbers of synonymous and nonsynonymous nucleotide substitutions provides a powerful test of the hypothesis that positive Darwinian selection has acted to favor changes at the amino acid level \[[@B26]\]. This approach has previously been applied to both α- and β-defensins of mammals and has revealed positive selection acting on the mature defensin but not on other regions of the gene \[[@B16],[@B17]\]. In the comparison of the chicken β-defensin sequences, synonymous sites were saturated with changes or nearly so, making it impossible to test the hypothesis of positive selection in every case. In pairwise comparisons among all sequences, mean p~S~in the propeptide region was 0.551 ± 0.036 (S.E.), while mean p~N~was 0.369 ± 0.040. In the mature defensin region, mean p~S~was 0.673 ± 0.027, while mean p~N~was 0.534 ± 0.051. Mean p~N~in the mature defensin was significantly greater than that in the propeptide (z-test; P \< 0.05), indicating lesser functional constraint on the amino acid sequence of the former. The high mean p~S~shows that chicken β-defensin genes have not duplicated recently, unlike β-defensin genes of the bovine \[[@B16]\]. In the comparison between the most closely related pair of sequences (*Gal6*and *Gal7*), mean p~S~in the mature defensin was 0.221 ± 0.082, while mean p~N~was 0.331 ± 0.076. While these values are not significantly different at the 5% level, the fact that p~N~was higher than p~S~suggested that positive selection may have acted to diversify the mature defensin region between these two genes. Genomic organization and chromosomal localization of the chicken β-defensin gene cluster ---------------------------------------------------------------------------------------- Searching through HTGS database led to identification of two overlapping bacterial artificial chromosome (BAC) sequences, TAM31-54I5 (accession no. AC110874) and CH261-162O9 (accession no. AC146292), both of which were sequenced and deposited earlier by one of us (J.F. Chen). Alignment of these two sequences allowed to re-order three DNA fragments in AC110874 and to construct a continuous, gap-free genomic contig that includes 11 *Gal*genes except for *Gal4*and *Gal5*. Later search of chicken WGS sequences released on February 29, 2004 confirmed the order of the genomic contig that we assembled and also revealed the locations of two remaining genes, *Gal4*and *Gal5*, both of which reside on a WGS (accession no. AADN01058096) that overlaps with AC110874 (Figure [3](#F3){ref-type="fig"}). The position and orientation of each *Gal*gene were obtained by comparing its cDNA with the assembled DNA sequence. As shown in Figure [3](#F3){ref-type="fig"}, all 13 *Gal*genes were clustered densely within a distance of 86.0 Kb on the genome. It was also confirmed by aligning such a contig with the chicken genome assembly, in which 13 *Gal*genes are located on six WGS contigs (Table [1](#T1){ref-type="table"}) of chromosome 3 that are only \~3.3 Mb from the distal end. Consistent with this, the *Gal*gene cluster was physically mapped to the tip of chicken chromosome 3 at the region of q3.5-q3.7 by fluorescence in situ hybridization (FISH) using the TAM31-54I5 BAC DNA as probe (Figure [4](#F4){ref-type="fig"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Genomic organization of the chicken β-defensin gene cluster.**The horizontal lines at the bottom represent the three overlapping genomic clones that were used to assemble the continuous, gap-free contig. The position of each gene is represented by a solid vertical bar and the width of each bar is proportional to the size of each gene. The direction of transcription is indicated by the triangle above each gene. The genes with solid triangles are transcribed in the direction opposite to the ones with open triangles. Slanted lines refer to the sequences omitted. Note that the three fragments of AC110874 sequence have been re-ordered and the gaps have been filled following alignment with AC146292. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Chromosomal localization of the chicken β-defensin gene cluster by fluorescence in situ hybridization.**The BAC clone TAM31-54I4, which harbors 11 *Gal*genes, was mapped to chicken chromosome 3q3.5-q3.7. Arrows indicate the hybridization signals. :::  ::: Comparing the cDNA with genomic sequences also revealed the structure of each *Gal*gene. Unlike most mammalian β-defensin genes, which primarily consist of two exons and one intron, the *Gal*genes were found to be composed of four short exons separated by three introns with variable lengths ranging from 117 bp to 3,322 bp (Table [1](#T1){ref-type="table"}). *Gal12*is an exception, in which the last two exons have been fused together. While the first exon of the *Gal*genes encodes 5\'-untranslated region (UTR) and the majority of the last exon encodes 3\'-UTR as well as a few C-terminal amino acids, two internal exons resemble mammalian β-defensin genes in that one exon encodes the signal and pro-sequence and the other encodes the mature sequence with six-cysteine motif \[[@B19],[@B27]-[@B29]\]. Apparently, the first two and the last two exons of the *Gal*genes have joined together during the evolution as a result of exon shuffling, which occurred in many other evolutionarily conserved gene families \[[@B30]\], including invertebrate defensins \[[@B5]\]. The fusion of defensin exons in mammals is presumably adaptive because it allows a faster mobilization of such host defense molecules to better cope with invading microbes. Tissue expression patterns of chicken β-defensins ------------------------------------------------- It has been shown that *Gal1*and *Gal2*are expressed in bone marrow and lung, while *Gal3*is more preferentially expressed in bone marrow, tongue, trachea, and bursa of Fabricius \[[@B23]\]. To study the tissue expression patterns of novel *Gal*genes that we identified, RT-PCR was performed with a panel of 32 different chicken tissues. Similar to *Gal 1--3*, *Gal 4--7*are highly restricted to bone marrow cells with *Gal5*also expressed in tongue, trachea, lung, and brain at lower levels (Figure [5](#F5){ref-type="fig"}). By contrast, the six remaining genes, *Gal 8--13*, were not found in bone marrow, but instead in liver, kidney, testicle, ovary, and male and female reproductive tracts (Figure [5](#F5){ref-type="fig"}). These results clearly suggested that all chicken β-defensin genes can be divided into two subgroups. Seven genes (*Gal 1--7*) are predominantly expressed in bone marrow and the respiratory tract, whereas the other six genes (*Gal 8--13*) are more restricted to liver and the urogenital tract. However, the functional significance and transcriptional regulatory mechanisms of these genes during inflammation and infection remain to be investigated. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Tissue expression patterns of 10 novel chicken β-defensins by RT-PCR.**See Materials and Methods for details. The number of PCR cycles was optimized for each gene, and the specificity of each PCR product was confirmed by sequencing. The house-keeping gene, GAPDH, was used for normalization of the template input. :::  ::: Comparative analysis of chicken and mammalian β-defensin gene clusters ---------------------------------------------------------------------- To study the origin and evolution of mammalian defensins, a comparative analysis of β-defensin gene clusters in the chicken, mouse, and human was performed by employing additional, more phylogenetically conserved gene markers surrounding the defensin clusters. As shown in Figure [6](#F6){ref-type="fig"}, two genes, *CTSB*(Cathepsin B, accession no. NP\_680093) and a human EST sequence (accession no. BE072524) immediately located centromeric to chicken defensins, were also found to be conserved in the defensin gene clusters on human chromosome 8p22 and mouse chromosome 14C3. Similarly, another gene, *HARL2754*(accession no. XP\_372011) that is 6-Kb telemetric to *Gal4*is also conserved in another defensin cluster in human (8p23) or mouse (8A1.3) (Figure [6](#F6){ref-type="fig"}). ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Comparative analysis of defensin clusters among the chicken, mouse, and human.**The gene clusters were drawn proportionally according to their sizes. Each vertical line/bar represents the position of a gene, and the width of each line/bar is proportional to the size of each gene. Three highly conserved genes (*CTSB*, BE072524, and *HARL2754*) surrounding the defensin clusters in the chicken, mouse, and human were connected by solid lines. The position of the α-defensin locus (*DEFA*) was indicated as an open square. Note that the human θ-defensin pseudogene resides in the *DEFA*locus. The positions and orders of defensin genes in human and mouse were drawn based on the genome assemblies released in July 2003 and October 2003, respectively. Abbreviations: *GGA*, chicken chromosome; *MMA*, mouse chromosome; *HSA*, human chromosome; *Tel*, telomere; *Cen*, centromere. :::  ::: These results strongly suggested that all vertebrate β-defensins are evolved from a single gene. This conclusion is further supported by the fact that there are three highly similar β-defensin-like sequences present in the largely finished zebrafish genome (G. Zhang, unpublished data). In addition, a group of homologous β-defensin-like sequences, namely crotamine and myotoxins, have been found in several *Crotalus*snakes \[[@B31]\], which are presumably derived from a single ancestral gene. The appearance of multiple β-defensin gene clusters on different chromosomal regions in mammalian species \[[@B12]\] is apparently a result of rapid gene duplication, positive diversifying selection, and chromosomal translocation following divergence of mammals from other vertebrate lineages. In addition to the structural conservation between β-defensin-like sequences in the rattlesnake and mammals \[[@B32]\], a growing body of evidence suggests that their functions appear to be largely conserved in that both are capable of interacting negatively-charged lipid membranes followed by formation of ion channels or pores \[[@B32]-[@B34]\]. It is noteworthy that the conservation of Cathepsin B (CTSB) adjacent to β-defensins is perhaps not surprising, given the recent finding that cathepsins are involved in the cleavage and inactivation of β-defensins \[[@B35]\]. Conclusions =========== We have showed that chicken genome encodes a total of 13 different β-defensin genes clustered densely within a 86-Kb distance on the chromosome 3q3.5-q3.7, but with no α-defensin genes. These peptides exhibit homology to different subgroups of mammalian β-defensins-, consistent with the hypothesis that α-defensins and β-defensins arose by gene duplication after the divergence of birds and mammals. The θ-defensins are specific to primates; and thus appear to have arisen from α-defensins by gene duplication specific to the primate lineage. Apparently, the evolution of defensins is rapid and driven by duplication and positive diversifying selection. Collectively, this study represents the first large-scale detailed investigation of defensins in non-mammalian vertebrates. There is no doubt that further analysis of these defensin genes will lead to a better understanding of host defense mechanisms and evolution of innate immunity. Methods ======= Computational search for novel chicken defensins ------------------------------------------------ To identify novel defensins in the chicken, all known cysteine-containing defensin-like peptide sequences discovered in plants, invertebrates, birds, and mammals were individually queried against the translated chicken NR, EST, HTGS, and WGS databases in the GenBank by using the TBLASTN program \[[@B21]\] with default settings on the NCBI web site \[[@B36]\]. All potential hits were then examined for the presence of the characteristic defensin motif. For every novel defensin identified, additional iterative BLAST searches were performed until no more novel sequences could be revealed. Because mammalian defensins tend to form clusters \[[@B12],[@B14],[@B15],[@B18]\], all chicken genomic sequences containing defensin sequences were also retrieved from the GenBank and translated into six open reading frames and curated manually for the presence of the defensin motif in order to discover potential sequences with distant homology. Alignment and phylogenetic analysis of chicken β-defensins ---------------------------------------------------------- Multiple sequence alignment was constructed by using the ClustalW program (version 1.82) \[[@B37]\]. A phylogenetic tree of amino acid sequences of mature β-defensins was constructed by the neighbor-joining method \[[@B38]\]. So that a comparable data set would be used for all pairwise comparisons, any site at which the alignment postulated a gap in any sequence was excluded from the analysis. To maximize the number of sites available for analysis, certain sequences with large deletions were excluded from the analysis. Because the sequences were very short (25 aligned sites), no correction for multiple hits was applied. The reliability of clustering patterns within the tree was assessed by bootstrapping; 1000 bootstrap pseudo-samples were used. The proportion of synonymous nucleotide differences per synonymous site (p~S~) and the proportion of nonsynonymous nucleotide differences per nonsynonymous site (p~N~) were estimated by the method of Nei and Gojobori \[[@B26]\]. Again, no correction for multiple hits was applied because a small number of sites were examined. Assembly of the chicken β-defensin gene cluster ----------------------------------------------- To generate a continuous defensin gene cluster, the HTGS and WGS sequences containing the putative defensin genes were retrieved from the GenBank, aligned to generate a longer contig, which was confirmed later by searching through the assembled chicken genome released on February 29, 2004, by using the BLAT program \[[@B39]\] under the UCSC Genome Browser web site \[[@B40]\]. The relative positions, orientations, and structural organizations of individual genes were determined by comparing its cDNA sequence to the continuous genomic contig that we assembled. Chromosome localization of the chicken β-defensin gene cluster -------------------------------------------------------------- Fluorescence in situ hybridization (FISH) was used for chromosomal assignment of the chicken β-defensin gene cluster by using the BAC clone TAM31-54I4 as probe, which harbors 11 *Gal*genes. Metaphase chromosome speads were prepared from mitogen-stimulated chicken splenocyte culture as we described \[[@B41],[@B42]\]. The BAC clone was labeled by nick translation with biotin 16-dUTP (Roche Diagnostics), hybridized to metaphase chromosome DNA, followed by detection with FITC-labeled avidin (Roche Diagnostics) and staining with propidium iodide to simultaneously induce the R-banding. RT-PCR analysis of the tissue expression patterns of chicken β-defensins ------------------------------------------------------------------------ Total RNA was extracted with Trizol (Invitrogen) from a total of 32 different tissues from healthy, 2-month-old chickens (see Figure [5](#F5){ref-type="fig"}). A total of 4 μg RNA from each tissue were reverse transcribed with random hexamers and Superscript II reverse transcriptase by using a first-strand cDNA synthesis kit (Invitrogen) according to the instructions. The subsequent PCR was carried out with 1/40 of the first-strand cDNA and gene-specific primers for each β-defensin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as described \[[@B28],[@B43]\]. Every pair of primers were designed to locate on different exons to aid in distinguishing PCR products amplified from cDNA vs. genomic DNA (Table [2](#T2){ref-type="table"}). The PCR program used was: 94°C denaturation for 2 min, followed by different cycles of 94°C denaturation for 20 sec, 55°C annealing for 20 sec, and 72°C extension for 40 sec, followed by a final extension at 72°C for 5 min. The number of PCR cycle was optimized for each gene to ensure linear amplification (Table [2](#T2){ref-type="table"}). A half of the PCR products were analyzed by electrophoresis on 1.2% agarose gels containing 0.5 μg/ml ethidium bromide. The specificity of each PCR product was confirmed by cloning of the PCR product into T/A cloning vector, followed by sequencing of the recombinant plasmid. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Primer sequences used for RT-PCR analysis of novel chicken β-defensins ::: Gene Primer Sequence Product Size (bp) Cycles Used ------- ------------------------- ------------------------ ------------- ------ ---- Gal4 CATCTCAGTGTCGTTTCTCTGC ACAATGGTTCCCCAAATCCAAC 321 899 36 Gal5 CTGCCAGCAAGAAAGGAACCTG TGAACGTGAAGGGACATCAGAG 300 1100 36 Gal6 AGGATTTCACATCCCAGCCGTG CAGGAGAAGCCAGTGAGTCATC 249 1203 36 Gal7 CTGCTGTCTGTCCTCTTTGTGG CATTTGGTAGATGCAGGAAGGA 230 665 35 Gal8 ACAGTGTGAGCAGGCAGGAGGGA CTCTTCTGTTCAGCCTTTGGTG 261 967 35 Gal9 GCAAAGGCTATTCCACAGCAG AGCATTTCAGCTTCCCACCAC 211 1802 33 Gal10 TGGGGCACGCAGTCCACAAC ATCAGCTCCTCAAGGCAGTG 298 2285 33 Gal11 ACTGCATCCGTTCCAAAGTCTG TCGGGCAGCTTCTCTACAAC 301 1299 33 Gal12 CCCAGCAGGACCAAAGCAATG GTGAATCCACAGCCAATGAGAG 335 731 36 Gal13 CATCGTTGTCATTCTCCTCCTC ACTTGCAGCGTGTGGGAGTTG 175 4514 50 GAPDH GCACGCCATCACTATCTTCC CATCCACCGTCTTCTGTGTG 356 876 30 ::: Note added in proof =================== Following submission of this manuscript, Lynn *et al.*reported independently discovery of seven novel chicken β-defensins in the chicken EST database by using homology search strategies \[[@B44]\]. Consistent with our conclusion, they also revealed occurrence of positive selection particularly in the mature region of chicken β-defensins following evolutionary analysis. Moreover, albeit the use of a different nomenclature, they confirmed that the expressions of *Gal 4--7*are primarily in bone marrow, while other genes are more restricted to liver and the genitourinary tract. List of abbreviations ===================== Abbreviations: Gal, Gallinacin; NR, nonredundant; EST, expressed sequence tag; HTGS, high throughput genomic sequence; WGS, whole-genome shortgun sequence; BAC, bacterial artificial chromosome; FISH, fluorescence in situ hybridization; UTR, untranslated region; GAPDH, glyceraldehyde-3-phosphate dehydrogenase. Authors\' contributions ======================= YX carried out the tissue collection, RT-PCR analysis of tissue expression patterns, and drafted the manuscript. ALH carried out the phylogenetic and molecular evolutionary analyses. JA and YM carried out the fluorescence in situ hybridization. JFC carried out the sequencing of two chicken defensin-containing BAC clones. DSN participated in tissue collection and preparation. GZ conceived of the study, carried out all computational analyses and annotation, drafted the manuscript, and participated in its design and coordination. All authors read and approved the final manuscript. Acknowledgements ================ This work was supported in part by Oklahoma Center for Advancement of Science and Technology Grant HR-136 (to GZ) and the Oklahoma Agricultural Experiment Station.

PubMed Central

2024-06-05T03:55:47.321950

2004-8-13

PMC515300

Background ========== Nosocomial bacterial pneumonia is frequently polymicrobial, with gram-negative bacilli predominating \[[@B1]\]. Because delays in antimicrobial treatment can lead to adverse outcomes, the choice of empirical therapy is vital. Many effective antimicrobial agents are available, but the treatment of nosocomial pneumonia remains challenging. We recently reported the antibiotic-resistance patterns of respiratory isolates of *Pseudomonas aeruginosa*in our region \[[@B2]\]. The current study investigates the distribution and drug resistance of other gram-negative bacteria in the respiratory secretions of hospitalized patients. Results ======= Table [I](#T1){ref-type="table"} and Table [II](#T2){ref-type="table"} present the antibiotic susceptibility patterns of our isolates. The most common organisms were *Klebsiella*spp (35%), *A. baumanii*(27%), and *E. coli*(15%). We also isolated rare organisms such as *Stenotrophomonas maltophilia, Burkholderia*spp, and *Hafnia alvei*. All studied Enterobacteriaceae (except *Enterobacter*spp) were far more susceptible to ticarcillin-clavulanate than to ticarcillin alone, which suggests that the primary mechanism of resistance in these organisms is β-lactamase production. *K. pneumoniae*accounted for 79% of *Klebsiella*isolates. *Klebsiella*spp were generally more susceptible to the tested antimicrobials than were *Enterobacter*spp, *Serratia*spp*,*or *E. coli*. The overall resistance rates to the third-generation cephalosporins (cefotaxime, ceftazidime, and ceftriaxone) were as follows: *Klebsiella*spp, 10%--19%; *Serratia*spp, 16%--33%; and *Enterobacter*spp, 22%--45%. *Serratia*spp were less resistant to third-generation cephalosporins than *Enterobacter*spp. *E. coli*isolates resistant to piperacillin, gentamicin, and the fluoroquinolones accounted for only 4% of all E. coli isolates. Imipenem was the most active agent against our isolates. After imipenem, ciprofloxacin, and the aminoglycosides, tetracycline was the most active agent against *A. baumanii*. Tobramycin was more effective against *A. baumanii*than against Enterobacteriaceae. Tobramycin and imipenem were the most active agents against both gentamicin- and ciprofloxacin-resistant *A. baumanii*(Table [III](#T3){ref-type="table"}). We observed the ESBL phenotype in 10 *E. coli*isolates (20.8%) and 11 *K. pneumoniae*isolates (12.2%). All *K. pneumoniae*and *E. coli*isolates with the ESBL phenotype were resistant to tetracycline. Regarding *K. pneumoniae*isolates, 2 were susceptible to tobramycin, 3 to gentamicin, and 4 to ciprofloxacin, but 8 were susceptible to amikacin and imipenem. Regarding *E. coli*isolates, 4 were susceptible to tobramycin, 7 to gentamicin, 5 to ciprofloxacin, and 8 to amikacin, but all were susceptible to imipenem. Discussion ========== Contrary to the findings of the Turkish antimicrobial resistance study group \[[@B3]\], our *Klebsiella*isolates were more susceptible to third-generation cephalosporins (42.6% vs. 81.6% for ceftazidime; 65.8% vs. 85.7% for cefotaxime), aztreonam (44.0% vs. 65.6%), and ticarcillin-clavulanate (37.0% vs. 60.3%). *Klebsiella*spp were 81.6% susceptible to ceftazidime in our study; these rates are 96.6% in North America \[[@B4]\], 86.7% in China \[[@B5]\], 80.5% in Korea \[[@B6]\], 69.4% in Latin America \[[@B7]\], and 51.9% in India \[[@B8]\]. Although the isolation of *Acinetobacter*spp in respiratory specimens may reflect colonization and not necessarily infection \[[@B9]\], the most common site of nosocomial *Acinetobacter*infection is the lower respiratory tract, especially in mechanically ventilated patients \[[@B10]\]. *Acinetobacter*spp were the second most frequent gram-negative bacilli isolated from patients with pneumonia in Latin America \[[@B7]\]. In our survey, all compounds tested showed decreased activity among the *A. baumanii*isolates. Susceptibility to imipenem was \>95% in Canada \[[@B11]\], India \[[@B8]\], and China \[[@B5]\]; susceptibility was 80.5% in our study and 55.5% in another study from Turkey \[[@B12]\]. The high prevalence of respiratory tract infections due to multiresistant *A. baumanii*will stimulate the use of carbapenems and possibly increase carbapenem resistance in our region. Only 75% of our *E. coli*isolates were susceptible to ciprofloxacin. However, this rate was greater in Europe (95.2%) \[[@B13]\], North America (93.3%) \[[@B4]\], and Latin America (93.9%) \[[@B7]\]. *E. coli\'s*susceptibility to ceftazidime was \>95% in Europe \[[@B13]\], North America \[[@B4]\], China \[[@B5]\], and Korea \[[@B6]\] but only 84.8% in Latin America \[[@B7]\], 69.6% in our study, and 42.1% in India \[[@B8]\]. Imipenem, the most active compound, inhibited 97.8% of our *E. coli*isolates. Conversely, *E. coli*strains were not resistant to imipenem in Europe \[[@B13]\], Latin America \[[@B7]\], India \[[@B8]\], China \[[@B5]\], or Korea \[[@B6]\]. *Enterobacter*spp showed high rates of resistance to broad-spectrum penicillins with or without β-lactamase inhibitors (41.7% resistance to ticarcillin-clavulanate) and third-generation cephalosporins (45.2% resistance to ceftazidime). The high rates of ceftazidime resistance among *Enterobacter*spp suggests a high prevalence of stably derepressed AmpC cephalosporinase-producing strains. Interestingly, resistance to third-generation cephalosporins, aztreonam, and ticarcillin-clavulanate was higher in our study (28.1%--48.0%) than with the Turkish antimicrobial resistance study group (13.3%--38.3%) \[[@B3]\]. In their study, no *Enterobacter*or *Serratia*isolates were resistant to imipenem. In our study, however, the rates of susceptibility to imipenem were 86.2% for *Enterobacter*spp and 76.5% for *Serratia*spp. Imipenem susceptibility for these two species was \>95% in other parts of the world \[[@B7]-[@B10],[@B13]\]. Moreover, *Serratia*spp were at least 95% susceptible to ceftazidime in the United States \[[@B14]\], Canada \[[@B11]\], India \[[@B8]\], China \[[@B5]\], and Korea \[[@B6]\]. From 1997 to 1999, ESBL detection rates in *K. pneumoniae*isolates were 45.4% in Latin America, 24.6% in the Western Pacific, 22.6% in Europe, 7.6% in the United States, and 4.9% in Canada \[[@B15]\]; this rate was 12.2% in our survey, but the other study from Turkey reported a rate of 60.5% \[[@B3]\]. During this same period, ESBL detection rates in *E. coli*isolates were 8.5% in Latin America, 7.9% in the Western Pacific, 5.3% in Europe, 4.2% in Canada, and 3.3% in the United States \[[@B15]\]; this rate was 20.8% in our study. The presence of the ESBL phenotype in *E. coli*isolates decreased susceptibility to the aminoglycosides, tetracycline, and ciprofloxacin but not imipenem, suggesting the presence of other resistance genes in ESBL-encoding plasmids. Despite the high percentage of ESBL production in *E. coli*isolates, antibiotics remained reasonably effective with these isolates. Imipenem was active against all ESBL-producing *E. coli*isolates. *E. coli*remained 30.0% resistant to gentamicin; this resistance rate was 75.9% in the Western Pacific, 57.8% in Latin America, 25.7% in Europe, and 21.1% in the United States \[[@B15]\]. Only 0.5%--0.7% of ESBL-producing *K. pneumoniae*isolates were resistant to imipenem in the United States, Latin America, the Western Pacific region, and Canada \[[@B15]\]. Our rate (27.2%) was very high in comparison. This finding may be due to our low number of isolates or our lack of a confirmation test for the ESBL phenotype. Resistance to tetracycline among ESBL-producing *K. pneumoniae*strains was 61.1% in Canada, 55.1% in the Western Pacific, 52.0% in Latin America, 49.5% in Europe, and 44.4% in the United States \[[@B15]\], but this rate was 100% in our study. Resistance to ciprofloxacin among ESBL-producing *K. pneumoniae*strains was 44.2% in the Western Pacific, 34.6% in the United States, 24.3% in Europe, 23.1% in Latin America, 22.2% in Canada, and 63.6% in our study. We found only one imipenem-resistant *E. coli*isolate. It was resistant to ampicillin, ticarcillin, and piperacillin but susceptible to ceftazidime, ceftriaxone, and aztreonam. This profile suggests an oxacillinase with carbapenemase properties. This finding is interesting because class D enzymes have been found only in *Acinetobacter*spp \[[@B16]\]. Two imipenem-resistant *Klebsiella*spp were resistant to all β-lactams, including aztreonam. These species were probably expressed a metallo-β-lactamase with additional mechanisms (efflux, cephalosporinase hyperproduction) \[[@B16]\]. The absence of a confirmation test for the ESBL phenotype limits the impact of our results. On the other hand, it is known that supplemented media (blood) can alter the zone diameters for several agents and bacterial species. Despite these limitations, our data can be used for local therapeutic choices. Conclusions =========== We previously presented the antibiotic susceptibility patterns of 249 respiratory isolates of *P. aeruginosa*during the same period \[[@B2]\]. When combined with our current data, these results show that, in our region, ceftazidime can still be used for managing respiratory infections due to gram-negative aerobic bacteria in combination with aminoglycosides. It appears that increasing imipenem resistance may cause serious therapeutic problems in future. Methods ======= We collected our data from 01/01/1999 to 01/01/2002 at the microbiology laboratory of the University of Cumhuriyet. We processed the data to eliminate duplicate registrations. We excluded any isolates collected within 7 days when they came from the same specimen source of the same patient. We initially identified the isolates using such routine methods as colonial/microscopic morphology and enzymatic characteristics. We confirmed species identification with API-bioMerieux products. We retrospectively analyzed antibiotic susceptibility patterns in 238 respiratory isolates of Enterobacteriaceae members and 90 respiratory isolates of *A. baumanii*. We accepted all consecutive isolates because we did not attempt to distinguish actual pathogens from colonizing strains. Specimen types consisted of sputum (54.2%), transtracheal/endotracheal aspirates (24.6%), and bronchial lavage fluid (21.2%). We cultured sputum samples that showed no oral contamination in the presence of sputum purulence or a suspected lower respiratory infection. We confirmed susceptibility to 18 antimicrobial agents using disk diffusion according to the National Committee for Clinical Laboratory Standards (NCCLS) guidelines \[[@B17]\], except insofar as we supplemented Mueller-Hinton agar with 5% defibrinated blood. We aerobically incubated the inoculated plates at 35°C and evaluated them after 24 h. For quality control of the disk diffusion tests, we used *E. coli*ATCC 25922 and *Staphylococcus aureus*ATCC 25923 strains. The disks (Oxoid) contained the following antimicrobials: ampicillin (10 μg), ampicillin/sulbactam (20 μg), piperacillin (100 μg), aztreonam (30 μg), cefazolin (30 μg), cefuroxime (30 μg), cefotaxime (30 μg), ceftriaxone (30 μg), ceftazidime (30 μg), amikacin (30 μg), gentamicin (10 μg), tobramycin (10 μg), ciprofloxacin (5 μg), imipenem (10 μg), tetracycline (30 μg), and cotrimoxazole (25 μg). Until November 1999, our microbiology laboratory based susceptibility rates on disk zone sizes; thereafter, we used a coordinating laboratory to determine the minimal inhibitory concentrations (MICs) of these 18 antimicrobial agents, accomplished with a standardized microdilution technique (Sceptor System, Becton Dickinson Microbiology System). We used this system to determine MICs for all strains. We used the NCCLS criteria to identify possible extended-spectrum β-lactamase (ESBL)-producing strains of *Klebsiella*spp and *E. coli*when MICs were increased (2 mg/mL) with ceftazidime and/or ceftriaxone and/or aztreonam \[[@B18]\], but we lacked a test to confirm the ESBL phenotype. We classified our results into two categories. We labeled strains deemed susceptible by the disk diffusion method or microdilution technique as *susceptible*. We labeled all resistant and intermediate isolates as *resistant*. We divided the number of resistant isolates by the total number of isolates that had undergone susceptibility testing. Authors\' contributions ======================= UG had primary responsibility for study design, collection of data, and writing the manuscript. IA, MZB, TE had intellectual contribution as well as the writing of manuscript. All authors read and approved the final manuscript. Acknowledgements ================ We thank Cem Celik, Nurgul Turkmen, and Muhtereme Kilavuz for their technical assistance. No portion of this work supported by a foundation. Figures and Tables ================== ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Susceptibility rates (percentages) of Enterobacteriaceae and *Acinetobacter baumanii*between January 1999 and November 1999 (Disk diffusion method). ::: Antibiotics Klebsiella spp (n: 28) E.coli (n: 18) Proteus spp. (n: 10) Enterobacter spp (n: 10) A. baumanii (n: 7) Serratia spp. (n: 6) ------------------- ------------------------ ---------------- ---------------------- -------------------------- -------------------- ---------------------- Ampicillin \- 33.3 \- \- \- \- Amoxicillin/clav. 7.1 33.3 \- \- \- \- Aztreonam 42.9 44.4 50.0 60.0 \- 33.3 Piperacillin 32.1 27.8 50.0 30.0 \- 33.3 Cefazolin 64.3 66.7 50.0 10.0 14.3 \- Cefuroxime 46.4 66.7 40.0 40.0 \- \- Cefotaxime 67.9 72.2 90.0 100.0 \- 66.7 Ceftazidime 71.4 72.2 70.0 90.0 14.3 50.0 Ceftriaxone 71.4 88.9 90.0 90.0 \- 50.0 Amikacin 92.9 100.0 100.0 100.0 28.6 100.0 Gentamicin 82.1 100.0 80.0 90.0 85.7 100.0 Tobramycin 10.7 16.6 \- 41.7 14.3 33.3 Ciprofloxacin 92.9 88.9 90.0 70.0 42.9 100.0 İmipenem 100.0 100.0 80.0 100.0 100.0 100.0 Tetracycline \- 22.2 20.0 20.0 \- 16.7 Cotrimoxazole 64.3 55.6 70.0 100.0 14.3 100.0 ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Susceptibility rates of Enterobacteriaceae and Acinetobacter baumanii between November 1999 and January 2002 (Microdilution technique). ::: Antibiotics Klebsiella spp (n: 86) A. baumanii (n: 83) E.coli (n: 30) Enterobacter spp (n: 24) Proteus spp. (n: 13) Serratia spp. (n: 13) ------------------- ------------------------ --------------------- ---------------- -------------------------- ---------------------- ----------------------- -------- ------- --------- ------- -------- ------ Ampicillin 8/\>16 10.5 16/\>16 7.2 8/\>16 10.0 8/\>16 4.2 8/\>16 15.4 8/\>16 \- Amoxicillin/clav. 4/\>16 77.9 16/\>16 7.2 4/\>16 13.3 8/\>16 4.2 4/\>16 84.6 8/\>16 \- Aztreonam 2/\>16 73.3 8/\>16 6.0 2/\>16 66.6 2/\>16 45.8 2/\>16 92.3 2/16 84.6 Piperacillin 16/\>64 60.5 16/128 39.7 16/128 30.0 16/128 41.7 16/\>64 84.6 16/128 61.5 Ticarcillin 32/\>64 25.6 16/128 7.2 16/128 26.7 32/128 54.2 32/\>64 53.8 32/128 38.5 Ticarcillin/clav. 16/128 60.5 16/128 9.6 16/128 40.0 16/128 58.3 16/\>64 92.3 16/128 61.5 Cefazolin 8/\>16 77.9 16/\>16 1.2 8/\>16 46.7 8/\>16 8.3 8/\>16 69.2 8/\>16 \- Cefuroxime 8/\>16 88.4 16/\>16 3.6 8/\>16 46.7 8/\>16 12.5 8/\>16 76.9 8/\>16 \- Cefotaxime 8/64 96.5 16/64 9.6 16/64 73.3 8/64 58.3 8/64 100.0 8/32 84.6 Ceftazidime 1/\>16 84.9 4/\>16 42.2 1/\>16 66.7 1/32 37.5 1/\>16 92.3 1/\>16 76.9 Ceftriaxone 16/\>32 90.7 16/64 12.0 16/\>32 73.3 8/\>32 50.0 16/\>32 100.0 8/32 84.6 Amikacin 16/\>32 93.0 16/\>32 54.2 16/\>32 93.3 8/\>32 100.0 16/\>32 100.0 8/\>32 92.3 Gentamicin 4/16 89.5 4/16 65.0 4/\>8 90.0 4/\>8 95.8 4/\>8 69.2 4/\>8 84.6 Tobramycin 4/16 84.9 4/16 91.5 4/\>8 93.3 4/\>8 41.7 4/\>8 84.6 4/\>8 76.9 Ciprofloxacin 1/\>2 89.5 1/\>2 66.2 1/\>2 66.6 1/\>2 100.0 1/\>2 100.0 1/\>2 92.3 İmipenem 2/\>8 95.3 4/\>8 79.5 2/\>8 96.6 2/\>8 79.2 2/\>8 100.0 2/\>8 69.2 Tetracycline 4/\>8 81.4 4/\>8 56.6 4/\>8 63.3 4/\>8 79.2 4/\>8 15.4 4/\>8 15.4 Cotrimoxazole 1/\>2 81.4 1/\>2 22.9 1/\>2 53.3 1/\>2 75.0 1/\>2 23.1 1/\>2 69.2 ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Susceptibility (numbers) of imipenem-, gentamicin-, and ciprofloxacin-resistant Acinetobacter baumanii. ::: All isolates (n: 90) Gentamicin-resistant (n: 30) Imipenem-resistant (n: 18) Ciprofloxacin-resistant (n: 32) ------------------------- ---------------------- ------------------------------ ---------------------------- --------------------------------- Ampicillin 6 2 \- \- Amoxicillin/clavulanate 6 3 \- \- Aztreonam 5 1 1 \- Piperacillin 33 4 1 9 Ticarcillin 6 2 \- 1 Ticarcillin/clavulanate 8 5 \- 1 Cefazolin 2 1 \- \- Cefuroxime 3 \- 1 \- Cefotaxime 8 1 1 \- Ceftazidime 36 5 6 7 Ceftriaxone 10 1 1 \- Amikacin 47 2 6 13 Gentamicin 60 \- 11 25 Tobramycin 58 24 17 28 Ciprofloxacin 58 21 11 \- İmipenem 72 25 \- 26 Tetracycline 47 17 10 6 Cotrimoxazole 20 10 5 6 :::

PubMed Central

2024-06-05T03:55:47.325399

2004-8-22

PMC515301

Background ========== For many years the vast majority of gene expression studies have employed non-quantitative or semi-quantitative RNA gel blots and RT-PCR analysis. Real-time PCR technology has removed many of the difficulties associated with quantitative gene expression studies \[[@B1]\], and real-time quantitative RT-PCR (qRT-PCR) is rapidly being adopted as a standard method for in-depth expression studies, including studies of alternative splicing, verification of microarray expression results, and molecular diagnostics \[[@B2]-[@B5]\]. Real-time qRT-PCR offers a robust means for precisely quantifying changes in gene expression over a wide dynamic range. It is also applicable to experiments where RNA amounts are limiting, such as for micro-dissected tissues. However, selection of an appropriate normalization method is crucial for reliable quantitative gene expression results \[[@B1],[@B6]\]. The purpose of normalization is to correct for non-specific variation, such as differences in RNA quantity and quality, which can affect efficiencies of the RT and PCR reactions. Normalization to total RNA content poses a number of problems. It is difficult to quantify small amounts of RNA, and variation in RT and PCR reaction efficiencies are not accounted for by this method. Similarly, normalization to an external RNA standard is problematic due to RNA instability. The most commonly used method is relative quantitation, whereby gene expression level is normalized to that of an internal reference gene. While this avoids the problems and limitations of absolute quantitation, selection of a proper internal control--gene expressed at a nearly constant level in all tissue samples being investigated--is required. Failure to use an appropriate control gene may result in biased gene expression profiles, as well as low precision. The consequences may be that only gross changes in expression level are declared statistically significant, or that patterns of expression are erroneously characterized. Until recently, internal controls (often referred to as housekeeping or maintenance genes), were selected based on stability of expression in qualitative studies (e.g., visual examination of RNA gel-blots), via low-sensitivity assays such as densitometry of hybridized blots, or via semi-quantitative RT-PCR. None of these will be adequate for identifying reliable internal controls for real-time qRT-PCR. For example, expression profiling via real-time qRT-PCR of 10 commonly used human internal control genes revealed different degrees and patterns of expression among 13 tissue types, and no single gene was a suitable universal control for all tissue types \[[@B7]\]. Although *18S*rRNA is frequently used as an internal control, it is far from ideal. It requires the use of total RNA and random primers for the RT reaction, and is expressed at very high levels; some means for attenuating *18S*expression might be needed when weakly expressed genes are studied. In addition, there can be imbalances in rRNA and mRNA fractions between different samples, and *18S*is not always expressed at a constant level in all conditions \[[@B1]\]. Finally, *18S*expression levels appear to be affected to a lesser extent by partial RNA degradation than are mRNA expression levels \[[@B8]\]. Studies in mammalian and microbial systems, where real-time qRT-PCR has been most extensively applied to date, have begun to include evaluations of various housekeeping genes for normalization \[[@B7]-[@B11]\]. Vandesompele et al. \[[@B7]\] recognized the importance of using statistical approaches to selecting the best internal controls for a given set of samples, and developed a procedure to select internal controls based on the mean pairwise variation of a gene from all other tested control genes. The adoption of real-time qRT-PCR methodology is somewhat reminiscent of the introduction of cDNA expression microarrays in that initial microarray studies did not identify differentially expressed genes by a statistical method, but by an arbitrary cut-off value of fold-change \[[@B12]\]. Similarly, the first real-time qRT-PCR studies have generally normalized expression levels to an internal control that is assumed to be valid rather than one that has been shown to be valid by statistical analysis of data. More rigorous methods will be needed as qRT-PCR is increasingly applied to study of regulatory genes, and for verifying patterns observed in microarray experiments. In this study, we used real-time qRT-PCR to examine the expression of 10 housekeeping genes in a diversity of poplar (*Populus trichocarpa*× *P. deltoides*, cottonwood hybrid) tissues collected at different developmental stages, and at different times of the year. The goal of our studies was to detect changes associated with seasonal development and tree aging for several regulatory genes. We therefore undertook a study to compare the stability of several potential control genes. We found that the genes tested exhibited very different degrees of variation in expression among tissue samples, and that a statistical and graphical method helped us to select the genes best suited for the developmental studies we were conducting. This approach, which is very similar to a classical method used by plant breeders to assess the relative stability in yield of different varieties \[[@B13]\], can be applied to any gene or set of tissues to identify the most stable internal controls. Results and Discussion ====================== Expression profiling of poplar housekeeping genes ------------------------------------------------- Ten housekeeping genes that represent different functional classes and gene families were chosen for study. These include ubiquitins, actins, tubulins, cytosolic cyclophilin (peptidyl-prolyl isomerase), translational initiation factor, elongation factor, and rRNA. Searches of the literature revealed that members of all classes have been used as internal controls for studies of plant gene expression using RNA gel blots or RT-PCR assays. Poplar genes belonging to these gene families were identified via searches of the EST database (Table [2](#T2){ref-type="table"}). The expression level of these genes was determined in eight tissue samples (Table [1](#T1){ref-type="table"}) collected over a seven month period from mature female poplar trees growing in plantations in Oregon, USA. Within a single experiment, aliquots of the same cDNA synthesis reaction were used for real-time PCR amplification of each of the 10 genes and all gene primer and cDNA combinations were amplified in triplicate in a single PCR run. The entire experiment was then repeated a second time and results combined for statistical analysis. Quantitation via real-time PCR is based on cycle threshold (C~T~). C~T~is the cycle at which a significant increase in amount of PCR product (measured by increase in fluorescence) occurs, generally the middle of the exponential phase of amplification. Mean C~T~values (average of both experiments) for each gene are given in Table [3](#T3){ref-type="table"}. We had previously used *18S*as an internal control for expression studies using these and other tissue samples and had noticed that *18S*C~T~values sometimes varied considerably (data not shown). This may have been largely due to the high abundance of *18S*transcripts. Use of *18S*as an internal control for studies of genes expressed at relatively low levels required additional dilution of the cDNA templates for 18S amplification relative to the gene being studied. In the present study, the amount of cDNA was the same for all PCR reactions, but *18S*primer concentrations were 50 nM, while all other gene primer concentrations were 600 nM. As expected, *18S*was the most abundant (lowest C~T~) housekeeping transcript; *TUA*was the least abundant. Statistical analysis of stability of gene expression level ---------------------------------------------------------- We used single-factor ANOVA and linear regression analyses of C~T~values to examine variation among tissues and RT-PCR experiments. Examination of the distribution of the residual values from ANOVA indicated that assumptions concerning homogeneity of variance and normality of data were adequately met (data not shown). The ANOVA F-test of differences among tissues indicated that five of the genes showed significant variation in expression among the tissue samples. The degree of residual variation, as reflected in the mean square error (MSE), residuals, or coefficient of variation (CV), varied widely. Four genes had CVs below 5%, and two had CVs at or above 25% (Table [3](#T3){ref-type="table"}). The mean absolute value of the residuals (Fig. [1](#F1){ref-type="fig"}) varied 4.2-fold, from a level of 0.72 for ACT11 to 0.17 for UBQ. To test whether this variation could be due to chance alone, we tested the variation in size of residuals via Levene\'s test (Levene 1960). The variation among genes was highly significant (P \< 0.004), and the difference in residuals between ACT11 and UBQ was also significant based on Tukey\'s Studentized Range Test and the Bonferroni t-test at the 5% confidence level. The mean expression level for each gene in each tissue sample was regressed against the overall means for the different tissue samples. This overall mean provides an index of RNA quality and quantity for that tissue sample, much as means over test sites provide an index of site fertility in yield trials \[[@B13]\]. The slope provides an estimate of the degree to which the gene is sensitive to general expression-promoting conditions, and the residuals (deviation from regression prediction) and mean squared residuals (MS-Reg) estimate the degree to which expression of a gene varies unpredictably after linear effects are removed. The residual variation after regression was substantial (Figure [2](#F2){ref-type="fig"}); MS-Reg varied approximately 14-fold (Table [3](#T3){ref-type="table"}). Assuming that both constancy over tissues (low slope) and high predictability (low CV) are desired, we created a stability index as the product of slope and CV. The genes with the lowest stability index will usually provide the best controls. In this study, *UBQ*had the lowest stability index, a result of both a very low slope and very low CV. Selection of internal controls ------------------------------ In addition to constancy of expression level, the expression level of an internal control compared to that of the genes being analyzed might be important to consider in certain cases. In our study, two of the most stably expressed genes represented opposite ends of the spectrum. *UBQ*is highly expressed (mean C~T~= 15.8), whereas *TUA*is expressed at a much lower level (mean C~T~= 28.9) (Table [3](#T3){ref-type="table"}). For the samples we tested, the high stability of *UBQ*and *TUA*expression indicate that use of either as a single internal control gene is appropriate. However, for some studies, no single gene may be adequate. In these cases, a method for normalization to two or more of the most stable internal control genes identified might be necessary. For example, normalizing to the geometric mean of selected internal control genes \[[@B7]\]. A potential strategy to avoid the additional expense and labor of using multiple internal control genes is to design a PCR primer pair that will amplify two or more members of a control gene family, whose combined expression level may exhibit the desired expression level and stability. Our primers were designed based on a limited EST set that likely did not include all family members, and ESTs vary in sequence quality. Thus, primers could have amplified more than one family member or primer mismatches due to EST sequence errors could have lowered PCR efficiency. Although gel and real-time PCR dissociation curve analyses did not indicate that multiple genes were amplified with our primer sets, these analyses might not detect multiple amplicons from different family members that are the same size and have the same PCR efficiency. As discussed above this is not necessarily a detriment--amplification of multiple family members might result in a more stable internal control than single gene amplification. In addition, the upcoming release of a large poplar unigene set and annotated genome sequence \[[@B14]\] will improve gene selection and primer design capabilities. Conclusions =========== Using ANOVA and linear regression analysis, we demonstrated that levels of expression stability among a number of potential control genes can vary widely, and that it is not difficult, costly or labor-intensive to test a number of genes. Moreover, such validation tests might have the additional benefit of revealing technical problems, such as excessive variability in RT and PCR efficiency due to RNA quality or inconsistent pipetting. For some experiments, choice of an internal control is straightforward. For example, a number of housekeeping genes should be satisfactory controls for comparisons of transgene expression level in the same tissue type from different transgenic lines grown under identical conditions. However, this is not the case for studies that compare gene expression among different tissue or cell types, at different developmental stages, or under different environmental conditions, as were represented in our study of trees in field environments over a period of seven months. For such studies, internal controls should be carefully tested and validated. Statistical confirmation of internal controls for qRT-PCR should enable previously indiscernible small changes in expression level to be reliability detected. Methods ======= Tissue collection and RNA extraction ------------------------------------ Tissues were collected from five or six year-old ramets (genetically identical trees) of a single female poplar hybrid clone (*P. trichocarpa*× *P. deltoidies*) over a seven-month period in 2001 (Table [1](#T1){ref-type="table"}). The trees had been growing in commercial plantations in the Columbia River basin northwest of Portland, Oregon USA. Bud scales were removed and tissues were frozen in liquid N~2~and stored at -80°C until RNA extraction. Total RNA was isolated using the RNeasy mini kit (Qiagen, Valencia, CA, USA) with modifications. Tissues (0.2 g) were ground to a fine powder with mortar and pestle in liquid N~2~. The powder was added to a tube containing 1 ml of RNeasy RLT buffer and 0.01 g soluble polyvinylpyrrolidone (PVP-40; Sigma, St. Louis, MO, USA), and homogenized using a polytron for approximately 30 sec. Four volumes of 5 M Potassium acetate, pH 6.5 was added to the homogenate, the mixture was incubated on ice for 15 min, and the precipitate removed by a 15 min centrifugation (12,000 rpm) at 4°C. Supernatant was transferred to two 1.5 ml microcentrifuge tubes and 0.5 volume of 100% EtOH was added. Samples were transferred to RNeasy mini columns and the remaining steps were as directed by the manufacturer\'s instructions for plant RNA isolation (steps 6--11). RNA was quantified using spectrophotometric OD~260~measurements and quality was assesed by OD~260~/ OD ~280~ratios and by electrophoresis on 1% formaldehyde agarose gels followed by ethidium bromide staining. RNAs were stored at -80°C. Selection of poplar sequences and PCR primer design --------------------------------------------------- To identify poplar homologs of genes commonly used as controls for plant gene expression studies, we queried poplar EST databases with Arabidopsis protein sequences using TBLASTN \[[@B15]\]. Selected poplar ESTs were then used to query the Arabidopsis protein database using BLASTX (Table [2](#T2){ref-type="table"}). Primers were designed using Primer3 software \[[@B16]\] or Primer Express (Applied Biosystems, Foster City, CA, USA) with melting temperatures of 59--60°C. By comparison to related poplar EST sequences, primers were designed to be as specific as possible for the selected gene family member. All primer pairs were initially tested via standard RT-PCR using the same conditions as described below for real-time RT-PCR. Amplification of single products of expected size was verified by electrophoresis on 3% agarose-1000 (Invitrogen, Carlsbad, CA, USA) and ethidium bromide staining. Real-time RT-PCR ---------------- Contaminating DNA was removed from RNA samples using the DNA-Free kit (Ambion, Austin, TX, USA) according to the manufacturer\'s protocol, and two-step real-time RT-PCR performed. cDNA was synthesized from 5 μg of RNA using the SuperScript first-strand synthesis system for RT-PCR (Invitrogen) with random hexamer primers according to the manufacturer\'s instructions, except that the initial 65°C denaturation step was omitted. The cDNAs were diluted 1:5 with nuclease-free water. Aliquots of the same cDNA sample were used with all primer sets for real-time PCR, and amplification reactions with all primer sets were performed in the same PCR run. Reactions were done in a 25 μl volume containing 600 nM of each primer, 6.5 μl of cDNA sample (≈320 ng of input RNA) and 1X SYBR Green PCR master mix (Applied Biosystems). For 18S amplification, primer concentration was 50 nm. Real-time PCR was performed on the ABI Prism 7700 Sequence Detection System (Applied Biosystems) in a 96-well reaction plate using the parameters recommended by the manufacturer (2 min. at 50°C, 10 min. at 95°C and 40 cycles of 95°C for 15 sec and 60°C for 1 min.). Each PCR reaction was performed in triplicate and no-template controls were included. Specificity of the amplifications was verified at the end of the PCR run using ABI Prism Dissociation Curve Analysis Software. The entire experiment, including both the RT and real-time PCR steps, was repeated, giving a total of two experimental replications. Statistical analyses -------------------- Results (C~T~values) from the ABI PRISM 7700 Sequence Detection System were analyzed in Microsoft Excel. Single factor ANOVA and regression analysis using the least squares method were performed using the Excel Analysis ToolPak. Assumptions concerning homogeneity of variance and normality were evaluated from inspection of residuals (the difference between an observed value and overall mean for all genes) from the ANOVA (Fig. [1](#F1){ref-type="fig"}). The level and significance of the difference between gene expression levels in different samples were evaluated by Fisher\'s F statistic \[F = between-tissue-sample mean square / error mean square\] assuming the three replicate PCR reactions approximated variance between fully independent observations. Other statistics are as defined in Table [3](#T3){ref-type="table"}. The general procedure for data analysis to compare genes for use as internal controls was: 1\) Generate data from multiple analyses of gene expression via quantitative RT PCR that can be assumed to be statistically independent (or nearly so), including from multiple independent samples that bracket the experimental conditions of interest. 2\) Conduct ANOVA to examine the extent of variation among samples, and (optionally) test their significance using appropriate F-ratios. Examine plot of residuals vs. mean expression level, or use a statistical test, to check normality of data. 3\) Genes showing high variance among tissues in ANOVA, especially if accompanied by large mean square errors (or coefficients of variation), are to be avoided as controls. 4\) Calculate mean expression level over all genes studied for each sample type as an index of both experimental and biological conditions that promote high levels of measured expression. Use of a large number of genes and tissue samples (e.g., at least five, and preferably many more) are desirable where estimates of stability are to be compared between studies. 5\) Regress mean expression level for each gene in each sample type over the mean for the sample type. The estimated slope and mean square residual (deviation from regression prediction) provide estimates of the degree to which the gene is sensitive to general expression-promoting conditions (slope) and whose expression continues to be difficult to predict (residual). 6\) Assuming that both constancy over sample types (low slope) and high predictability (low coefficient of variation) are desirable, a stability index can be created as their product (or via other mathematical means), and the gene with the lowest value chosen. 7\) Alternatively, visually inspect regression and residual plots to select genes that would be most suitable as controls for specific sets of experimental conditions. Authors\' contributions ======================= AMB developed the molecular methods, participated in design of the study, and drafted the manuscript. IAY designed primers, carried out the real time RT-PCR experiments, performed the statistical analysis and drafted the manuscript. SHS conceived the analysis, and participated in its design. All authors read, helped to edit, and approved the final manuscript. Acknowledgements ================ We thank Olga Shevchenko for isolating the RNAs, Dr. Brian Stanton and Greenwood Resources (formerly Fort James Corporation) for permission to study their plantation trees, and Ove Nilsson and Frances Martin for providing poplar EST sequences prior to public release. I.A.Y. was supported by a Fullbright Fellowship. Partial funding for this work came from the Tree Biosafety and Genomics Research Cooperative at Oregon State University and the USDA NRI Competitive Grants Research Program (No. 2002-35301-12173). Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Scatterplot of residuals after regression of tissue means from each experiment on overall mean for all genes.**Genes were ordered based on absolute value of mean residuals (increasing from left to right). :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Regression lines for several genes showing predicted regression lines and actual means over both experiments.**The most stable and consistent control genes would have the lowest slope and closest fit to the regression line. *UBQ*(second from bottom) had the highest and *TUA*(top) the second highest stability indices in this experiment. *CYP*(third from bottom) had the lowest stability index. See Table 1 for descriptions of tissue samples. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Poplar tissues used for gene expression studies. ::: **Sample name** **Collection date (2001)** **Tissue description** ----------------- ---------------------------- ------------------------------------------------------------------------------- VB1 March 20 Overwintered terminal vegetative buds VB2 April 3 Overwintered terminal vegetative buds approximately 1 week prior to bud flush S April 18 Newly expanding shoots (average shoot elongation = 38 mm) ST1 May 3 Shoot tips, including unexpanded leaves ST2 June 19 Shoot tips, including unexpanded leaves VB3 August 7 Terminal vegetative buds VB4 October 15 Terminal vegetative buds FB October 15 Inflorescence buds ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Description of poplar genes and primers for real-time PCR. ::: ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- **Name**^a^ **GenBank accession number**^b^ ***Arabidopsis*homolog locus**^c^ ***Arabidopsis*locus description** **BLASTX score/ E value** **Primer sequences (forward/reverse)** ------------- --------------------------------- ----------------------------------- ----------------------------------------------------------- --------------------------- ------------------------------------------------ *ACT11* [CA824001]{.underline} [AT3G12110]{.underline} Actin 11 363/ e-101 CACACTGGAGTGATGGTTGG / ATTGGCCTTGGGGTTAAGAG *ACT2* [BU879695]{.underline} [AT5G09810]{.underline} Actin 2/7 320/ 5e-088 CCCATTGAGCACGGTATTGT / TACGACCACTGGCATACAGG *CYP* [BU875027]{.underline} [AT2G21130]{.underline} cyclophilin (CYP2) 284/ 6e-077 GGCTAATTTTGCCGATGAGA / ACGTCCATCCCTTCAACAAC *TUA* [CA822230]{.underline}\ [AT5G19780]{.underline} tubulin alpha-3/alpha-5 chain 439/ e-130 AGGTTCTGGTTTGGGGTCTT / TTGTCCAAAAGCACAGCAAC [CA825391]{.underline} *TUB* [CA824237]{.underline} [AT4G20890]{.underline} tubulin beta-9 chain 154/ 4e-038 GCACCAACTTGTTGAGAATGC / TTTCAACTGACCAGGGAACC *UBQ* [BU879229]{.underline} [AT4G05050]{.underline} polyubiquitin (UBQ11) 416/ e-117 GTTGATTTTTGCTGGGAAGC / GATCTTGGCCTTCACGTTGT *UBQ-L* [BU871588]{.underline} [AT2G35635]{.underline} ubiquitin-like (UBQ7) 291/ 4e-079 TGAGGCTTAGGGGAGGAACT / TGTAGTCGCGAGCTGTCTTG *EIF4B-L* [CA825614]{.underline} [AT4G38710]{.underline} similarity to eukaryotic translation initiation factor 4B 80/ 1e-015 AAAAAGGGGATTTGGGATTG / AACTTCGTCCTCGGTAGCAA *EF1β* [BI125345]{.underline} [AT2G18110]{.underline} elongation factor 1-beta, putative 122/ 1e-028 AAGAGGACAAGAAGGCAGCA / CTAACCGCCTTCTCCAACAC *18S* [AF206999]{.underline} [18RRNA]{.underline} 18S ribosomal RNA 2949/ 0.0 AATTGTTGGTCTTCAACGAGGAA/ AAAGGGCAGGGACGTAGTCAA ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ^a^All poplar sequences except 18S are ESTs, and were named based on similarity to *Arabidopsis*proteins determined via BLASTX (Altschul et al. 1997). In most cases, the name indicates only a gene family or subfamily rather than a specific member of a gene family because partial poplar sequences and BLAST will not necessarily identify the putative *Arabidopsis*ortholog. ^b^Two accession numbers indicate that two EST sequences were used to design the primer set. ^c^Closest *Arabidopsis*homolog identified using Tair BLAST 2.0 <http://www.arabidopsis.org/Blast/>. AGI proteins database was queried with poplar nucleotide sequences using BLASTX or in the case of 18S, *Arabidopsis*genome database with BLASTN. ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Summary of statistics measuring stability of gene expression ::: **GENE** **MEAN**^a^ **F**^b^ **MSE-ANOVA**^c^ **CV (%)**^d^ **SLOPE**^e^ **INTERCEPT** **R**^2^ **MS-REG**^f^ **STABILITY INDEX**^g^ --------------- ------------- ----------- ------------------ --------------- -------------- --------------- ---------- --------------- ------------------------ ***UBQ*** 15.8 1.42 0.53 3.4 0.11 15.3 0.56 0.49 0.37 ***TUA*** 28.9 1.33 1.56 5.4 0.10 28.7 0.28 1.65 0.54 ***18S*** 14.0 4.72\*\* 0.58 4.1 0.20 13.2 0.52 0.38 0.83 ***ACT2*** 17.3 1.29 1.25 7.2 0.16 16.6 0.57 1.14 1.16 ***UBQ-L*** 19.8 1.97 1.57 7.9 0.17 19.1 0.35 1.62 1.35 ***EF1β*** 25.1 3.76\*\* 1.87 7.5 0.28 23.8 0.6 1.41 2.09 ***TUB*** 17.7 4.45\* 2.7 15.3 0.37 16 0.63 1.92 5.64 ***ACT11*** 22.8 1.01 5.71 25.0 0.24 21.7 0.37 5.29 6.01 ***EIF4B-L*** 20.6 11.44\*\* 3.22 15.6 0.52 18.2 0.68 1.79 8.13 ***CYP*** 19.2 7.35\*\* 5.14 26.8 0.67 16.2 0.84 2.66 17.94 ^a^Data based on analysis of C~T~values. Genes are ordered, top to bottom, from those tending to show the highest stability to those showing the lowest, based on the stability index. ^b^Approximate F-tests of variance among tissue samples tested. \*, P \< 0.05; \*\*, P \< 0.01. Degrees of freedom for numerator were 7 and for denominator were 40, except for 18S RNA where they were 7 and 16, respectively. ^c^MSE-ANOVA represents variance among experiments and RT-PCR reactions within experiments. 18S amplification was only included in one experiment; thus, MSE-ANOVA for 18S only represents within experiment variance. ^d^Coefficient of variation (MSE divided by mean multiplied by 100). ^e^Slope of regression of gene means (over experiments and samples within experiments) against overall means for the different samples. Intercepts and coefficient of determination (R^2^) are also given for the estimated regression lines. ^f^Mean square of deviation of means from estimated regression line (MS-reg), which estimates the degree to which genes deviate from the linear model in their level of mean expression for a particular tissue sample. ^g^Stability index is the product of CV and slope (multiplication of columns 4 and 5). Genes whose expression shows the lowest random variation within tissue samples due to variation among experiments or PCR reactions (MSE-ANOVA), and whose expression depends least in a predictable way on tissue sample (slope), are preferred as controls. :::

PubMed Central

2024-06-05T03:55:47.328219

2004-8-18

PMC515302

Background ========== Cluster randomised trials are those where research subjects are not allocated to treatments independently, but as a group. For example, in a study of counselling patients on physical activity in general practice, practices were allocated to counselling or control and patients aged 40--79 years who attended during a five day period and who did not take regular exercise were invited to take part. Patients in the same practice received the same treatment, counselling or usual care, depending on how the practice was allocated. \[[@B1]\] The group of patients within the general practice formed a cluster. Members of a cluster will be more like one another than they are like members of other clusters and we need to take this into account in the analysis, and preferably the design, of the study. Methods which ignore clustering may mislead, because they assume that all subjects provide independent observations. Applying simple statistical methods to such data, without taking the clustering into account, can lead to confidence intervals which are too narrow and P values which are too small. There has been an increasing interest in cluster randomised trials over the past 20 years. For example, by the end of 2003 the *British Medical Journal*Statistics Notes on this topic \[[@B2]-[@B7]\] had been cited 121 times. There have been several reviews of published cluster randomised trials \[[@B8]-[@B13]\] (Table [1](#T1){ref-type="table"}). All but Puffer *et al.*\[[@B10]\] reported that very few trials had sample size calculations which included clustering and about half took clustering into account in the analysis, fewer in the African studies reported by Isaakidis and Ioannidis. \[[@B11]\] Puffer *et al.*\[[@B10]\] did not mention whether trials failed to take clustering into account in the analysis. My own review of their trials as listed on the *British Medical Journal*website found that only 3 out of 36 ignored clustering. The review of the *American Journal of Public Health*and *Preventive Medicine*in 1998 -- 2002 \[[@B13]\] is especially interesting because it attempted to replicate an earlier study \[[@B9]\] in the same journals. There was an increase in the number of reports of cluster randomised trials: 12.3 studies were reported per year in 1998 -- 2002 compared to 5.3 studies per year in 1990 -- 1993. \[[@B9]\] The quality of the analysis may have improved, but such assessments are subjective and very difficult to compare between reviews. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Some reviews of published cluster randomised trials ::: Authors Source of trials Years Clustering allowed for in sample size Clustering allowed for in analysis -------------------------------- ----------------------------------------------------------------------------------------- ------------------------------- --------------------------------------- ----------------------------------------------- Donner *et al.*\[8\] 16 non-therapeutic intervention trials 1979 -- 1989 \<20% \<50% Simpson *et al.*\[9\] 21 trials from *American Journal of Public Health*and *Preventive Medicine* 1990 -- 1993 19% 57% Isaakidis and Ioannidis \[11\] 51 trials in Sub-Saharan Africa 1973 -- 2001 (half post 1995) 20% 37% Puffer *et al.*\[10\] 36 trials in *British Medical Journal*, *Lancet*, and *New England Journal of Medicine* 1997 -- 2002 56% 92% ^a^ Eldridge *et al.*\[12\] 152 trials in primary health care 1997 -- 2000 9% 59% Varnell *et al.*\[13\] 60 trials in *American Journal of Public Health*and *Preventive Medicine* 1998 -- 2002 20% 54% (all analyses) + 25% (some analyses only) ^a^My review of trials identified by Puffer *et al.*\[10\] ::: It is understandable that papers do not report sample size calculations, as often these are omitted from papers entirely, sometimes by the request of the journal to save space. It can be argued (though I would not do so) that once we have carried out a study, the sample size calculations are not particularly informative. Analysis which ignores the clustering, however, can be highly misleading, finding significant differences where there are none. We may have incorrect conclusions in the literature, which are then uncritically repeated and become false knowledge. We should not be surprised that clustering is ignored. In the past, few textbooks have cautioned against this and the assumption of independence of observations is seldom stressed. Many statisticians will admit to having incorrectly ignored clustering in the analysis of clustered designs, including myself when I was younger and more ignorant than today. However, it can be very important. In this paper I attempt to chart the changes in both the number of cluster randomised trials reported and the proportion of these reports where clustering has been taken into account in the analysis. Methods ======= I first carried out a search on the ISI Web of Science, looking for papers on cluster randomisation and reports of trials. I classified these by type (trial report or methodological article), year of publication, and journal. To identify cluster randomised trials we have to read the papers. We cannot tell whether a trial is cluster randomised from title, keywords, or abstract. Many authors are not aware of the importance of clustering and do not mention it. In this paper I report the results of a hand search of the *British Medical Journal*. I identified and scanned all papers reporting trials for the years 1983, 1988, 1993, 1998, and 2003, recording any where subjects were allocated in clusters. I excluded any studies where subjects were not allocated to groups by the investigator, for example several comparisons of fund-holding and non-fund-holding general practices. For each trial identified, I noted whether clustering had been taken into account in the analysis. There are several approaches which can be used to allow for clustering. The easiest is to calculate a summary statistic for each cluster. \[[@B4]\] This is usually a mean for a continuous outcome or a proportion for a dichotomous outcome. We can also use robust variance estimates, general estimating equation models (GEEs), multilevel modelling, Bayesian hierarchical models, and several other techniques. Any method which takes into account the clustering should be an improvement compared to methods which do not. I also noted whether ignoring the possible effects of clustering might have an important effect on the conclusions. Clustering may result in P values and confidence intervals which are sufficiently biased to have a major effect if any of the following are true: the cluster size is large, the number of clusters is small, or the intra-cluster correlation coefficient is large. Whether any of these applies in a trial which ignores clustering is a matter of judgement. Results ======= A computer search for cluster randomised trials ----------------------------------------------- Figure [1](#F1){ref-type="fig"} shows the result of a search on the ISI Web of Science, looking for papers on cluster randomisation and reports of trials. I found that other terms, such as \'group randomised\' did not work, as I got hundreds of abstracts with \'patients were in two groups, randomised to active or control treatments\'. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Results of a Web of Science search.**Results of a Web of Science search on: randomi\* in clusters OR cluster randomi\*, up to the end of 2003. :::  ::: There are many potential biases. In the first part of the period, the Web of Science database did not include abstracts, so there was less opportunity to pick up the search terms. More recently, several journals began to include a description of the trial design in the title of the paper, for example \'Effect on hip fractures of increased use of hip protectors in nursing homes: cluster randomised controlled trial\'. \[[@B14]\] This will increase the detection rate. These design descriptions are not always correct, nor does the cluster randomised nature of the trial necessarily appear in the description. Also, many authors will not be aware of the importance of clustering and will not mention it. These factors will reduce detection. Hence this is not a thorough search and will have missed many studies, but it might give an idea of the increase in activity. I divided the papers into those which were methodological, either educating researchers into the appropriate design and analysis of cluster randomised trials or developing new methods of analysing such trials, and those reporting actual trials. The data for 2000 and 2001 includes special issues of *Statistics in Medicine*and *Statistical Methods in Medical Research*on cluster randomisation, so there were a larger number of methodological papers than might be expected in those years. The numbers of papers found in the two categories were similar in each year before 2001: as many papers were about how to do such trials as were reports of actual trials. It is hard to believe that there are so few such trials being reported and it is likely that many have been reported without any acknowledgement of the importance of clustering. All the papers up to 1990 are due to Donner and his colleagues. \[[@B8],[@B15],[@B16]\] However, it was impossible to identify papers which used older terminology. A paper by Cornfield \[[@B17]\] \'Randomisation by group: a formal analysis\' includes the following statement \'Randomization by cluster accompanied by an analysis appropriate to randomization by individual is an exercise in self-deception, however, and should be discouraged.\' This would not be found by the search. The book on cluster randomization by Murray \[[@B18]\] is called *The Design and Analysis of Group-Randomized Trials*. The *British Medical Journal*was the journal most frequently represented in the survey, no fewer than 43 of the 332 publications found, 13%, appearing there. The next was *Statistics in Medicine*with 39 publications (12%), all methodological, then the *British Journal of General Practice*with 17 (5%), *Controlled Clinical Trials*with 16 (5%), and *Family Practice*with 10 (3%). Figure [1](#F1){ref-type="fig"} shows that papers in the *British Medical Journal*reflect the literature as a whole. The first paper in the *BMJ*was a report of a trial, \[[@B19]\] which was followed four years later by a series of short educational articles. \[[@B3],[@B4],[@B6],[@B7]\] When I first did the search reported in Figure [1](#F1){ref-type="fig"}, I was surprised by how few trials were reported. My subjective impression was that there were many more cluster randomised trials than I had found. I therefore decided to carry out a small survey of journals to find out whether the dramatic increase shown in Figure [1](#F1){ref-type="fig"} was real. As the *British Medical Journal*had most reports and had been published for many years, this was the obvious journal with which to begin. A survey of papers in the British Medical Journal ------------------------------------------------- The results of the search are shown in Table [2](#T2){ref-type="table"}. As Table [2](#T2){ref-type="table"} shows only reports of trials, it does not include all the *BMJ*papers in Figure [1](#F1){ref-type="fig"}, which also includes methodological papers. A list of all papers reviewed is given in the additional file: papers in the survey. Only one of the trials in survey \[[@B1]\] cited any of the *BMJ*Statistics Notes on clustering. \[[@B2]-[@B7]\] ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Result of a hand search for cluster randomised trials in the *British Medical Journal* ::: Year Vol Trials Clustering ignored Ignoring clustering judged as important Found in Web of Science search ------ ------- -------- -------------------- ----------------------------------------- -------------------------------- 2003 326-7 9 0 0 5 1998 316-7 4 1(?) 1 0 1993 306-7 4 3 2 0 1988 296-7 0 0 0 0 1983 286-7 1 1 1 0 ? doubtful whether clustering taken into account ::: The noted query relates to a paper in which the authors stated that \'Univariate comparisons were calculated by t test and χ^2^analysis. The role of potential covariates was explored using linear regression specified as a two level model (practice and individual) using the software package MLn\'. \[[@B20]\] I could find no multilevel modelling in this paper, but a lot of t and χ^2^tests. This was a trial of community based management in failure to thrive by babies. Thirty eight primary care teams were randomly allocated to intervention or control and all children identified in the practice were offered the same intervention, so clearly clustering should be taken into account. The trials which I regarded as failing to take the clustering into account were as follows. Russell *et al.*\[[@B21]\] investigated the effect of nicotine chewing gum as an adjunct to general-practitioners advice against smoking. Subjects were \'assigned by week of attendance (in a balanced design) to one of three groups (a) non-intervention controls (b) advice and booklet (c) advice and booklet plus the offer of nicotine gum.\' There were 6 practices, with recruitment over 3 weeks, one week to each regime. The study was analysed by chi-squared tests. As the clusters were large, with 1938 subjects in 18 clusters, clustering should have been taken into account. Rink *et al.*\[[@B22]\] investigated the impact of introducing near patient testing for standard investigations in general practice. Twelve practices were used, and some given the equipment and some not in a cross-over design. Analysis used paired t tests, unpaired t tests, odds ratios, ratios of proportions with confidence intervals, and chi squared tests, none of which took clustering into account. In a trial of clinical guidelines to improve general-practice management and referral of infertile couples, Emslie *et al.*\[[@B23]\] randomised 82 general practices in Grampian region and studied 100 couples in each group. However, the main outcome measure was whether the general practitioner had taken a full sexual history and examined and investigated both partners appropriately. The cluster size may be small but the cluster effect may be large. The GP should be the unit of analysis here as opposed to the couple, as done in the paper. The trial where I judged ignoring clustering to be unimportant had many very small clusters. Wetsteyn and Degeus \[[@B24]\] compared 3 regimens for malaria prophylaxis in travellers to Africa. Members of one family were allocated to one regimen and the results analysed using a chi-squared test. Only five of the 18 trials had been found in the Web of Science search, showing that that was indeed an underestimate. However, the growth in numbers of trials is indicated by both electronic and hand searches. Discussion ========== A bibliometric survey has suggested a rapid increase in the number of cluster randomised trials, many of which appeared in the *British Medical Journal*. A hand search of the *British Medical Journal*has confirmed this increase, at least in this journal. Although the effects of clustering have often been ignored in trials, producing potentially misleading conclusions, the situation has certainly improved in the *British Medical Journal*. This has followed many articles on the topic in the Journal. Perhaps statistician pressure works. Identification of cluster allocation is subjective. I included one year, 1998, also searched by Puffer *et al.*\[[@B10]\] and identified four trials where Puffer *et al.*\[[@B10]\] identified only one. My assessments of whether clustering has been taken into account and whether ignoring it might be important are also subjective. Nevertheless, I think that the general conclusion of increasing activity and better reporting of trials, at least in the *British Medical Journal*, is valid. Whether we would find a similar improvement in other journals is less certain. It is likely that reporting of cluster randomised trials in the *British Medical Journal*is especially good, as the journal reports many such trials, has carried many articles on their correct analysis and reporting, has a fairly rigorous statistical refereeing system, \[[@B25]-[@B27]\] and is generally of a relatively high methodological standard. The *BMJ*\'s current statistical checklist \[[@B28]\] does not mention clusters, however. It would be possible to extend the survey to other journals where such trials are frequently reported, but these, too, might be more likely to adhere to sound principles of analysis and reporting than would journals where few such studies appear. The thought of hand-searching journals where no trials might be found does not appeal. There are still many other aspects of trial reporting where improvement is possible \[[@B10],[@B12]\] but the picture drawn by this survey is encouraging. Methodologists need to keep up the pressure and to extend it to specialist journals. The recently published extension of the CONSORT statement to cluster randomised trials is to be welcomed. \[[@B29]\] We should also pursue other types of study where the unit of analysis is doubtful, such as those involving observations of multiple body parts in the same patient or multiple measurements on the same tissue treated as independent. Conclusions =========== Cluster trials have become much more frequent since the mid 1990s. Reporting of these trials has improved and in the journal which publishes more than any other the quality had improved greatly. This improvement has followed a large number of articles advocating methods of analysis which take clustering into account, Perhaps statistician pressure works. Competing interests =================== The author has been published frequently in the *British Medical Journal*, including articles written for payment. Authors\' contributions ======================= J. M. Bland is the sole contributor. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2288/4/21/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional File 1 **Papers in the survey**List of papers in the survey of the BMJ, Word file. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ I thank Douglas Altman, Sandra Eldridge, Sally Kerry, Jeremy Miles, and David Torgerson for helpful comments on a draft.

PubMed Central

2024-06-05T03:55:47.331562

2004-8-13

PMC515303

Background ========== Acute renal failure requiring hemodialysis (HD) is a common clinical problem in critically ill patients that is independently associated with increased mortality\[[@B1]\]. Dialysis disequilibrium syndrome (DDS) is the clinical phenomenon of acute central nervous system dysfunction attributed to cerebral edema that occurs during or following HD. The precise epidemiology of DDS is poorly defined\[[@B2]\]. Review of MEDLINE (January 1966 -- March 2004) suggested that DDS in critically ill patients has rarely been reported\[[@B3],[@B4]\]. We report a case of DDS-induced cerebral edema that resulted in irreversible brain injury and death following acute HD. Further, we review the relevant literature of the association of DDS and HD. Case presentation ================= A 22-year-old homosexual male presented to hospital with progressive dyspnea, productive cough, generalized malaise and fever. He had a known history of intravenous cocaine abuse and recent serology in prior 3 months was negative for human immunodeficiency virus (HIV). Results of a physical examination showed signs of tachypnea, tachycardia, accessory muscle use and left lung base crackles. Tympanic temperature was 34.7°C. The remainder of the examination was unremarkable except for urethral meatus stenosis. Initial laboratory investigations are presented in Table [1](#T1){ref-type="table"}. Arterial blood gases showed pH of 6.95, PaCO~2~10 mmHg, PaO~2~109 mmHg, HCO~3~2 mmol/L, and lactate 0.6 mmol/L consistent with high anion gap metabolic acidosis with respiratory compensation. Serum creatinine and blood urea nitrogen were 587 μmol/L and 46.7 mmol/L, respectively. Toxicology and drug screen was negative. The metabolic acidosis was partially accounted for by acute renal failure with retained unmeasured anions and ketonemia. Urinalysis showed pyuria. Electrocardiogram (ECG) showed normal sinus rhythm. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Laboratory values at prior to and following initiation of hemodialysis in the intensive care unit. ::: **Laboratory test** **Pre-dialysis Value** **Post-dialysis Values** **Reference range** ---------------------- ------------------------ -------------------------- --------------------- **Hemoglobin** 96 78 137--180 g/L **White cell count** 25.8 16.1 4.0--11.0 × 10^9^/L **Band count** 3.1 \- 0.0--1.3 × 10^9^/L **Platelets** 603 486 150--400 × 10^9^/L **Sodium** 132 132 133--145 mmol/L **Potassium** 3.1 1.8 3.5--5.0 mmol/L **Chloride** 107 93 98--111 mmol/L **Bicarbonate** 2 19 21--31 mmol/L **Glucose** 6.3 9.0 3.6--11.1 mmol/L **Magnesium** 0.88 0.57 0.65--1.15 mmol/L **Osmolality** 330 \- 280--300 mosmol/kg **Urea** 46.7 13.7 3.0--7.6 mmol/L **Creatinine** 537 \- 61--111 μmol/L **Lactate** 0.6 1.2 \< 2.0 mmol/L **Serum ketones** 2+ \- Undetected **Anion gap** 23 20 12--14 **Osmolar gap** 14.5 \- 0--10 ::: Chest radiograph revealed right middle lobe and lingular patchy opacification. An abdomino-pelvic CT scan demonstrated moderate to severe bilateral hydronephrosis, bladder wall thickening with multiple diverticuli, and retroperitoneal streaking consistent with acute infection. A provisional diagnosis of severe sepsis was made with multiple potential foci of infection. The patient was given empiric ceftriaxone, metronidazole and vancomycin. Sputum specimen cultured heavy methicillin-sensitive *Staphylococcus aureus*, blood cultures were positive for *S. aureus*, *Escherichia coli*, and Group B *Streptococcus*. Urine cultured greater than 10^8^CFU/L of multiple gram positive and negative organisms. The patient was admitted to the intensive care unit (ICU). The metabolic acidosis persisted (pH 7.00) a despite 100 mEq of 8.4% sodium bicarbonate bolus and infusion of three liters of normal bicarbonate solution (150 mEq of 8.4% sodium bicarbonate in 1000 mL D5W). The patient had a suprapubic bladder catheter inserted by angiography. However, due to concern the patient remained oliguric following 4 L crystalloid resuscitation, hemodialysis was organized. Hemodialysis parameters included: F160 membrane (surface area 1.5 m^2^and KUf 50 mL/hr/mmHg), dialysate sodium 136 mmol/L, potassium 3 mmol/L, calcium 1.25 mmol/L, bicarbonate 40 mmol/L, and Q~D~500 mL/min, Q~B~250--300 mL/min via a 25 cm left femoral double-lumen Uldall catheter. The patient had 71.5 L of blood processed over four hours with no fluid removal. Although the patient was alert and appropriate (Glasgow Coma Scale 15) with tachycardia and stable normal range blood pressure before the initiation of dialysis, he was demonstrating an increased work of breathing and oxygen requirements suggestive of worsening sepsis syndrome. Approximately 2.5 hrs after start of dialysis the patient became rapidly unresponsive prompting intubation for airway protection. At completion of HD and over the subsequent 4 hours the patient\'s neurologic status deteriorated with evidence of loss of all brainstem reflexes. Head CT-scan is shown in Figure [1](#F1){ref-type="fig"}. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Computerized Tomography (CT) head showing diffuse cerebral edema with effacement of basal cisterns and generalized loss of gray-white differentiation :::  ::: Repeat laboratory investigations immediately following hemodialysis revealed a pH 7.36, HCO3 19 mmol/L, sodium 132 mmol/L, potassium 1.8 mmol/L, and urea 13.7 mmol/L (urea-reduction-ratio was 71%) (Table [1](#T1){ref-type="table"}). The patient rapidly progressed to refractory shock and multi-organ dysfunction Diagnosis of brain death was declared independently by an intensivist and a neurologist. At autopsy, the brain showed evidence of diffuse cerebral edema. Cardiac assessment showed left ventricular enlargement consistent with systemic hypertension likely as a result of chronic kidney disease. Both lungs showed patchy acute bronchopneumonia with edema and congestion. Both kidneys appeared grossly pyonephrotic with dilated, thickened ureters and suggested the presence of acute on chronic pyelonephritis. The meatal aperture was scarred and stenosed. Discussion ========== The immediate indication for renal replacement therapy was correction of refractory metabolic acidosis in the setting of oliguria; however, following initiation of HD this patient developed irreversible symptoms consistent with DDS. DDS occurs most commonly following initiation of chronic HD for patients with end-stage renal disease\[[@B2]\]. Patients with pre-existing neurologic disease, such as head trauma, stroke or malignant hypertension, may be at greater risk for developing DDS\[[@B5],[@B6]\]. The precise epidemiology of DDS is poorly defined and may be under-reported due to the wide spectrum of clinical manifestations. Mild symptoms such as headache, nausea, blurred vision, muscle cramps, disorientation, anorexia, restlessness, hypertension and dizziness are common during or following HD and may be attributed to DDS\[[@B2]\]. More severe symptoms consistent with central nervous system dysfunction such as seizures, central pontine myelinolysis, coma and death are rare\[[@B7]\]. The temporal profile for DDS is not well described. DDS has been credited for acute electroencephalographic (EEG) abnormalities and structural changes on diagnostic imaging following rapid hemodialysis \[[@B8]-[@B10]\]. Likewise, brain MRI studies immediately following hemodialysis in chronic dialysis patients have shown quantitative increases in brain volume consistent with cerebral edema\[[@B11]\]. The pathogenesis remains debated and incompletely understood; however, two central hypotheses have emerged. First, acute urea removal occurs more slowly across the blood-brain barrier than from plasma, generating a \'reverse osmotic gradient\' promoting water movement into the brain and cerebral edema\[[@B12]\]. Absolute increases in brain water content have been demonstrated in a rat model of uremia undergoing rapid hemodialysis that was accounted for by an increase in the ratio of brain to plasma urea\[[@B13],[@B14]\]. Down-regulation of central nervous system urea transporters have been proposed as a mechanism contributing to the delay in urea clearance from the brain\[[@B15]\]. The second hypothesis states that the increased osmolality of the extracellular fluid in uremia stimulates an adaptive accumulation of intracellular organic osmolytes to limit cerebral cell dehydration\[[@B16]\]. During hemodialysis, retention of these organic osmolytes contributes to a paradoxical reduction in intracellular pH resulting in increased brain osmolality and cerebral edema\[[@B17],[@B18]\]. The patient in this case unfortunately may have been susceptible to both proposed pathophysiologic mechanisms. The likely and under-appreciated presence of pre-existing kidney disease (chronic obstructive nephropathy and pyelonephritis) with an increased serum osmolality would have resulted in adaptive changes in the central nervous system. Ensuing hemodialysis correction of the plasma metabolic acidosis may have eclipsed a more severe cerebral intracellular acidosis. Further, urea clearance by hemodialysis was efficient at approximately 70% and probably generated a sufficient plasma-to-brain urea gradient for development of cerebral edema, intracranial hypertension and DDS. A less efficient initial course of hemodialysis would have diminished the osmolar gradient of urea across the central nervous system reducing the likelihood of symptoms of DDS. Other variables may have contributed. The patient was compensating for the severe metabolic acidosis by hyperventilation (PaCO~2~10 mmHg); however, initial post-intubation PaCO~2~was 42 mmHg. Rapid elevations in PaCO~2~can alter cerebral autoregulation resulting in exacerbated intracranial hypertension\[[@B19]\]. Concomitant sepsis syndrome with poly-microbial bacteremia resulting in widespread immune activation may alter blood-brain-barrier permeability and contribute to cerebral edema\[[@B20],[@B21]\]. These factors likely contributed to an increased risk for DDS-induced cerebral edema. The symptoms of DDS have been ameliorated by several interventions targeted to reduce the hemodialysis-induced plasma-to-brain osmotic gradient promoting cerebral edema\[[@B22]\]. A similar case of severe DDS requiring intubation was prevented from recurring during subsequent hemodialysis by use of modified dialysate containing 10.1 mmol/L of urea\[[@B23]\]. Likewise, the administration of intravenous mannitol and hyperventilation reversed a case of severe DDS-induced central nervous system dysfunction in a patient undergoing initial hemodialysis for acute renal failure\[[@B3]\]. Conversely, sodium profiling, high sodium or hyperglycemic dialysate have been attempted with variable results\[[@B24],[@B25]\]. Prevention of DDS is traditionally the mainstay of therapy, particularly during initiation of hemodialysis in new patients. Despite the absence of evidence-based guidelines, the conventional aim is for a gradual clearance of urea. This can be accomplished with intermittent hemodialysis by use of a smaller, less efficient dialyzer and by reducing the duration of initial dialysis to approximately 2 hours with targeted lower blood flow rates of 150--200 mL/min, use of sustained low-efficiency dialysis (SLED), or initiation of continuous renal replacement therapy (CRRT) with more gradual and stable clearance of urea\[[@B2],[@B26]-[@B28]\]. As a result, DDS has not been reported with the use of SLED or CRRT in critically ill patients. By providing a shorter, less efficient trial of initial hemodialysis, the severe DDS and brain death in this case may have been altogether prevented. In summary, the precise epidemiology and pathophysiology of DDS remain unclear. Although DDS usually presents in end-stage renal disease patients undergoing initial therapy, critically ill patients may represent a unique population where co-existing illnesses such as sepsis, brain injury or other central nervous system disease, multiorgan dysfunction, and need for sedation can present obstacles for prompt diagnosis of DDS. Furthermore, for similar reasons, critically ill patients may have increased susceptibility to DDS conditions. Competing interests =================== None declared. Authors\' contributions ======================= SMB wrote and revised the manuscript. ADP, MH, PJEB, KBL and CJD provided critique of successive drafts of the manuscript. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2369/5/9/prepub> Acknowledgements ================ The authors would like to thank Dr. Bruce Culleton, (Division of Nephrology, Department of Medicine, University of Calgary) for critical review of this manuscript.

PubMed Central

2024-06-05T03:55:47.333599

2004-8-19

PMC515304

Background ========== The S100 protein family consists of at least 21 small, acidic, Ca^2+^-binding proteins with different expression patterns and apparently diverse functional and biological properties \[[@B1]\]. The S100A4 protein has been associated with increased metastatic capacity of cancer cells \[[@B2]-[@B4]\], but how S100A4 exerts the metastasis-promoting effects is still incompletely understood. S100A4 has been shown to interact with a number of cytoskeleton-associated proteins, including non-muscle myosin \[[@B5]\], actin \[[@B6]\] and non-muscle tropomyosin \[[@B7]\], thereby possibly affecting cell motility. In addition, we have previously demonstrated an association between S100A4, matrix metalloproteinases and the invasive capacity of osteosarcoma cells \[[@B8]\]. It has also been reported that S100A4 can be secreted, and suggested extracellular functions include induction of angiogenesis \[[@B9]\], stimulation of cell motility \[[@B10]\], and stimulation of neurite outgrowth \[[@B11]\]. Moreover, we have recently demonstrated nuclear localization of S100A4, and an association between nuclear expression and tumor stage in colorectal cancer \[[@B12]\], indicating other, yet undiscovered, functions of S100A4. Finally, S100A4 has been reported to interact with the tumor suppressor protein p53 and to modulate transcription of downstream target genes, thus influencing p53-mediated apoptosis \[[@B13]\]. Interferon-gamma (IFN-γ) is a pleiotropic cytokine secreted from activated T lymphocytes and natural killer cells. It can activate and suppress a number of target genes, leading to various effects, including inhibition of viral replication, activation of immune cells and induction of cell cycle arrest and apoptosis \[[@B14],[@B15]\]. Apoptosis is a process of programmed cell death characterized by chromatin condensation and DNA fragmentation, plasma membrane blebbing and cell shrinkage. One of the central executioners of the apoptotic pathway are the caspases \[[@B16]\]. It has, however, become evident that a number of caspase-independent pathways exist, leading to programmed cell death involving cathepsins \[[@B17]\], apoptosis inducing factor \[[@B18]\], calpains, AP24 and other serine proteases \[[@B19]\]. Evidence has been provided showing that IFN-γ is able to induce programmed cell death by both caspase-dependent and -independent pathways, e.g. through cathepsin D \[[@B20]\], caspase-1 \[[@B21]\], TNF-related apoptosis-inducing ligand (TRAIL) \[[@B22]\] and Fas/FasL \[[@B23]\]. Searching for cytokines and signal transduction modulators affecting S100A4 expression, we recently discovered that IFN-γ downregulated S100A4 transcription and induced apoptosis in the human osteosarcoma cell line OHS \[[@B24]\]. In the corresponding anti-S100A4 ribozyme-transfected cell line II-11b \[[@B3]\], apoptosis induction was remarkably lower, and these findings prompted us to investigate whether S100A4 could affect IFN-γ-mediated apoptosis. In the present study, we show that S100A4 can be secreted from these human osteosarcoma cells, and that extracellular addition of rS100A4 sensitizes cells to apoptosis induced by IFN-γ. Furthermore, we demonstrate that this cell death pathway is independent of caspases, but possibly dependent upon NF-κB transactivation and generation of reactive oxygen species (ROS). Methods ======= Cell culture and treatment -------------------------- The human osteosarcoma cell line OHS and its anti-S100A4 ribozyme transfected counterpart II-11b have been described previously \[[@B8]\]. Cell lines were cultivated in RPMI-1640 (Bio Whittaker, Verviers, Belgium) containing 10% fetal bovine serum (FBS; Biochrome KG, Berlin, Germany), 1 mM Hepes, and 2 mM Glutamax (GIBCO BRL, Life Technologies, Paisley, UK). For all experiments, subconfluent cultures were trypsinated and seeded at 1.5 × 10^4^cells/cm^2^. After overnight incubation, the culture medium was replaced with medium in presence or absence of IFN-γ and recombinant S100 proteins, and harvested as indicated in the text. Where indicated, protease inhibitors or antioxidants were added to the cell culture simultaneously as IFN-γ. Materials --------- Human recombinant IFN-γ, L-NMMA, zVAD-fmk, zYVAD-fmk, zVDVAD-fmk, zDEVD-fmk, zVEID-fmk, zIETD-fmk, zLEHD-fmk, zFA-fmk and E64 were purchased from Calbiochem (San Diego, CA, USA). N-acetylcysteine was from Sigma Chemical Co (St Louis, MO, USA). Production of proteins ---------------------- Histidine-tagged mouse recombinant S100A4 was cloned into the pQE30 vector (kindly provided by E. Lukanidin, Institute of Cancer Biology, Copenhagen, Denmark), expressed in E. coli, purified on a Ni^2+^-column and dialyzed against phosphate-buffered saline. A satisfactory degree of purity was confirmed by SDS polyacrylamide gel electrophoresis (SDS-PAGE) where a single band was visualized by silver staining. Verification of the identity of the protein was performed using matrix-assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI). Human recombinant S100A10 and S100A13 tagged with maltose-binding protein were kindly provided by C. S. Skjerpen (Dep. of Biochemistry, The Norwegian Radium Hospital, Oslo, Norway). Cell viability -------------- Cell viability was estimated by measuring metabolic activity using CellTiter 96^®^AQ~ueous~One Solution Reagent (MTS) (Promega, Madison, WI, USA) according to the manufacturer\'s manual. The absorbance at 490 nm was recorded using a Wallac 1420 Victor^2^Multilabel counter (Wallac Oy, Turku, Finland). Western blot analysis --------------------- Protein lysates were prepared in 50 mM Tris-HCl (pH 7.5) containing 150 mM NaCl and 0.1% NP-40 with 1 mM PMSF and 2 μg/ml each of pepstatin, aprotinin (Sigma Chemical Co, St Louis, MO, USA) and leupeptin (Roche Diagnostics, Mannheim, Germany). Total protein lysate from each sample was separated on 8--12% SDS-polyacrylamide gel electrophoresis, and transferred onto Immobilon-P membranes (Millipore, Bedford, MA, USA) according to the manufacturer\'s manual. As a loading and transfer control, the membranes were stained with 0.1% amidoblack. After blockage of non-specific binding sites with 10% non-fat dry milk in Tris-buffered saline with 0.25% Tween (TBST), blots were incubated for 1 h at room temperature with rabbit polyclonal anti-poly (ADP-ribose) polymerase (PARP; diluted 1:1000, Roche Diagnostics, Mannheim, Germany), mouse monoclonal anti-lamin B (0.3 μg/ml; Oncogene Research Products, Boston, MA, USA), mouse monoclonal anti-α-tubulin (0.3 μg/ml; Oncogene Research Products, Boston, MA, USA), rabbit polyclonal anti-caspase-1 (1:100; Santa Cruz Biotechnology, Santa Cruz, CA, USA), goat polyclonal anti-caspase-3 (1:2000; R&D Systems, Minneapolis, MN, USA), mouse monoclonal anti-caspase-8 (1:5000; Alexis Biochemicals, Montreal, Canada), mouse monoclonal anti-caspase-9 (1:1000; R&D Systems, Minneapolis, MN, USA), rabbit polyclonal anti-caspase-12 (1:2000; Oncogene Research Products, San Diego, CA, USA), mouse monoclonal anti-NFκB p65 (1:1000, Cell Signaling Technology Inc. Beverly, MA, USA) or mouse monoclonal anti-phospho-IκBα (Ser32/36) (1:2000, Cell Signaling Technology Inc. Beverly, MA, USA). After washing, the blots were incubated for 1 h at room temperature with a horseradish peroxidase-conjugated secondary antibody (DAKO, Glostrup, Denmark) diluted 1:5000. Anti-lamin B, anti-α-tubulin, anti-PARP, anti-caspase-1, anti-caspase-8, anti-NFκB p65 and anti-phospho-IκBα were diluted in TBST with 5% non-fat dry milk, anti-caspase-3 and anti-caspase-12 were diluted in TBS containing 0.05% Tween and 2% non-fat dry milk and anti-caspase-9 was diluted in TBS with 0.05% Tween, 1% non-fat dry milk and 1% BSA. Signals were visualized using the ECL chemiluminescence substrate (Amersham Pharmacia Biotech, Buckinghamshire, UK). DNA fragmentation analysis -------------------------- Cells were resuspended in lysis buffer (10 mM Tris-HCl pH 7.4, 10 mM NaCl, 10 mM EDTA, 0.5 % SDS and 0.5 μg/ml proteinase K) and incubated for 1 hour at 50°C. High molecular weight DNA was precipitated by adding 1 M NaCl and subsequently incubated overnight at 4°C. Samples were centrifuged for 30 min, 2700 g at 4°C, and supernatants collected. 95 % ethanol was added, and DNA precipitated overnight at -20°C. Samples were centrifuged for 10 min, 10 000 rpm at 4°C, and pellets were washed in 70 % ethanol and re-centrifuged. Thereafter, pellets were air-dried and dissolved in 10 mM Tris-HCl (pH 7.0), 15 mM NaCl, 1 mM EDTA and 0.2 mg/ml RNAse A, and incubated for 1 h at room temperature. DNA was separated by electrophoresis in 1.0 % agarose gel containing ethidium bromide and photographed under ultraviolet illumination. Detection of S100A4 in cell culture medium ------------------------------------------ Cell culture medium from OHS and II-11b cells were harvested at 24, 48 and 72 hours, and immediately filtered using a 0.45 μm filter to discard cells in suspension. From each sample, 1.5 ml of medium was subjected to immunoprecipitation using 15 μg of rabbit polyclonal anti-S100A4 (DAKO, Glostrup, Denmark) or 15 μg rabbit polyclonal anti-p300 (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Primary antibody binding was carried out in cell culture medium at 4°C overnight. Thereafter, antibody was precipitated using Protein A Sepharose beads for 1 h at 4°C and eluted in loading buffer. Samples were subjected to Western blot analysis as described above using rabbit polyclonal anti-S100A4 diluted 1:300 in TBST. Transient transfection and plasmid constructs --------------------------------------------- The NF-κB reporter construct containing three NF-κB response elements (i.e., sites identical to the κB site from the Igκ light chain promoter) driving luciferase expression has been described previously \[[@B25]\]. 1.0 × 10^6^II-11b cells were transiently transfected with 10 μg of the NF-κB reporter construct using electroporation (240 V). Thereafter, cells were seeded in 96-well plates at a density of 4.5 × 10^4^cells/cm^2^and incubated in the absence or presence of rS100A4 and zFA-fmk. Cells were harvested 48 h later and assayed for luciferase activity using the Luciferase Assay System (Promega, Madison, WI, USA) according to the manufacturer\'s manual. Statistical analysis -------------------- All statistical analyses were performed using Student\'s t-test. P-values less than 0.05 were considered to be statistically significant. Results ======= OHS cells are more sensitive than II-11b cells to IFN-γ-mediated suppression of cell viability ---------------------------------------------------------------------------------------------- We have previously observed that the human osteosarcoma cell line OHS is more sensitive to IFN-γ induced apoptosis than the S100A4-ribozyme transfected counterpart II-11b \[[@B24]\]. To examine the effects of IFN-γ on cell viability in more detail, time- and dose-response experiments were performed (Fig. [1A](#F1){ref-type="fig"} and [1B](#F1){ref-type="fig"}). In OHS cells, addition of 1 u/ml IFN-γ had no effect on cell viability, while an increasing amount of cell death was observed with higher concentrations. On the other hand, II-11b cells only showed a marked reduction in cell number after treatment with 1000 u/ml IFN-γ for 72 hours. IFN-γ induces apoptosis in OHS Cells ------------------------------------ Previously, we have shown that treatment of OHS cells with IFN-γ induced a significant increase in the fraction of apoptotic cells, while the cell cycle distribution remained unchanged \[[@B24]\]. Thus, it seemed plausible that the reduced cell viability could be due to induction of apoptosis, and to confirm this hypothesis, we investigated some of the known hallmarks of the apoptotic process. As seen in Fig. [2A](#F2){ref-type="fig"}, IFN-γ induced DNA fragmentation in OHS cells. Additionally, treatment with IFN-γ induced cleavage of poly (ADP-ribose) polymerase (PARP) and lamin B (Fig. [2B](#F2){ref-type="fig"}). Taken together, these results clearly demonstrated that IFN-γ-treatment induces apoptosis in OHS cells. Detection of S100A4 in cell culture medium ------------------------------------------ When performing the OHS cell culture experiments, we observed that, when keeping the concentration of IFN-γ constant, decreasing the volume of cell culture medium or increasing the number of cells resulted in increased cell death (data not shown). This observation suggested that the cells might secrete a soluble factor with impact on the level of apoptosis induced by IFN-γ. Since the II-11b cells are derived from the OHS cells by transfection with an S100A4-specific ribozyme, this finding prompted us to investigate whether S100A4 is secreted from these cells, and whether extracellular S100A4 could cooperate with IFN-γ in induction of apoptosis. Fig. [3](#F3){ref-type="fig"} shows an immunoprecipitation of cell culture medium from OHS and II-11b cells grown for 24, 48 and 72 hours. This clearly demonstrated that S100A4 is present in cell culture medium from OHS cells, and to a much lesser extent from II-11b cells. A small amount of S100A4 was also detected in the cell culture medium control. This possibly reflects bovine S100A4 present in fetal calf serum added to the medium, but the exact nature of this signal has not been further investigated. To exclude cell lysis as the cause of extracellular S100A4, lactate dehydrogenase activity was measured in the conditioned culture medium. No significant increase in enzyme activity was detected, supporting the hypothesis that S100A4 is actively secreted (data not shown). Importantly, cell culture medium from OHS cells cultured for 72 hours immunoprecipitated with a rabbit polyclonal anti-p300 antibody did not give any S100A4-signal, indicating that the detection of S100A4 was not due to unspecific binding. Sensitization of IFN-γ-mediated apoptosis by extracellular S100A4 ----------------------------------------------------------------- In order to investigate whether extracellular S100A4 could increase apoptosis induced by IFN-γ, recombinant S100A4 (rS100A4) was added to the cell culture medium of OHS and II-11b cells in addition to IFN-γ. Addition of rS100A4 alone had no effect on cell viability at concentrations ranging from 400 pg/ml to 20 μg/ml (Fig. [4](#F4){ref-type="fig"} and data not shown). On the other hand, when 20 μg/ml of rS100A4 was added together with 100 u/ml IFN-γ to II-11b cells, a significant decrease in cell viability was observed (Fig. [4B](#F4){ref-type="fig"}). In fact, when II-11b cells were incubated with 20 μg/ml rS100A4, the level of apoptosis induced by IFN-γ was similar as in OHS cells. Addition of rS100A4 to OHS cells also increased the IFN-γ-mediated cell death, thus seemingly adding to the effects of the endogenously produced extracellular S100A4 (Fig [4A](#F4){ref-type="fig"}). Accordingly, a significant decrease in cell viability was seen at lower concentrations of added rS100A4 in OHS cells than in II-11b cells. Furthermore, we made attempts to neutralize the secreted S100A4 protein by adding anti-S100A4 antibodies to the cell culture medium of IFN-γ treated OHS cells, but were unable to detect any decrease in the amount of cell death (data not shown). Possibly, the antibodies used (DAKO, Glostup, Denmark, as well as three in-house produced monoclonal antibodies \[[@B26]\]) were not able to block the activity of extracellular S100A4. S100 proteins share a high degree of sequence homology; therefore it was of interest to investigate whether other S100 proteins were able to sensitize tumor cells to IFN-γ-mediated cell death. Addition of 20 μg/ml recombinant S100A10 or S100A13 had no effect on apoptosis induction by IFN-γ in II-11b cells. Thus, it seems unlikely that the observed effects are a general feature of S100 proteins, but whether other S100 proteins than those tested here could possess IFN-γ-sensitizing properties requires further investigation. The possibility exists that rS100A4 and IFN-γ induce a different type of cell death in II-11b cells than the IFN-γ-mediated apoptosis in OHS cells. DNA fragmentation (Fig. [5A](#F5){ref-type="fig"}) as well as PARP and Lamin B cleavage (Fig. [5A](#F5){ref-type="fig"}) were demonstrated also in II-11b cells, suggesting that addition of recombinant S100A4 actually mimics the biological effects of endogenous S100A4 in these experiments. IFN-γ-mediated apoptosis in OHS dells is caspase-independent ------------------------------------------------------------ Caspase-mediated apoptosis is the most important program of cell death. In order to investigate whether IFN-γ activated the caspase cascade in our cell system, we have followed two different approaches: detection of activated caspases by Western blotting and treatment of cell cultures with caspase inhibitors. Using specific antibodies detecting both the proenzyme and the active form of caspase-1, -3, -8, -9 and -12, we were not able to detect any activation of caspase-1, -3, -8 or -12 (Fig. [6](#F6){ref-type="fig"} and data not shown). However, the caspase-9 antibody detected a weak, but distinct, \~37 kDa band in IFN-γ treated cells at 48 and 72 h and also in control cells at 72 h. This probably corresponds to one of the active forms of caspase-9. Additionally, we observed a lower level of expression of the proenzyme of caspase-3 in IFN-γ treated cells at 48 and 72 h (Fig. [6](#F6){ref-type="fig"}). This could possibly indicate a suppression of caspase-3 expression by IFN-γ or activation of caspase-3 with resulting cleavage of the proenzyme. On the other hand, no active forms of caspase-3 were detected, even when exposing the film for several hours. To ensure that the antibodies were able to detect the mature forms, immunotoxin-treated breast cancer cells were included as a positive control \[[@B27]\]. Furthermore, the pan-caspase inhibitor zVAD-fmk inhibited IFN-γ-induced cell death, albeit only at high concentrations (Fig. [7](#F7){ref-type="fig"}). zVAD-fmk suppressed IFN-γ-mediated apoptosis significantly at 125 μM, whereas 50 μM inhibited cell death non-significantly, and 10 μM zVAD-fmk had no effect on apoptosis inhibition. The latter concentration has been reported to be sufficient to inhibit most caspases \[[@B28]\], therefore it seems likely that the observed effect of zVAD-fmk could be due to inhibition of other proteases or signaling events than caspases. In addition, inhibition of caspase-1 (zYVAD-fmk), caspase-2 (zVDVAD-fmk), caspase-3 and -7 (zDEVD-fmk), caspase-6 (zVEID-fmk), caspase-8 (zIETD-fmk) and caspase-9 (zLEHD-fmk) had no effect on IFN-γ-induced cell death at concentrations up to 50 μM. Taken together, we have not observed any activation of caspases-1, -3, -8 or -12, whereas a low-level activation of caspase-9 was detected. Inhibitors against a range of caspases including caspase-9 did not suppress induction of cell death; hence, from these experiments we concluded that the cell death program induced by IFN-γ in OHS cells most likely is caspase-independent. Reactive oxygen species mediate IFN-γ-induced apoptosis ------------------------------------------------------- The caspase inhibitors used in the experiments above were solved in dimethyl sulfoxide (DMSO), and therefore DMSO was included as a negative control. Surprisingly, a significant inhibition of IFN-γ-induced apoptosis was observed using 0.25 % DMSO (Fig. [7](#F7){ref-type="fig"}). One of the known biological properties of DMSO is scavenging of hydroxyl radicals, indicating that the generation of reactive oxygen species (ROS) could be implicated in the observed induction of apoptosis. Furthermore, treatment with the commonly used antioxidant N-acetylcysteine significantly suppressed the IFN-γ-induced cell death (Fig. [7](#F7){ref-type="fig"}), whereas the nitric oxide synthase inhibitor L-NMMA had no effect (data not shown). Notably, zVAD-fmk was recently shown to possess potent antioxidant effects at high concentrations \[[@B29]\] possibly explaining its inhibitory effects shown above. IFN-γ-mediated apoptosis in OHS cells is independent of cathepsin B, but possibly mediated by NF-κB --------------------------------------------------------------------------------------------------- The pan-caspase inhibitor zVAD-fmk has previously been shown to bind the cysteine protease cathepsin B and to suppress cathepsin B activity *in vitro*\[[@B30]\]. Since cathepsins are able to act as mediators of programmed cell death \[[@B17]\], zVAD-fmk could possibly inhibit IFN-γ-induced apoptosis through inhibition of cathepsin B. However, 50--250 μM E64 (a cysteine protease inhibitor) or 1--10 μM zFA-fmk (a cathepsin B inhibitor) failed to suppress induction of the observed cell death (data not shown), arguing against cathepsin B as a mediator of IFN-γ-induced apoptosis. On the other hand, high concentrations (125 μM) of zFA-fmk completely inhibited induction of apoptosis (Fig. [7](#F7){ref-type="fig"}). Cathepsin B activity is most likely suppressed at concentrations below 10 μM \[[@B28]\]; thus, we believe that the observed inhibition by zFA-fmk is due to non-specific effects. Indeed, zFA-fmk was recently shown to suppress transactivation by NF-κB at concentrations similar to those used in these experiments \[[@B30]\]. Interestingly, addition of rS100A4 to II-11b cells lead to a marked induction of NF-κB activity in transient transfection assays using an NF-κB reporter construct, and this induction was blocked by zFA-fmk (Fig. [8A](#F8){ref-type="fig"}). Furthermore, IκBα was phosphorylated upon addition of extracellular S100A4 (Fig. [8B](#F8){ref-type="fig"}), confirming activation of the NF-κB pathway. Taken together, these results demonstrate that extracellular S100A4 activates NF-κB, and suggest that the apoptosis induced by S100A4 and IFN-γ could be mediated through an NF-κB-dependent pathway, but whether these two events are actually causally connected remains to be investigated. Discussion ========== IFN-γ induces apoptosis in a variety of different cell types, but some cells are less susceptible or even resistant to IFN-γ-mediated apoptosis. The reasons for this variation have not been clearly elucidated, and given the biological importance of IFN-γ, a better understanding of the mechanisms of IFN-γ-induced cell death is warranted. In the present study, we have demonstrated that the Ca^2+^-binding protein S100A4 sensitizes human osteosarcoma cells to IFN-γ-induced apoptosis. Furthermore, we have shown that the IFN-γ-mediated cell death is independent of caspases, but possibly mediated by reactive oxygen species. Finally, we have demonstrated that extracellular S100A4 can activate NF-κB through increased phosphorylation of IκBα, but whether this NF-κB transactivation is required for apoptosis induction remains unknown. Several of the S100 proteins are released into the extracellular space, mostly by unknown mechanisms \[[@B1]\]. S100A4 secretion by human osteosarcoma cells is in agreement with previous reports showing that S100A4 can be released from non-malignant as well as tumor cells \[[@B9],[@B31],[@B32]\]. The apoptosis-inducing properties of extracellular S100A4 are, however, novel findings, emphasizing the multiple functions of this protein. Other extracellular S100 proteins also exert multiple effects, including proinflammatory activity, chemotaxis, neurite extension, stimulation of angiogenesis, and regulation of cell survival \[[@B1]\]. Proapoptotic activity has, to our knowledge, only been demonstrated previously for S100A1, S100B \[[@B33]\] and S100A8/S100A9 \[[@B34]\]. S100B induces apoptosis by interaction with the multi-ligand cell surface receptor RAGE (a member of the immunoglobulin superfamily) \[[@B33]\]. RAGE also binds S100A12 \[[@B35]\], whereas no cell surface receptor for S100A4 has been identified so far. In view of the significant homology between S100 proteins, S100A4 could possibly bind to and activate RAGE. However, treatment with anti-RAGE IgG \[[@B35]\] up to 50 μg/ml had no inhibitory effect on apoptosis induction (data not shown), and higher concentrations were toxic to the cells. Hence, another cell surface receptor might mediate S100A4-induced sensitization of apoptosis, and in line with this, RAGE-independent actions were recently demonstrated for S100B \[[@B36]\]. Notably, mechanisms distinct from ligand-receptor interactions can also propagate proapoptotic signals. S100A8/S100A9 induced apoptosis even when a dialysis membrane hindered contact with the cells, possibly by binding Zn^2+^or other divalent cations \[[@B34]\]. Addition of 20 μg/ml rS100A4 (1.5 μM) to the cell medium of II-11b cells was required to induce apoptosis upon IFN-γ-treatment. Comparably, micromolar amounts of S100B and S100A8/A9 were needed to induce cell death \[[@B33],[@B34]\]. Also S100A4-mediated neurite extension \[[@B11]\] as well as stimulation of angiogenesis \[[@B9]\] were shown to require micromolar concentrations of recombinant protein. Whether native extracellular S100A4 is present at these concentrations in our experiments is unknown, as secreted S100A4 was not quantitatively estimated. In many cases, stimulation of tumor cells with a single proapoptotic signal is not sufficient for induction of programmed cell death. On the other hand, a number of cytokines and other proapoptotic signal substances can increase the cells\' sensitivity to other apoptosis-inducing agents, thus initiating the cell death program. IFN-γ induces apoptosis in synergism with a wide variety of such agents, and particularly members of the tumor necrosis factor (TNF) superfamily have been studied, e.g., TNF-α \[[@B37]\], FasL \[[@B38]\], LIGHT \[[@B39]\] and TRAIL \[[@B22]\]. Our work adds a novel proapoptotic signaling molecule to the list of agents acting in synergism with IFN-γ to induce programmed cell death. The mechanisms by which S100A4 sensitizes the tumor cells to apoptosis induction are not completely elucidated. Recently, Grigorian *et al*demonstrated that S100A4 interacts with the tumor suppressor protein p53 to enhance p53-dependent apoptosis. The proposed mechanism was based on modulation of p53-dependent transactivation of target genes \[[@B13]\]. The osteosarcoma cells used in our experiments harbor, however, a mutated p53 protein \[[@B40]\], thus, S100A4 seemingly enhance apoptosis through at least two distinct mechanisms. One possible explanation for the apoptosis sensitizing effects of S100A4 is that the protein positively modulates IFN-γ-signaling in general. By studying the regulation of S100A4 expression, we have demonstrated that extracellular S100A4 enhances IFN-γ-mediated suppression of S100A4 transcription (Pedersen *et al.*, in preparation). These findings, in addition to the present work, suggest that secreted S100A4 could sensitize IFN-γ-signaling, and this possibility is currently under investigation. Previously identified sensitizers of IFN-γ-signaling include type I interferons (IFN-α/-β) \[[@B41]\] and IFN-γ itself \[[@B42]\]. Hu *et al*. discusses, based on their experiments using peripheral blood mononuclear cells, that other factors probably also exist to sensitize IFN-γ-signaling. Interestingly, S100A4 is strongly expressed by T lymphocytes , but whether lymphocytes also secrete S100A4 remains to be investigated. Our results suggest that S100A4-mediated NF-κB activation might be required for triggering the apoptotic process induced by IFN-γ. To support this hypothesis, we made attempts to inhibit NF-κB signaling by treatment with SN50, a peptide that contains the p50 NLS and selectively blocks nuclear translocation of NF-κB, but this lead to an increased fraction of apoptotic cells. Schotte *et al.*suggest that zFA-fmk has a promoter-specific NF-κB inhibitory effect \[[@B30]\], whereas SN50 potently blocks all NF-κB activity, possibly explaining the divergent results. In general, activation of NF-κB is regarded as an anti-apoptotic event and as such, a total suppression of NF-κB activity could be deleterious to cells. However, NF-κB apparently has a dual role in apoptosis regulation, and cell death induction by NF-κB has been reported in a variety of cell systems \[[@B43]\]. IFN-γ-induced apoptosis in OHS cells was shown to be caspase-independent by two different approaches: (i) no active forms of a number of caspases were detected; and (ii) caspase inhibitors were not able to suppress cell death triggered by IFN-γ. Even though the caspase cascade was not activated, well-known hallmarks of apoptosis were identified. DNA fragmentation and Lamin B cleavage have previously been shown to occur independent of caspase activation \[[@B44],[@B45]\], whereas PARP cleavage has been regarded as a caspase-indispensable event. Cleavage of PARP to an 89-kDa fragment was however evident, but the observed low level activation of caspase-9 was, due to the lack of effect when adding the specific caspase-9 inhibitor, not considered crucial, possibly implicating other proteases than caspases in PARP inactivation. Finally, experiments using the antioxidant NAC indicated that ROS are involved in S100A4/IFN-γ-mediated triggering of apoptosis. In light of recent reports demonstrating that both S100B and S100A8/S100A9 induced apoptosis in an oxidant-dependent manner this result is highly interesting \[[@B33],[@B46]\]. Conclusions =========== In conclusion, we have demonstrated that extracellular S100A4 induces NF-κB transactivation and possesses proapoptotic activity through sensitization of IFN-γ-induced apoptosis, but whether activation of NF-κB is required for the observed sensitization of cell death is still unknown. Even though the precise physiological role of the apoptosis sensitizing effect of S100A4 awaits further studies, it highlights yet another area of research in the attempts to elucidate the complex biological roles of S100A4. Competing interests =================== None declared. Authors\' contributions ======================= KBP carried out the cell culture experiments, immunoprecipitations and western blots used in the detection of S100A4 in cell culture medium. In addition he did the dose-response experiments in II-11b cells, all the experiments using caspase inhibitors and NAC, western blot analysis of caspase activity, and drafted the manuscript. KA carried out dose-response experiments in OHS cells, and in addition all the cell culture experiments used in subsequent western blot analysis of caspase- and NF-κB activity and DNA fragmentation analysis. In addition she did transient transfections and all cell culture experiments using other S100 proteins and anti-RAGE antibodies. ØF participated in the design of the study. GMM conceived the study, and participated in its design and coordination. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2407/4/52/prepub> Acknowledgements ================ We gratefully acknowledge the generous gift of the pQE30-S100A4 expression construct from Dr. E. Lukanidin (Institute of Cancer Biology, Copenhagen, Denmark), recombinant S100A10 and S100A13 from C. S. Skjerpen (The Norwegian Radium Hospital, Oslo, Norway) and anti-RAGE IgG from Dr. A. M. Schmidt (Columbia University, New York, USA). The authors would also like to thank Dr. Dave Warren for production of recombinant S100A4, Ida Grotterød for excellent technical assistance and Dr. Yvonne Andersson for helpful discussions. The Norwegian Cancer Society, The Research Council of Norway and Stiftelsen Sophies Minde financially supported the study. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **OHS cells are more sensitive than II-11b cells to IFN-γ-mediated suppression of cell viability.**Cell viability was measured in OHS cells (*A*) and II-11b cells (*B*) after treatment with the indicated concentrations of IFN-γ for 24, 48 and 72 hours. In each experiment, untreated control cells were included at all time points, and all values are given as a percentage of corresponding untreated control cells. The results are presented as mean values ± S.D. of at least three independent experiments performed in triplicate. \*, p \< 0.0001 and \*\*, p \< 0.05 as compared to untreated control cells. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **DNA fragmentation, PARP and Lamin B cleavage in IFN-γ-treated OHS cells.***A*, Ethidium bromide stained agarose gel of DNA isolated from OHS cells grown with or without 100 u/ml IFN-γ for 72 hours. *B*, Western blot analysis of the expression of PARP and Lamin B in total cell lysates. OHS cells were left untreated (-) or treated with 100 u/ml IFN-γ (+) for 24, 48 and 72 hours, and lysates subjected to Western blot analysis as described in the \"Materials and Methods\" section. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **OHS cells secrete S100A4.**Immunoprecipitation of cell culture medium alone (control) or cell culture medium from II-11b and OHS cells cultured for 24, 48 and 72 hours as indicated. 1.5 ml of medium was immunoprecipitated using rabbit polyclonal anti-S100A4 or anti-p300, and subsequently immunoblotted with anti-S100A4. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Addition of recombinant S100A4 sensitizes OHS and II-11b cells to IFN-γ-induced apoptosis.**Cell viability was measured in OHS and II-11b cells after addition of the indicated concentrations of recombinant S100A4 to the cell culture medium with or without 100 u/ml IFN-γ. All values are given as a percentage of viable cells relative to untreated control cells. The results are presented as mean values ± S.D. of at least three independent experiments performed in triplicate. \*, p \< 0.0001 and \*\*, p \< 0.05 as compared to untreated control cells. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **DNA fragmentation, PARP and Lamin B cleavage in rS100A4/IFN-γ-treated II-11b cells.***A*, Ethidium bromide stained agarose gel of DNA isolated from II-11b cells. Cells were stimulated or left untreated with 100 u/ml IFN-γ for 72 hours in the presence or absence of 20 μg/ml rS100A4 as indicated. *B*, Western blot analysis of the expression of PARP and Lamin B in total cell lysates. II-11b cells were left untreated or treated with 100 u/ml IFN-γ and/or 20 μg/ml rS100A4 for 24, 48 and 72 hours as indicated, and lysates subjected to Western blot analysis as described in \"Materials and Methods\". :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Western blot analysis of caspase-3 and caspase-9 in total cell lysates.**OHS cells were left untreated (-) or treated with 100 u/ml IFN-γ (+) for 24, 48 and 72 hours, and lysates subjected to Western blot analysis as described in \"Materials and Methods\". As a positive control for caspase activation, an immunotoxin-treated breast cancer cell line was included. α-tubulin serves as a loading control. The results shown are representative of three independent experiments. :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Inhibition of cell death by zVAD-fmk, zFA-fmk, DMSO and NAC.**OHS cells were cultivated in the presence or absence of 100 u/ml IFN-γ and the indicated inhibitors for 72 hours, and cell viability was measured as described in \"Materials and Methods\". All values are given as a percentage of viable cells relative to corresponding cells treated with the indicated inhibitor, but without IFN-γ-treatment. Data are mean values ± S.D. The results represent at least three independent experiments performed in triplicate. \*, p \< 0.001 and \*\*, p \< 0.05 as compared to IFN-γ-treated cells without inhibitor. :::  ::: ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **Recombinant S100A4 induces NF-κB transactivation.***A*, II-11b cells were transiently transfected with an NF-κB reporter construct and incubated with or without 4 μg/ml rS100A4 and 50 μM zFA-fmk for 48 hours. Luciferase activity was expressed as fold induction of activity compared to the corresponding untreated control. Data are mean values ± S.D. The results represent at least three independent experiments performed in duplicate. \*, p \< 0.001 and \*\*, p \< 0.01 as compared to untreated control. \*\*\*, p \< 0.005 as compared to rS100A4-treatment without zFA-fmk. *B*, Western blot analysis of NF-κB p65 and phospho-IκBα (Ser32/36) in total cell lysates from II-11b cells, untreated or treated with 100 u/ml IFN-γ and/or 20 μg/ml rS100A4 for 24 and 48 hours as indicated. :::  :::

PubMed Central

2024-06-05T03:55:47.335770

2004-8-19

PMC515305

Background ========== Glioblastoma Multiforme (GBM) are the most common and aggressive tumors of the central nervous system (CNS) \[[@B1]\]. Current treatments (e.g. chemo-and radiotherapy) have been relatively unsuccessful and no significant improvement in the prognosis has been recorded over the last 20 years \[[@B1]\]. Cancer cells are capable of inducing a specific immune response against the tumor and this property has been used in an attempt to design new therapeutic strategies \[[@B2],[@B3]\]. In particular, anti-tumor vaccination strategies using known tumor-associated antigens or whole tumor extracts have been developed over the last few years \[[@B2],[@B3]\]. However, a growing body of evidence has recently shown that tumors are capable of suppressing anti-cancer immune responses by the induction of tolerance, anergy or by selective killing of immune cells, thereby preventing their destruction by the immune system \[[@B4],[@B5]\]. Dead cells have been also widely used as a source of tumor antigen but contradictory results have been reported on their effect on tumor growth \[[@B6]\]. We \[[@B7]\] and others \[[@B8],[@B9]\] have shown that apoptotic bodies are capable of inducing a long-lasting and efficient immune response against tumors whereas others have suggested that necrotic cells could be anergic and tolerogenic \[[@B10],[@B11]\]. Thus, the ability of dead cells to generate an immune response against a tumor could be associated with the nature of the death inducer used and/or the modus operandi of cell death (i.e. necrosis vs. apoptosis). Cytotoxic T lymphocytes (CTLs) and natural killer cells (NK), the major actors of the immune surveillance, have the ability to induce cell death by apoptosis mainly through two mechanisms: the death receptor pathway (i.e. CD95/Fas/APO-1, TRAIL) or the cytotoxic granules (i.e. perforin/granzyme pathway) \[[@B4],[@B12]\]. Activation of death receptors appears to be sufficient to induce the cytosolic activation of caspases, the main proteolytic enzymes of apoptosis, in some tumors (class I) while, in class II tumors, amplification of the death signal occurs through mitochondrial activation of caspases \[[@B13]\]. Proteins of the BCL-2 family play a major role in the control of apoptosis both *in vitro*and *in vivo*in the latter pathway \[[@B14]\]. These proteins can be divided into anti-apoptotic proteins such as Bcl-2 and pro-apoptotic proteins such as Bax \[[@B14]\]. Inhibition of apoptosis through the overexpression of Bcl-2 promotes oncogenesis as demonstrated in some follicular B-lymphomas \[[@B15]\] while, on the other hand, the loss of Bax function has been associated with tumor progression and a bad prognosis in colon and gastric tumors having a microsatellite mutator phenotype \[[@B16]\]. We have recently observed that the expression of a gain of function variant of Bax can be associated with a longer survival in GBM patients \[[@B17]\]. Note that the expression of the anti-apoptotic molecule Bcl-2 and that of the pro-apoptotic Bax increased in parallel in low grade to high grade tumors of glial origin, suggesting that Bax and Bcl-2 could play an antagonistic but essential role in these tumors \[[@B18]\]. One of the most powerful mechanisms of control of tumor growth is exercised by the immune system. The immune surveillance hypothesis suggests that potentially dangerous cells could be eliminated through induction of cell death \[[@B3]\]. Although the CNS is usually considered an immune privileged site \[[@B5]\], specific cellular immune responses against tumoral antigens have been achieved in some animal models \[[@B19],[@B20]\]. In GBM, the absence of Bax protein is compensated by an increased expression of Bak, another multidomain pro-apoptotic protein, which also maintains the immune-induced cell death \[[@B21]\]. Thus, manipulation of the expression of the BCL-2 family members could also be involved in the sensitivity of glial tumors to the immune system. We have tested this hypothesis by establishing cell lines, which stably express transgenes encoding either for human *Bax*or *Bcl-2*in a rat glioma model and analyzed the effects of these transgenes on the *in vitro*and *in vivo*growth of these cell lines. Methods ======= Reagents -------- Unless specified, all reagents used in this study were from Sigma (St Quentin-Fallavier, France). Monoclonal anti-human Bax antibody (clone 4F11) was from Immunotech (Villepinte, France) and monoclonal anti-human Bcl-2 antibody (M 0887) was from Dako (Trappes, France); antibodies against rat Bcl-2 or Bax were respectively from Oncogene (Ab5) (Fontenay sous Bois, France) and from Pharmingen (13456E) (Le Pont de Claix, France). The fluorogenic peptide Ac-DEVD-AMC was from Bachem (Voisins les Bretonneux, France) and the lactate dehydrogenase (LDH) activity was measured using the Cytotox 96^®^assay from Promega (Charbonnières, France) as described previously \[[@B17],[@B18],[@B21]\]. Experimental research on animals have been conducted according to recommendations of the French National Ethics committee, and are in compliance with the [Helsinki]{.underline}[Declaration]{.underline}. In vitro transfection, proliferation and induction of apoptosis --------------------------------------------------------------- The rat glioma cell line, A15A5, was obtained from the European Collection of Animal Cell Culture (Valbonne, France). The cell line was maintained in RPMI-1640 (Invitrogen, Cergy-Pontoise, France) supplemented with 10% heat-inactivated FCS (Eurobio, Les Ulis, France), 100 μg/ml streptomycin, 100 U/ml penicillin and 2 mM L-glutamate in a 5% CO~2~air-humidified atmosphere at 37°C. Plasmids encoding for human Bcl-2 or Bax were subcloned into pRcCMV (Invitrogen) as described by the manufacturer. A15A5 cells were transfected with either 2 μg pCMV vector, pCMV Bcl-2 or pCMV Bax. Plasmid DNA was introduced into 10^6^cells by electroporation (GenePulser, BioRad, Yvry sur Seine, France) using 200 V/cm and 250 μF. Transfected cells were selected and cloned in a medium containing neomycin (250 μg/ml) for several weeks before clonal dilution. Apoptosis was induced by a short UV-treatment. Both untreated and UV-treated cells were cultured for a further 24 h under serum-free conditions. Cell death was also induced with FasL (0.5 μg/ml; a gift of Dr P. Saas EPI 119, Besançon, France), doxorubicin (doxo; 2 μM), staurosporine (STS; 1 μM), Na-Butyrate (NaB; 10 mM) or serum deprivation (d-serum) for 3 days. A MTT assay was used according to the manufacturer\'s instructions (Promega) to determine *in vitro*cell proliferation of transfected A15A5 cells. Tumor and cell extracts and Western blots ----------------------------------------- A15A5 transfected cells (10^5^cells) or tumors established in rats or mice were homogenized vol./vol. in RIPA buffer (PBS containing 1% NP-40, 0.5% Na-deoxycholate, 0.1% SDS, 10 nM PMSF, 10 nM aprotinin, 1 nM Na-orthovanadate). After several passages in a 2 ml glass Dounce homogenizer, the homogenates were centrifuged at 4°C at 13,000 g for 30 min. The resulting supernatants were assayed for protein concentration using the Bradford technique prior to analysis on 15% SDS-PAGE. Western blots were performed as described earlier \[[@B21]\], using primary antibodies anti-Bcl-2 (1 μg/ml), anti-Bax (2 μg/ml) and actin (0.5 μg/ml). The antibodies bound to Immobilon-P (Millipore, France) were detected by enhanced chemiluminescence (Amersham, Aylesbury, UK) using a second peroxydase-labelled antibody. The amount of immunoreactive protein was quantified using IP-Lab Gel Program (Signal Analytics, Vienna, USA) after scanning with an Imager (Q-Biogene, Strasbourg, France). Animal experiments ------------------ Inbred BDIX rats and Swiss *nude*mice were purchased from Iffa-Credo (L\'Abresle, France) and were housed under standard conditions in our laboratory. huBax and huBcl-2 as well as pCMV A15A5 cells were injected into the brain of BDIX rats weighing between 200 and 240 g. All animal procedures were performed with approved protocols and in accordance with published recommendations for the proper use and care of laboratory animals. Briefly the rats were anaesthetised with an intraperitoneal injection of pentobarbital (50 mg/kg) and positioned in a stereotactic head frame. Aseptic surgical techniques were used to open the scalp in the midline and to expose the frontal and temporalis bones. A 1.0 mm aperture for implanting tumor cells was drilled through the skull. The stereotaxic position of this injection site was 2.5 mm anterior to the bregma and 2.0 mm to the right of midline. 10^4^tumor cells were implanted stereotactically at a depth of 3.0 mm into the cerebral parenchyma using a 10 μl Hamilton syringe with a 26-gauge needle. The volume injected was 5 μl and the hole was sealed with sterile bone wax. Alternatively, rats were injected *sc*with 10^5^cells into the hindlimbs and tumor growth was monitored every week by measuring the volume of the growing tumors. Swiss *nude*mice were treated similarly except that 10^4^cells were injected subcutaneously. In order to evaluate the implication of huBax A15A5 cells preventive anti-tumoral treatment, we designed a protocol consisting of three *sc*injections of 3.3 × 10^4^huBax A15A5 cells 15, 10 and 5 days before injection of pCMV or huBcl-2 A15A5 cells. As a control PBS, A15A5 cell oncolysate, or apoptotic bodies derived from Na-Butyrate (NaB)-treated A15A5 cells as previously described \[[@B7]\]. Briefly, apoptosis was induced *in vitro*by a 10 mM NaB treatment in subconfluent A15A5 cultures. When signs of apoptosis were observed under a microscope (changes in cell morphology, detachment from dishes, chromatin condensation as viewed with Hoechst 33342) apoptotic bodies were collected, centrifuged and conserved at -80°C prior to use. Typically, 250 μg apoptotic bodies were mixed with 5 mg/ml BCG and injected subcutaneously three times over 15 days. As a control we used oncolysates obtained after several cycles of rapid freezing/ thawing of A15A5 cells and the lysates was injected together with 5 mg/ml BCG to the animals as above. On day 0, 5 days post-treatment, four groups of rats were *sc*challenged with 10^5^pCMV or huBcl-2 A15A5 cells. Tumor growth was evaluated daily over 60 days. Immunohistochemical analysis of tumor cell injection site --------------------------------------------------------- Immunochemical analysis was performed as on 12 μm brain frozen sections. Briefly, sections were fixed with 4% paraformaldehyde in PBS for 30 min at room temperature. Endogenous peroxydase activity was inhibited by a treatment with 0,3% H~2~O~2~in methanol for 20 min. The sections were incubated overnight at 4°C with anti-rat CD8 (hybridoma supernatant, Ox 8) diluted 1 in 2 in 1% BSA in PBS, then a secondary antibody coupled to peroxydase was added. The staining was revealed with an AEC substrate. Flow cytometry determination of intratumoral immune population -------------------------------------------------------------- Tumors were resected and minced into 1--2 mm^3^pieces, which were incubated in extraction buffer (30 U/ml hyaluronidase; 500 U/ml DNase, 0.01% w/v collagenase in PBS) at room temperature for 45 min and under constant agitation. The cell suspension was filtered through a sterile grid and washed three times with RPMI and maintained overnight in RPMI before analysis. For the phenotypic analysis, monoclonal antibodies obtained from Pharmingen raised against the following molecules were used: CD3 (556970; cl. 1F4), CD4 (554835, cl ox35), CD8 (554854, cl. Ox 8), CD161 (555006, cl. 10/78), CMH I (22301 D), OX 41/ CD172 (552297) and OX 62 (555010). Monoclonal antibodies raised against granzyme B (GrB, cl. 2C5/F5) was obtained from Chemicon (France) and Fas (AF 126) from R&D Systems (Lille, France). Cells were incubated with the primary antibodies for 30 min at 4°C and washed twice in PBS + 0.1% BSA. For the intracellular detection of GrB, cells were first fixed in 4% paraformaldehyde for 10 min at room temperature, washed with PBS + 0.1% BSA, then permeabilized with 0.1% saponin then incubated with the anti-GrB antibody for 30 min. The secondary antibody was then added for 30 min at 4°C and the cells washed 3 times with PBS + 0.1% BSA. Cells were analyzed on FACScalibur (Becton Dickinson, Le Pont de Claix France) using Cell Quest Pro software. A total of 5000 cells were counted in each experiment. Results ======= Charaterization of rat glioma cells transfected with human Bax or Bcl-2 transgenes ---------------------------------------------------------------------------------- A15A5 cells transfected with the pCMV vector, human *Bax*or *Bcl-2*transgenes, were obtained as described in materials and methods. Two different clones were used in each experiment. The expression of the transgenes was monitored by immunoblot analysis using antibodies specific for human Bcl-2 or Bax and, as shown in figure [1A](#F1){ref-type="fig"}, transfections of the rat glioma cell line with human transgenes were efficiently achieved. Note that the overexpression of Bax did not induce apoptosis in the A15A5 cells, suggesting that the different clones selected expressed sublethal amounts of Bax. To examine the effect of the different transgene expression to cell death, we studied their sensitivity toward different inducers of apoptosis using both drugs such as doxo (20 ng/ml) and STS (20 μM) or treatments such as d-serum, NaB (10 mM) or a short (1 min) UV irradiation. Cell death was monitored and quantified by measuring the activity of caspase 3 (namely the cleavage of the peptide Ac-DEVD-AMC) and the activity of the cytosolic enzyme LDH released into the culture medium (see materials and methods). As shown in figure [1B](#F1){ref-type="fig"}, when compared to pCMV A15A5 cells, huBax A15A5 cells were more sensitive to cell death in all cases. Conversely and as expected, huBcl-2 A15A5 cells were more resistant to all death inducers. Note that the resistance of pCMV A15A5 cells to apoptosis was already high, suggesting that these cells were naturally resistant to apoptosis. Nevertheless, these results suggest that human Bax and Bcl-2 transgenes were functional in the rat glioma A15A5 cells. The immune system exerts its anti-tumoral surveillance mainly through cell death induced by CTLs and NK \[[@B12]\]. These cells use different effectors to mediate apoptosis in target cells: the death receptor mechanism such as the FasL/Fas receptor system or the perforin/GrB cytotoxic pathway \[[@B12]\]. The ligation of the death ligands to their receptors initiates cell death by the activation of the intracellular initiator caspase 8, which in turn can induce apoptosis either through the direct activation of caspase 3 in type I cells or by using mitochondria as an obligatory amplifier of the death signal in type II cells \[[@B4]\]. Cytolytic granules function through the serine protease GrB, which activates apoptosis mainly through the mitochondrial pathway \[[@B12]\]. We investigated the *in vitro*response to externally added FasL or by transient transfection of pCMV GrB, huBax A15A5 and huBcl-2 A15A5 cells as described previously for human glioma cells \[[@B21]\]. As shown in figure [1C](#F1){ref-type="fig"}, the expression of Bax sensitized the A15A5 cells to apoptosis induced by both FasL and GrB while apoptosis was inhibited by the presence of Bcl-2 under the same conditions. Incidentally, the fact that huBax or huBcl-2 transfection modulated apoptotic sensitivity toward FasL suggests that the A15A5 cells belong to the type II group. Bcl-2 expression has been associated with retardation in cell cycle entry \[[@B22]\] and as such we examined the proliferation rate of the different transfected cells using a MTT assay as described in material and methods. All transfected cells appeared to have a similar doubling time, which means that the expression of human Bax or Bcl-2 transgenes did not affect their proliferative capacity *in vitro*(figure [2A](#F2){ref-type="fig"}). Similarly, no significant differences in the effect of the transfection of pCMV, huBcl-2 or huBax A15A5 in A15A5 cells on the clonogenicity of the glioma cells were observed as shown in figure [2B](#F2){ref-type="fig"}. Tumorigenicity of Bax and Bcl-2 transfected cells ------------------------------------------------- To determine the influence of Bax and Bcl-2 on tumoral growth *in vivo*, 10^4^pCMV, huBcl-2 or huBax A15A5 cells were delivered intracranially (ic) into syngenic BDIX rats as described in materials and methods. The survival rates of the different groups of rats were different as the median survival for the group of rats which had received huBcl-2 A15A5 and pCMV A15A5 cells were respectively 10.2 and 15.6 days, a difference, which was highly significant (*P*= 0.0086) (figure [3A](#F3){ref-type="fig"}). On the other hand, 80% of rats, which were injected with huBax A15A5 cells, had a disease-free survival of at least 30 days (figure [3A](#F3){ref-type="fig"}). It should be noted that 20% of the death, which occurred after injection of huBax, huBcl-2 or pCMV-A15A5 cells appeared to be due to operation-induced traumatisms since sham-operated animals gave similar results (**data not shown**). Brain sections from rats injected with huBcl-2 or A15A5 cells were histologically examined at the time of their death or after 30 days for rats injected with huBax A15A5 cells. As illustrated in figure [3B](#F3){ref-type="fig"}, brain sections from rats injected with huBax A15A5 cells showed little or no tumoral growth contrary to that observed in rats injected with pCMV or huBcl-2 A15A5 cells. This result suggested that tumoral growth was severely impaired in huBax A15A5 tumors but stimulated in huBcl-2 A15A5 tumors when compared to pCMV A15A5 tumors. To gain information about the kinetics of tumoral growth, cells were subcutaneously (s.c.) injected into the hindlimbs of syngenic BDIX rats and the volume of the tumors evaluated every 10 days. As shown in figure [3C](#F3){ref-type="fig"}, results similar to that obtained with i.c. experiments were observed as no or little growth was observed with huBax A15A5 tumors whereas huBcl-2 A15A5 tumors exhibited a faster growth than pCMV A15A5 tumors. To test the involvement of the innate immunity in the control of proliferation, the different types of transfected cells (10^5^) were inoculated s.c. into Swiss *nude*mice (see materials and methods). Tumor growth in the mice was measured for a period of 45 days (figure [4](#F4){ref-type="fig"}), tumors developed rapidly with similar kinetics for all the A15A5 transfected cells including huBax A15A5 cells. Immunoblot analysis of Bax and Bcl-2 in tumors did not reveal any changes in the expression of Bcl-2 between the established tumors and the huBcl-2 A15A5 cells (**data not shown**). This result showed that the growth of the huBax A15A5 was not due to the induction of rat Bcl-2 expression during tumoral growth in the Swiss *nude*mice. Accumulation of CD8+ T cells in huBax A15A5 tumors -------------------------------------------------- The latter result suggested that immune-induced apoptosis could be involved in the control of tumoral progression of A15A5 cells in BDIX rats, a feature partially lost in athymic mice. To assess local anti-tumoral response, we first investigated the presence of the CD8 marker in the different rat brain sections (figure [3B](#F3){ref-type="fig"}). The immunochemical analysis revealed a significant number of infiltrating CD8+ cells in huBax A15A5 tumors whereas huBcl-2 A15A5 or pCMV tumors showed very few CD8+ cells (figure [5A](#F5){ref-type="fig"}). Moreover, we noticed that necrotic tissue was absent, thus excluding a non-specific recruitment of lymphocytes due to an inflammatory process. Next, we also quantified the accumulation of CD8+ cells by flow cytometry and examined the phenotypes of the infiltrating immune cells in *ic*tumors for the presence of intra-tumoral lymphocytes, monocytes and dendritic cells or the loss of the major histocompatibility (MHC) class I molecules. Tumoral cells were dissociated and enzymatically treated as described in materials and methods. We observed a significant increase in double positive cells CD3/CD8 in huBax A15A5 tumors (\~17%) compared to huBcl-2 (\~4%) and pCMV (\~6%) tumors (figure [5B](#F5){ref-type="fig"}). An in-depth analysis of these CD3/CD8 cells (figure [5C](#F5){ref-type="fig"}) showed that only the lymphocyte T markers and the NK marker CD161 were significantly increased in huBax A15A5 tumors. On the other hand, no differences were found in the expression of class I MHC among the different tumors. The latter result suggested that T-cell mediated immunity could be involved in the rejection of the huBax A15A5 cells in syngenic rats. However, NK cells could play an auxiliary role in this process as suggested by the results obtained in *nude*mice. huBax A15A5 cells and A15A5 apoptotic bodies confer a protection against pCMV A15A5 cells and to a lesser extent against huBcl-2 A15A5 cells -------------------------------------------------------------------------------------------------------------------------------------------- The A15A5 rat glioma cells are highly immunogenic in the syngenic host \[[@B23]\], this immune response could eradicate tumors prone to apoptosis such as huBax A15A5 cells without affecting the viability of the cells resistant to apoptosis such as the pCMV or the huBcl-2 A15A5 cells. We have previously shown that apoptotic bodies derived from cells obtained from a rat colon carcinoma were a source of anti-tumoral antigens \[[@B7],[@B24]\]. It is conceivable that the enhanced rate of cell death observed in huBax A15A5 tumors *in vivo*could generate a stronger immune response to pCMV or huBcl-2 A15A5 tumors. To address this question, we compared the efficacy of an anti-tumoral immune response induced by huBax A15A5 cells to that observed with apoptotic bodies or oncolysates obtained from A15A5 cells (see protocol in figure [6A](#F6){ref-type="fig"}). Apoptotic bodies were generated from A15A5 cells treated with 10 mM NaB as described earlier \[[@B7]\] and their molecular characterization will be published elsewhere (Bougras et al. in preparation). BDIX rats were vaccinated with these apoptotic bodies using as a control A15A5 cell oncolysates or PBS (cf. materials and methods). The rats were then challenged either with pCMV A15A5 cells (figure [6B](#F6){ref-type="fig"}) or huBcl-2 A15A5 cells (figure [6C](#F6){ref-type="fig"}). As shown in figure [6B](#F6){ref-type="fig"}, the tumoral growth of pCMV A15A5 cells was reduced after the apoptotic body treatment although the effect was limited in amplitude and in time. On the other hand, no effect on tumor growth was observed after a similar treatment with A15A5 oncolysates (compared figure [6B](#F6){ref-type="fig"} with figure [3A](#F3){ref-type="fig"}). Interestingly, the treatment with apoptotic bodies triggered a weaker response to huBcl-2 A15A5 cells, which grew rapidly in animals treated with apoptotic bodies or oncolysate (figure [6C](#F6){ref-type="fig"}). As shown in figure [6B](#F6){ref-type="fig"}, rats that had received huBax A15A5 cells, then challenged with pCMV A15A5 cells developed small palpable tumors, which did not progress after 35 days post-challenging (i.e. pCMV A15A5 cell proliferation was abolished after 35 days by a huBax A15A5 cell pretreatment). Note that huBcl-2 A15A5 cell proliferation was reduced but not abolished in these rats (figure [6C](#F6){ref-type="fig"}). These results suggested that rats that had received huBax A15A5 cells were specifically protected against tumor growth and that the overexpression of Bcl-2 could not completely overcome this protection. Discussion ========== Tumor recurrence after surgical resection is often observed in GBM patients and additional chemo-or radiotherapy has not been shown to substantially improve survival in these patients \[[@B1]\]. In animal models, cancer vaccines have been shown to procure an appropriate immune response, to be highly specific and to favor tumor rejection \[[@B25]\]. In the case of CNS tumors, intensive research has been performed in the field of immunotherapy since the discovery that the CNS could not be regarded any longer as a completely immunologically privileged site \[[@B25]\]. Phase I studies have demonstrated the feasibility and the safety of this approach in human gliomas (see for example \[[@B26]\]). However, although some promising results have been obtained in preclinical studies, so far most clinical attempts have been disappointing. This setback in the application of immunotherapy is, however, not restricted to CNS tumors and new strategies are now being elaborated to enhance the efficiency of this approach. Others and we have observed that apoptotic bodies, the entities derived from apoptotic cells, could be a source of « new » anti-tumoral antigens and as such could be a source of potent tumor vaccines \[[@B7]-[@B9]\]. However, the nature of the cell death program, which gives an appropriate anti-tumoral immune response remains controversial \[[@B6]\]. Apoptosis is thought to be critical for the development and the progression of cancer and it appears to be involved in numerous steps in tumor progression \[[@B27]\]. Impairment or dysregulation of apoptosis clearly provides a selective advantage to tumoral cells. This resistance could allow the neoplastic cells to evade immunosurveillance as well as environmental changes inherent to tumoral transformation. Indeed, this could explain why, at diagnosis, most tumors have already acquired a certain resistance to apoptosis \[[@B28]\]. This positive selection for apoptosis-resistant tumor cells is accentuated by current therapies (chemo-and radio-therapies), which also use the apoptotic program to kill cancer cells, and thus tumors resistant to these treatments are often highly resistant to apoptosis \[[@B29]\]. Thus, all therapeutic approaches should take into account this innate or acquired resistance in the design of new anti-cancer strategies. We have addressed the question of the role of apoptosis in tumor progression by using human Bax or Bcl-2 transfected cells and then analyzing tumoral growth in rats. Several clones of the rat glioma A15A5 cells stably transfected with human Bcl-2 or Bax were used (figure [1A](#F1){ref-type="fig"}). *In vitro*experiments suggest that the expression of the trangenes confer the expected different sensitivities toward apoptosis (figure [1B](#F1){ref-type="fig"}). We also show that neither the *in vitro*proliferation nor clonogenicity of these cells was affected by the expression of Bcl-2 or Bax (figure [2](#F2){ref-type="fig"}). Quite remarkably, the expression of Bax suppresses the growth of these tumors in syngenic rats while that of Bcl-2 seems to stimulate the growth (figure [3](#F3){ref-type="fig"}). The control of tumor growth appeared to be under the control of a specific immune response against tumors. i)Since all tumors proliferated at the same rate in nude mice (figure [4](#F4){ref-type="fig"}). ii) In the rejected tumors (figure [5](#F5){ref-type="fig"}), a specific increase in CD8+ cytotoxic lymphocytes, which have the potential to recognize and attack the major histocompatibility complex (MHC) class I-expressing brain cells including tumoral cells \[[@B30],[@B31]\] were detected. iii) Syngenic rats that have received huBax A15A5 cells develop an anti-tumoral reaction against A15A5 cells (figure [6](#F6){ref-type="fig"}). However, this response was partially occluded by the presence of Bcl-2, a result consistent with the fact that its expression rendered the cells more resistant to apoptosis including that induced by the immune system (figure [1C](#F1){ref-type="fig"}). However, the accumulation of CD8+ CTL at the site of injection of huBax A15A5 cells could also suggest that induction of cell death in tumors facilitated or triggered a greater specific response. This could explain why an immune response to tumors was specifically observed in animals treated with huBax A15A5 cells (figure [6](#F6){ref-type="fig"}). On the other hand, the transfection with huBax could modify the phenotype of the rat cells as suggested by previous results \[[@B32]\]. However, MicroArray analysis of huBax A15A5 versus huBcl-2 A15A5 cells did not reveal any changes in the transcriptome of the cells (Cartron and Jézéquel, unpublished observation). Of note, in human glioblastomas expressing Bax ψ, a highly apoptogenic variant of Bax α \[[@B17]\], we observed an accumulation of intra tumoral CD8+ cells when compared to Bax α tumors (Bougras et al., unpublished results). The apoptosis index has been found to be of significant prognostic significance in patients with high-grade astrocytomas \[[@B33]\]. Our study provides three new findings, which should be considered in immunotherapy: i) the expression of pro-or anti-apoptotic molecules can control the response of the tumor to the immune system during the course of tumor progression (at least in class II). Thus, we suggest that the over-expression of Bcl-2, which often occurs during tumorigenesis could account for the escape from the immune surveillance. ii) Anti-apoptotic mechanisms that often impede the success of treatments could also be an obstacle to immunotherapy. In addition other anti-apoptotic processes such as the existence of soluble decoy receptor that impairs FasL induced-apoptosis in malignant gliomas \[[@B34]\] or the existence of an inactive granzyme \[[@B35]\] could also be involved in the resistance to the immune system. iii) The type of cell death (e.g. apoptosis vs. necrosis) that induces the best stimulation of the immune system is still a matter of controversy but our results suggest that cells sensitized to apoptosis are capable of providing a long lasting and efficient protection against tumoral growth. Interestingly, a rat model of colon carcinoma has been described in which some clones gave rise to tumors that constantly expanded in the animal to eventually formed metastasis (Pro) while others progress for several days before complete regression (Reg) \[[@B36]\]. The disappearance of the latter clone has been shown to be controlled and to be triggered by the immune system and the transfection of Reg cells by Bcl-2 has been shown to prevent apoptosis and to restore its tumorigenicity \[[@B36]\]. However, the effect of the ectopic expression of Bax in the progressive counterpart cell line Pro was not investigated in this work \[[@B36]\]. Our results show thus for the first time, in the same type of cells, the adverse effects of Bax and Bcl-2 in the antitumoral role of the immune system. Conclusion ========== Taken together, our data provide evidence that BCL-2 family members control tumor growth through their sensitivity to immune-induced cell death and enhancement of the immunogenicity of tumor cells. Competing interests =================== The authors declared that they have no competing interests. Abbreviations ============= Bax: Baxα ; CNS: central nervous system; CTL: cytotoxic T lymphocytes; GBM: Glioblastoma Multiforme; *ic*: intra-cerebral; *sc*: subcutaneous. Author\'s contributions ======================= Gwenola Bougras carried out the phenotyping characterization of the cell lines and in vivo experiments with tumors, Pierre Francois Cartron the molecular genetic studies, Fabien Gautier carried out the experimental studies with apoptotic bodies and Stephane Martin the intracerebral implantation of the rat glioma cell lines. Marité LeCabellec participated in immunohistochemical characterization of the tumors. Marc Grégoire and Khaled Meflah participated in the coordination of the study. Francois M. Vallette conceived of the study, and participated in its design and coordination and drafted the manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2407/4/54/prepub> Acknowledgements ================ We thank Dr L. Oliver (INSERM 419) and Dr B. Melchior (INSERM 437) for fruitful discussions throughout this work. We are indebted to Dr P. Saas and Mr S Perruche (EPI 119, Besançon, France), for advice on solid tumors dissociation. This work was supported by grants from the Institut National de la Santé Et de la Recherche Médicale, the Université of Nantes, the \"Association pour la Recherche sur le Cancer\" and the \"Ligue Départementale Loire-Atlantique contre le Cancer\". Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **A.**Western blot analyses of human Bcl-2 and Bax transgenes expression in stable A15A5 transfectants. Two clones per transfection with pCMV (E12 and C4), pCMV-Bcl-2 (Bcl-2: D6 and D4), pCMV-Bax (Bax: H11 and E4), which expressed approximately the same amount of trangenes were selected to be used in this study. **B.**The effect of huBcl-2 and huBax on the sensitivity of A15A5 cells to apoptosis, transfected cells was analysed after incubated with various cell death inducers as described in materials and methods. Apoptosis was determined by measuring the specific DEVDase and LDH activities after treatment of the cells with doxorubicin (doxo), staurosporine (STS), Na-Butyrate (NaB), serum deprivation (d-serum) or UV irradiation (UV). S.D. was calculated from 3 different experiments using 2 independent clones for each transgene. **C.**Cell death was induced by FasL and GrB. A15A5 cells were treated with FasL (0.5 μg/ml) or were transfected with a plasmid encoding for human GrB (200 μg plasmid) as described in **21**. In the latter case, we verified by FACS that similar amounts of human GrB were expressed in all transfected clones (data not shown). S.D. was calculated from 3 different experiments using 2 independent clones for each transgene. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **A.***In vitro*proliferation of transfected A15A5 cells was determined using a MTT assay. S.D. were calculated from 3 different experiments using for each, one of the 2 different clones for pCMV, Bcl-2 or Bax transfected cell lines. **B.**For the clonogenicity assay: 400 cells were plated into 6-well plates and colonies formed were counted as described in **21**. Data were given as number of CFU (colony forming units) with S.D. calculated from 3 independent clones for each transgene. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **A.**Growth of pCMV, Bcl-2 and Bax clones in BDIX syngenic rats. 10^5^cells from each clone were injected *ic*and rat survival was quantified using a Kaplan-Meier analysis. The curve illustrated is representative of 3 independent experiments. Each point represents the mean growth of 6 rats (3 rats per clone). **B.**The macroscopic observation of the tumors induced by the *ic*injection of huBcl-2, pCMV and huBax A15A5 cells into rat brains. Rat brains were stained with hematoxylin. **C.**Subcutaneous growth was followed by measuring the tumoral volume every 10 days after the injection of the cells. The curve illustrated is representative of 3 independent experiments. Each point represents the mean growth of 6 rats (3 rats per clone). :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Growth of pCMV, huBcl-2 and huBax A15A5 cells in Swiss *nude*mice. 10^4^cells from each clone were injected *sc*and tumor growth was estimated every week by measuring tumor volume. The curve illustrated is representative of 2 independent experiments. Each point represents the mean of the growth of 6 mice (3 mice per clone). (□) pCMV; (○) Bax; (●) Bcl-2. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **A.**Qualitative analysis of CD8+ infiltrating lymphocytes by immunohistochemistry. **B.**Flow cytometric analyses of intra-tumoral CD8+ CTLs. Dot plots show the expression of CD3+ and CD8+ cells in the dissociated tumors. The percentages indicate the double positive cells for the three types of tumors. Results are representative of three independent experiments. **C.**Phenotypes of the intra-tumoral cells by flow cytometry. Tumors were treated as described in materials and methods. Results presented here are the mean of 3 independent experiments. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **A.**Vaccination protocol for the anti-tumoral immune response induced by huBax A15A5 cells, apoptotic bodies or oncolysates to pCMV or huBcl-2 A15A5 cells in rats. Rats were pretreated with PBS, oncolysate (OL) or apoptotic bodies (AB) from pCMV A15A5 cells or huBax A15A5 cells, 15, 10 and 5 days before being challenged with pCMV or huBcl-2 A15A5 cells. **B.**Comparison of tumor growth of pCMV A15A5 cells injected into syngenic BDIX rats. Four groups of 6 BDIX rats were treated as in **A**then on day 0 were challenged with 10^5^pCMV A15A5 cells injected *sc*and tumoral growth was followed for an 60 days. Data presented are illustrative of 3 independent experiments. **C.**Comparison of tumor growth of huBcl-2 A15A5 cells injected into syngenic BDIX rats. Four groups of 6 BDIX rats were treated as in **A**then on day 0 were challenged with 10^5^huBcl-2 A15A5 cells injected *sc*and tumoral growth was followed for an 60 days. Data presented are illustrative of 3 independent experiments. :::  :::

PubMed Central

2024-06-05T03:55:47.338441

2004-8-24

PMC515306

Background ========== Although a rare disease, uveal melanoma of the eye is the most common primary intraocular malignancy in adults, with an incidence rate of up to 1.0 per 100,000 person years (age-standardized, world standard) in Europe \[[@B1]\]. Only a few consistent risk factors have been identified for this disease. One set of uncommon risk factors include predisposing diseases like the dysplastic nevus syndrome, atypical ocular nevi, ocular and oculardermal melanocytosis \[[@B2],[@B3]\]. Another set of host risk factors are ancestry, light skin and iris pigmentation \[[@B4]-[@B6]\]. In addition, a number of environmental factors including UV radiation \[[@B7],[@B8]\] are weakly or inconsistently associated with uveal melanoma. Some uveal melanoma are associated with neurofibromatosis. However, the vast majority of familial cases reported are non-syndromic \[[@B9]\]. Some recent studies suggests that mutations in the breast cancer susceptibility locus, BRCA2 on chromosome 13, may be involved in the development of uveal melanoma \[[@B9],[@B10]\]. Occupation may be also relevant, and may include chemical work \[[@B11],[@B12]\], arc welding \[[@B8],[@B12]\] and agriculture and farming work \[[@B13],[@B14]\]. Two recent studies found an increased risk of uveal melanoma among cooks \[[@B8],[@B15]\]. Electromagnetic waves with frequencies of 300 kilohertz (kHz) to 300 gigahertz (GHz) are called radio-frequency radiation. Typical occupational sources transmitting radio-frequency radiation in Germany include walkie-talkies in the military and security services, in plants, radio sets on ships, transporters, freight trains, police cars and wireless phones including cellular phones (C-net: 450--465 MHz, since the 1990 ies D-net: 890--960 MHz and E-net: 1710--1800 MHz) and cordless phones (800--1900 MHz) with different modulation types. The population-wide introduction of analog and digital mobile phone techniques in the recent years, which has been coined as the mobile revolution \[[@B16]\], has resulted in an increasing number of people who fear that radio-frequency radiation (RFR) may have adverse health effects \[[@B17]\]. There is currently much uncertainty about the role, if any, of radio frequency transmitted by radio sets or mobile phones in human carcinogenesis. The assessment of the potential association of radio-frequency radiation and cancer risk is hampered by uncertainties about effective electromagnetic frequency ranges, the lack of a clear biological mechanism, as well as by difficulties of exposure assessment. Until now, the majority of epidemiological cancer studies focussed on brain cancer because the brain may be exposed to RFR \[[@B18]-[@B20]\]. With the exception of one study by Hardell et al. \[[@B21]\], all brain cancer studies showed no association between RFR as emitted by mobile phones and brain tumour risk until now. In contrast, the pooled analysis of two recent German case-control studies on the aetiology of uveal melanoma showed that frequent use of radiofrequency radiation devices including radio sets and mobile phones at the work place is associated with an about 4.2-fold elevated risk for uveal melanoma \[[@B22]\]. However, several methodological limitations including a small study size and a crude exposure assessment complicated the interpretation of these findings. Here we present the study design of an ongoing incident case-control study on uveal melanoma (acronym: RIFA study) that focuses on radiofrequency radiation as transmitted by radio sets and wireless telephones. We expect to publish the results of the study in summer 2005. Methods/Design ============== Study questions --------------- The RIFA study is planned to answer several etiologic questions with a special focus on electromagnetic radiation especially radio-frequency radiation as emitted by mobile phones and radio sets. First, is the finding of an increased risk of uveal melanoma among subjects with frequent use of RFR devices reproducible? Second, if there is an association, is this association site-specific in terms of laterality of the uveal melanoma and major site of mobile phone use? Third, if there is an association, is there a dose-response relationship between RFR and uveal melanoma risk? Another set of etiologic question relates to pigmentation characteristics including iris colour, hair colour, tendency of the skin to burn and to tan, freckling, number of cutaneous nevi. A further study questions relates to exposure to work and leisure time related ambient ultraviolet radiation and uveal melanoma risk. In addition, our study focuses on several occupational exposures or jobs that are suspected to be associated with an increased risk of uveal melanoma \[[@B23],[@B15]\] including working in the chemical industry, farming, coal mining, welding, cooking, working in the health service sector etc. Finally, we focus on the association between cancer history of the index persons and their relatives (especially breast cancer) and risk of uveal melanoma. Case recruitment ---------------- The case recruitment is hospital-based and takes place in the Division of Ophthalmology, University of Essen which is a the referral centre for eye cancer in Germany, currently treating about 400--500 eye cancer patients per year. Eligible uveal melanoma cases have to fulfil several criteria. Patients with newly diagnosed first uveal melanoma located in the choroid, iris, and/or ciliary body \[[@B24]\] during the recruitment period from September 25^th^, 2002 to September 24^th^, 2004 are eligible, if they are referred to the Division of Ophthalmology, University of Duisburg-Essen during the recruitment period, are aged 20--74 years at diagnosis, are living in Germany, and are capable to complete the interview in German. The majority (about 70--80%) of uveal melanoma treated at the University Hospital Essen receive episcleral plaque therapy without histological verification. For this reason we did not include a reference pathologist who reviews the diagnostic certainty of the cases. Experiences from our previous case-control studies showed that there is a nearly perfect agreement between the local eye doctors in the reference centre and the international reference pathologist (Dr. Ian Cree, London) \[[@B22]\]. We considered a diagnosis of uveal melanoma as definite if the results of the clinical examination of the eye (ophthalmoscopy) and ultrasound (sometimes supplemented by fluorescence angiography, computertomography or magnetic resonance imaging) were unambiguous. The inclusion and exclusion criteria of the cases are listed in table [1](#T1){ref-type="table"}. Control recruitment ------------------- Interim analyses showed that the majority of cases comes from the territory of former West Germany. Figure [1](#F1){ref-type="fig"} displays the geographic distribution of cases treated for uveal melanoma at the Division of Ophthalmology, University of Duisburg-Essen, from September 24^th^, 2002 through March 31^th^, 2004. Assuming comparable incidences of uveal melanoma in the federal states of Germany, the crude rate (referred cases divided by population at risk aged 20--74 years) may be considered as an indicator of the referral effect. Obviously, the referral effect varies by federal states. However, it is difficult to judge whether the case referral to the Division of Ophthalmology, University of Duisburg-Essen is a random sample of all newly diagnosed uveal melanoma cases in Germany. We therefore decided to recruit three different control groups. First, if we assume that cases treated in Essen are a random sample of all cases in Germany, a population-based control group would be the most appropriate control group. For this approach, we randomly select controls from mandatory lists of residence that cover the total population of the city or local district. These lists are regarded as the most complete sampling frame for population-based studies in Germany. Second, if the referred cases are not a random sample of all newly diagnosed cases of uveal melanoma in Germany, a control group sampled from those ophthalmologists who referred cases to the Division of Ophthalmology, University of Duisburg-Essen, would be most appropriate \[[@B25],[@B26]\]. To increase the statistical power, we decided to include two controls per case. In addition, we recruit sibling controls of cases (matching ratio 1:1) in order to assess whether genetic factors may confound the effect of exposure. The sibling controls are matched in genetic background. The inclusion criteria of the three control groups are presented in table [2](#T2){ref-type="table"}. Power calculations ------------------ Based on our former uveal melanoma case-control studies \[[@B15],[@B22]\] we estimated to identify 480 eligible cases within a recruitment period of 24 month. With an anticipated response proportion of about 80%, we expect to interview 380 cases overall. Population-based prevalence estimates of mobile phone use in the general population are scant. A recent telephone survey from 2001 showed that 82% of male and 74% of female participants aged 14--44 years use mobile phones; within the age group 45--59 years, 63% of male and 58% of female participants use mobile phones. The oldest age group (\> = 60 years) shows considerable lower prevalences of mobile phone use (men: 49%, women: 25%) \[[@B27]\]. To determine the statistical power of the case-control study, we assumed several mobile phone prevalence estimates in the control group. We conducted all power calculations two-sided according to formulas of Woodward \[[@B28]\]. We chose α to be 5% and 1-β to be 90%. Detectable increases of odds ratio estimates by varying prevalences of mobile phone use in the control group are presented in table [3](#T3){ref-type="table"}. A case-control interview ratio of 380 to 760 would enable us to detect increased odds ratios in the range of 1.5 to 2.2 depending on the exposure prevalence in the control group (table [3](#T3){ref-type="table"}). Exposure assessment ------------------- Table [4](#T4){ref-type="table"} presents a list of exposures that are assessed in the RIFA study. The questionnaire on mobile phone use is the same instrument which has been used by the international case-control study on brain cancer and mobile phone use sponsored by the International Agency for Research on Cancer, called Interphone study \[[@B29]\]. In contrast to the Interphone study, we do not perform personal interviews but telephone interviews. For this reason, we cannot show photographs of different types of mobile phones in order to assess the detailed type of mobile phone used. Compared to other studies on the aetiology of uveal melanoma, the RIFA study uses a detailed assessment of pigmentary characteristics. The self-administered questionnaire contains a eye and hair colour card that allows the participants to choose the most appropriate colour of their eyes and hairs at age 20 years. During the telephone interview, the skin reaction to sun exposure (tanning ability, burning tendency) is asked according the concept of Fitzpatrick \[[@B30]\]. Fitzpatrick\'s original question contains both items (tanning ability, burning tendency) within a single question which is of methodological concern because the categorical answers given by the Fitzpatrick question may not presents all types of skin reactions as has been demonstrated by Rampen et al. \[[@B31]\]. To reduce this potential misclassification, we separated the Fitzpatrick items into two questions that separately ask about burning tendency and tanning ability. For the assessment of nevi of the upper arms and the dorsum of the feet with a diameter of at least 3 mm, participants receive a template with a 3 mm hole that enables them to count all nevi of this minimum size. In addition, the CATÍ (computer assisted telephone interview) includes a detailed history of sunburns in the recent 15 years before interview and tendency to freckle as a child. The CATI contains a question on eye colour with the typical categorical answers as has been used by several others (blue, grey, green, hazel, brown, black, \[[@B22]\] that will enable us to study the agreement between eye colour assessment by colour cards and colour categories. The course of the exposure assessment is displayed in figure [2](#F2){ref-type="fig"}. The exposure assessment starts with a self-administered questionnaire among subjects who agreed to participate and gave written informed consent. Subjects are chronologically asked about each job held for at least six months and included questions on the job tasks and industries as has been done in previous studies on the aetiology of uveal melanoma \[[@B22]\]. In addition, subjects are asked questions related to eye colour, hair colour, ever use of mobile phone, wireless telephone and radio set, and number of nevi. Subjects who have sent back the self-administered questionnaire undergo a computer-assisted telephone interview which takes about 30--40 minutes. For subjects who reported selected work tasks (e.g. cooking and food processing, welding and others), we use 16 job-specific supplementary questionnaires to obtain details of the job tasks and materials used. Quality assurance ----------------- The quality control program includes several procedures. The study is designed to fulfill the recommendations of the German Good Epidemiologic Practises \[[@B32]\]. Eight month before the main study started, a manual of standard operating procedures (MOP) was written and a pilot study of four weeks was conducted to test the field work and exposure questionnaires. After some minor revisions of the MOP, report forms, and questionnaire instruments, the principal investigator and participating epidemiologists had to sign that they fully agree with the final version of the MOP. Interviewers of the study were introduced into the field work and were blinded against our study hypotheses. After an initial interviewer training course, interviewers are regularly monitored and receive regular training courses. The recruitment progress, given as number of registered cases and controls, distribution of inclusion and exclusion criteria, response proportions, is monitored monthly. The analysis of nonresponse reasons is supplemented by an short questionnaire for subjects not willing to participate. This questionnaire includes few demographic and exposure items that help us to assess potential selection effects due to nonresponse. A plausibility control of the interview data is done quarterly and is the basis for the regular training courses of the interviewers. The completeness of case registration is checked by regular comparison of the list of registered cases with lists of admissions to the referral centre. In addition we compare our list of cases with data of the hospital information system that includes information on diagnoses. The self-administered short questionnaires are visually edited by the study personnel before the telephone interview starts. The visual editing includes a completeness check and coding of the life-long job history. For each job period, the occupation and branch of industry is coded according to ISCO-68 \[[@B33]\] and NACE 1993 \[[@B34]\]. These classifications have been repeatedly used in occupational case-control studies. Self-administered questionnaires with incomplete information or missing data are marked and questions are prepared for the telephone interviewer who is responsible to ask these questions before the main telephone interview starts. The CATI contains internal quality checks that prevent data entry errors. For example, interviewers are not able to fill in the detailed questions on mobile phones, if the entry question on ever having used mobile phone has been answered with no. Planned analyses ---------------- At the end of the field phase, the data are quality-checked. To estimate the effect of exposures on uveal melanoma risk, we will use conditional logistic regression that accounts for the matching factors and allows to control for potential confounding. We will classify people exposed to an occupational category if they ever worked within this category for at least six months. The quantification of mobile phone use will be based on average number of phone calls and average duration of phone calls per time unit. The association between pigmentary characteristics and uveal melanoma risk will be assessed by detailed matrix containing information on hair colour at age 20 years, eye colour, freckling tendency, and skin colour. Final results of these analyses are scheduled to be published in summer 2005. Competing interests =================== None declared. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2415/4/11/prepub> Acknowledgement =============== Sponsored by „Deutsche Forschungsgemeinschaft (DFG), Förderkennzeichen KFO 109/1-1\" and „Bundesamt für Strahlenschutz (BfS), Förderkennzeichen: M8811\" Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Geographic distribution of 368 uveal melanoma cases referred to the Division of Ophthalmology, University of Duisburg-Essen, September 25^th^, 2002 through March 31, 2004. Federal states, not listed in this table provided no cases. NRW = North Rhine-Westphalia, RP = Rhineland-Palatinate, HE = Hessen, NI = Lower Saxony, HB = Bremen, SL = Saarland, BW = Baden Wuerttemberg, SH = Schleswig-Holstein, TH = Turingia, BY = Bavaria, HH = Hamburg, ST = Saxony Anhalt, SN = Saxony :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Course of exposure assessment of the RIFA case-control study :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Inclusion criteria of the cases ::: **Characteristics** ----------------------------- -------------------------------------------------------- Diagnosis Posterior and anterior uveal melanoma ^1)^ ICD10   C69.3 (Choroid)   C69.4 (Ciliary body) ICD-O-3: Localisation   C69.42 (Iris)   C69.43 (Ciliary body +/- further choroid sections)   C69.3 (Choroid) ICD-O-3 Histological types   8720/3 (Malignant melamoma, NOS)   8770/3 (Mixed epithelioid and spindle cell melanoma)   8771/3 (Epithelioid cell melanoma)   8773/3 (Spindle cell melanoma, type A)   8774/3 (Spindle cell melanoma, type B) Date of the first diagnosis 25.09.2002 -- 24.09.2004 (24 months) Age at diagnosis 20--74 years^2)^ Sex Man or woman Residence Germany Language Being capable to complete the interview in German 1\) Also patients with uveal melanoma who were never referred to Division of Ophthalmology, University of Duisburg-Essen, for diagnostics. 2) Reference date for calculation of age in cases in the first visit at Division of Ophthalmology, University of Duisburg-Essen due to uveal melanoma, as long as there is no clue that the uveal melanoma was diagnosed more than three month earlier. ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Inclusion and exclusion criteria of three different control groups of the RIFA case-control study ::: **Characteristics** --------------------- --------------------------------------------------- ----------------------------------------------------------------------------------------------------------------------- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- **Population control** **Ophthalmologist control** **Sibling control** Age Within 5 year age group of case^1^ Within 5 year age group of case^1^ Age within +/- ten years compared to age of case. If more then one sibling lies in this range, the one closest to the case is chosen, unless he/she refuses. If so, the sibling fitting second best in age is chosen. Sex The same sex as case The same sex as case The same sex as the case; when a same sex sibling does not exist or refused, then the opposite -- sex sibling is chosen. Reference period Visit at the ophthalmologist\'s within 10 working days (possibly more)^2^before reference date set by case diagnosis. Diagnosis No history of cancer No history of cancer No history of cancer Language Being capable to complete the interview in German Being capable to complete the interview in German Being capable to complete the interview in German Region of residence Residence matched to case^3^ Germany 1\) The reference date is the date of the case diagnosis 2) With small ophthalmologist\'s practices it may become necessary to increase the span of ten working days in order to recruit ten eligible patients. 3) As residence matching is not performed at the level of single cities, controls will be taken from a sample in a city comparable size within a radius of 60 km around the case\'s habitation, if such a sample exists. ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Detectable increased Odds Ratios in relation to expected study size and expected prevalences of mobile phone use ::: Mobile phone prevalence in the control group Detectable Odds Ratio (Cases : Controls) ---------------------------------------------- ------------------------------------------ ----------- ----------- ----------- 480:960 380 : 760 300 : 600 250 : 500 0.05 2.0 2.2 2.4 2.6 0.10 1.7 1.8 2.0 2.1 0.15 1.6 1.7 1.8 1.9 0.20 1.5 1.6 1.7 1.8 0.25 1.5 1.6 1.7 1.7 0.30 1.5 1.5 1.6 1.7 0.35 1.5 1.5 1.6 1.7 Calculations according to Woodward; α = 0.05, β = 0.10; two-sided ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Overview of exposure assessment, RIFA case-control study ::: Measurement of exposure ------------------------------ ----------------------------------------------------------------------------------------- Phenotypical characteristics Hair colour with 20 year, eye colour, skin colour Ability to tan, propensity to burn Number of nevi \> 3 mm on the upper arms & dorsum of the feet Freckling as a child Sun exposure Occupational outside work, holidays in the last 15 years Sunburns in the last 15 years Artificial UV radiation Exposure sources to artificial UV radiation as welding & tanning devices Use of protective devices Use of sunglasses or hats Eye burns Snow blindness, welding burns or sunburns of the eyes Radio-frequency radiation Use of mobile phones, wireless telephones Jobs-specific supplementary Main task Cooking, welding, radar, working with electronic equipment, working with microwave Job groups Mining, farming, working with farm animals, agriculture, chemical industry, health care Medical history Index person and family cancer histories Social class Highest school degree, highest professional degree, income :::

PubMed Central

2024-06-05T03:55:47.341530

2004-8-19

PMC515307

Background ========== The term anetoderma (anetos = slack) refers to a circumscribed area of slack skin associated with a loss of dermal substance on palpation and a loss of elastic tissue on histological examination \[[@B1]\]. In the past, cases of primary anetoderma were divided into the Jadassohn-Pellizari type, in which the lesions are preceded by erythema or urticaria, and the Schweninger-Buzzi type, in which there are no preceding inflammatory lesions. This is now of historical interest only, because in the same patient some lesions may be preceded by inflammation and the others may not, and the prognosis and histology are identical in the two types \[[@B2],[@B3]\]. Anetoderma is a rare disorder that in the most usual form develop on the trunk, thighs and upper arms, less commonly on the neck and face and rarely elsewhere. The scalp, palms and soles are usually spared. We report a patient with anetoderma whose lesions present on distal extremities consisting of hands and calves. Case presentation ================= A 14-year-old boy was noticed to have had multiple white, non pruritic area on his distal extremities for two years. The lesions consisted of whitish papules and depressed areas with central protrusion. On clinical examination, the otherwise healthy looking patient\'s general appearance and mental state had lumbar lordosis, laxity in large joints, tibia vara, high-arched palate, and dental misalignment. Skin lesions consisted of scattered, white to skin-colored papules, less than 1 centimeter in diameter, and with central protrusion, that were distributed on the dorsum of fingers (Fig [1](#F1){ref-type="fig"}), forearms (Fig [2](#F2){ref-type="fig"}), distal portion of the thighs and on the calves (Fig [3](#F3){ref-type="fig"}). No lesions on the trunk or proximal areas of extremities were detected. Palms, soles, dorsum of feet and mucosal membranes were spared. No sensory changes associated with the lesions. He did not have any history of medication consumption. Family history was negative. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Solitary lesion over dorsum of left index finger. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### An anetodermic lesion on forearm. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### An anetodermic lesion on lateral aspect of left lower leg (before cryotherapy) :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Anetoderma lesions some minute after cryotherapy :::  ::: Laboratory examinations consisting of complete blood count, urinalysis, and blood chemistry including erythrocyte sedimentation rate and liver function tests were within normal limits. Antinuclear antibody was negative. The patient had no risk factor for AIDS or syphilis, so we did not request HIV or VDRL test. Hepatitis B surface antigens were not detected. Immunological assays consisting of IgM, IgA, IgG, IgE, C3, and C4 levels were normal. An induration of 0.5 centimeter in diameter was observed after tuberculin testing. Chest x-ray film was normal. The skin biopsy was done. Haematoxylin and eosin stained section showed faintly eosinophilic separated collagen fibers in the upper and mid-dermis (Fig [6](#F6){ref-type="fig"}). Verhoeff-vanGieson stained sections showed a marked decrease or in some areas total absence of elastic fibres, in both superficial and mid-dermis. Elastic fibres around the vessels in the affected areas are also fragmented and markedly decreased (Fig [7](#F7){ref-type="fig"},[8](#F8){ref-type="fig"}). We tried to treat the patient with liquid nitrogen cryotherapy by means of cotton-tip applicator for two freeze-thaw cycles (freeze time, 10--15 seconds), in 6--8 sessions weekly; which obtained moderate improvement in some early-onset lesions with no frank atrophy (Fig [4](#F4){ref-type="fig"} and [5](#F5){ref-type="fig"}). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### The same view as Fig. 4 after several sessions of cryotherapy (the 2 lower lesions near completely been resolved). :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### Faintly eosinophilic separated collagen bundles in upper dermis (H&E ×250). :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### Severely decreased elastic fibres in superficial and mid dermis (Verhoeff-van Gieson stain ×250). :::  ::: ::: {#F8 .fig} Figure 8 ::: {.caption} ###### Severely decreased elastic fibres in superficial dermis (Verhoeff-van Gieson stain ×400). :::  ::: Discussion ========== Anetoderma, which was first described by Jadassohn in 1892, is characterized by localized areas of loss of substance and elastic tissue with flaccid skin and often leads to a herniation phenomenon \[[@B2]\]. We could not find similar reports (other than anetoderma-like changes on distal extremities secondary to hamartomatous congenital melanocytic naevi) \[[@B4]\] of anetoderma developing on distal extremities without involvement of the upper trunk and proximal arms, in the medical literature. This rare disorder occurs mainly in women aged 20--40 years, but is occasionally reported in younger and older patients of both sexes. It is perhaps more frequent in central Europe than elsewhere, which suggests a possible relationship to chronic atrophic acrodermatitis (due to *Borrelia*species) in some cases. In the most usual form, crops of round or oval, pink macules 0.5--1 centimeter in diameter develop on trunk, thighs and upper arms, less commonly on the neck and face and rarely elsewhere\[[@B3]\]. The scalp, palms and soles are usually spared. Each macule extends for a week or two to reach the size of 2--3 centimeter \[[@B3]\]. Sometimes there are larger plaques of erythema, and nodules have also been reported as primary lesion \[[@B5]\]. The number of lesions varies widely, from less than five to one hundred or more \[[@B3]\]. The lesions remain unchanged throughout life, and new lesions often continue to develop for many years. If the lesions coalesce, they form large atrophic areas, which are indistinguishable from acquired cutis laxa \[[@B3]\]. They may become confluent, to cover large areas, especially at the roots of the limbs and on the neck \[[@B3]\]. Although infrequently reported, anetoderma may occur in families, and patient must be examined for associated systemic abnormalities for thorough assessment of their skin disorders. In familial anetoderma, there were associated ocular, gastrointestinal or orthopedic anomalies in the affected patients or in any other family members, but causes without them have been reported \[[@B8]\]. Although isolated and perhaps coincidental, these abnormalities could be related to the same process that produces the lesions of anetoderma \[[@B3]\]. Primary anetoderma can be inherited, but it has also been described in association with prematurity, lupus erythematosus, antiphospholipid syndrome, and with decreased serum levels of alpha-1-antitrypsin \[[@B6],[@B7],[@B9],[@B10]\] and \[[@B11]\]. Secondary anetoderma develops over other dermatoses are shown in Table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Dermatoses Associated with Anetoderma ::: --------------------------------------------- ---------------------------------------------- ------------------------------------------------------- Syphilis \[3\] Sarcoidosis \[17\] Granuloma annulare \[25\] Tuberculosis \[3\] Acne vulgaris \[17\] Hepatitis B virus immunization \[26\] Xanthomas \[3\] Leprosy \[17\] Primary Sjogren\'s syndrome \[27\] Nodular amyloidosis \[3\] Lupus erythematosus \[17, 21\] Lichen planus \[28\] Melanocytic naevi \[4\] Pilomatricoma \[18, 19\] Insect bites \[28\] Low serum level of α-1-antitrypsin \[6\] Prurigo nodularis \[20\] Lupus profundus \[30, 31\] Antiphospholipid syndrome \[6, 11, 1nd 24\] Cutaneous plasmacytoma \[21\] Discoid lupus (with herediyary C2 defficiency) \[32\] Recurrent deep vein thrombosis \[7\] Benign cutaneous lymphoid hyperplasia \[21\] Pityriasis versicolor \[33\] History of Graves\' disease \[7\] Urticaria pigmentosa \[22, 23\] Dermatofibroma \[34\] Familial type \[8\] Perifolliculitis \[23\] Penicillamine-induced \[34, 40\] Prematurity \[10\] Varicella \[24\] HIV-infection \[42\] --------------------------------------------- ---------------------------------------------- ------------------------------------------------------- ::: The differential diagnosis of anetoderma includes other focal dermal atrophies and miscellaneous diseases that must be differentiated from the skin herniation phenomenon of anetoderma \[[@B12]\], are shown in Table [2](#T2){ref-type="table"}. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Differential Diagnosis of Anetoderma \[12-16\] ::: ------------------------------------- ------------------------------------ Atrophic scars Discoid lupus erythematosus Lichen sclerosus et atrophicus Atrophoderma of Pasini and Pierini Corticosteroid-induced atrophy Perifollicular macular atrophy Morphea Perifollicular atrophoderma Atrophoderma vermiculare Striae distensae Focal dermal hypoplasia Naevus lipomatosus Connective tissue naevus Neurofibromas Cutis laxa postinflammatory elastolysis mid-dermal elastolysis Granulomatous slack skin acrodermatitis chronica atrophicans ------------------------------------- ------------------------------------ ::: Atrophoderma of Pasini and Pierini is a major source of confusion both etymologically and clinically. Patients have larger lesions with a sharp peripheral border dropping into a depression with no outpouching. On biopsy, elastin is normal, while collagen may be thickened, but this finding is difficult to quantify \[[@B12]\]. Perifollicular atrophoderma is most prominent on the dorsa of the hands and often is associated with multiple basal cell carcinomas and hair abnormalities in the Bazex syndrome \[[@B13]\]. Perifollicular atrophy also has been described in extreme forms of keratosis pilaris, in which large keratin plugs may produce a dilated patulous follicle. This condition usually found on the cheeks of young children. Both of these lesions mimic perifollicular anetoderma but lack elastin changes \[[@B12]\]. In focal dermal hypoplasia thinning or absence of dermis, rather than changes in elastin fibres, accounts for the proximity of the subcutis to the epidermis \[[@B12]\]. Cutis laxa, postinflammatory elastolysis \[[@B14]\], and mid-dermal elastolysis \[[@B15]\] share with anetoderma the property of cryptogenic loss of elastic fibres. Elastase-producing strains of staphylococcus epidermidis have been held responsible for perifollicular macular atrophy. Anetoderma has also been reported in 5 patients with false-positive syphilis serology, 3 of who also fulfilled the criteria for the antiphospholipid syndrome \[[@B29]\]. Its pathogenesis is not yet clearly established, but immunological mechanisms could play an important role in dermal elastolysis \[[@B35]\]. The association of primary anetoderma with decreased levels of alpha-1-antitrypsin may be of significance: Alpha-1-antitrypsin inhibits neutrophil elastase and its reduction may cause increased elastic activity and elastin breakdown. Phagocytosis of elastic fibres by macrophages has been found in primary anetoderma \[[@B36]\]. No antibodies have been demonstrated against elastic fibres \[[@B37]\]. Venencie et al. \[[@B38]\], suggested that the degradation of elastic fibres in patients with anetoderma is caused by enhanced expression of progelatinases A and B and production of the activated form of gelatinase A, and that the lack of control of these enzymes by tissue inhibitors of metalloproteinases is probably a key factor in the development and duration of anetodermic lesions. Ghomrasseni et al. \[[@B39]\], demonstrated that for the five samples of anetodermic skin, matrix metalloproteinase-1 (MMP-1) levels were significantly higher compared with the uninvolver cultures and the healthy samples. A significant increase of tissue inhibitors of metalloproteinase (TIMP-1) expression was also observed in the affected cultures of explants. The study demonstrated a significant increase in the production of gelatinase A (MMP-2), and no significant production of TIMP-2 in lesional skin compared with the samples from the two healthy donors. Penicillamine-induced anetoderma has also been reported \[[@B3],[@B34]\] and \[[@B40]\]. Penicillin and the antifibrinolytic drug ε-aminocaproic acid have been advocated \[[@B41]\], but Venencie et al. \[[@B3]\] studied 16 patients and found that no treatment was beneficial once the atrophy had developed. However, the wrinkled skin appearance in our patient had been present for 2 years, and his lesions did not show any signs of inflammation or pre-existing conditions like melanocytic naevi. Conclusions =========== In summary, we report a case of anetoderma with lesions on unusual sites. We did not find similar reports of acral anetoderma in the medical literature. According to this paper liquid nitrogen cryotherapy has moderate efficacy in the treatment of some of the early anetoderma lesions, without frank atrophy. Competing interest ================== None declared. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-5945/4/9/prepub> Acknowledgements ================ We would like to thank Dr. N. Shokrpour for editing the manuscript. Also we thank patient\'s family for permission us to publish the patient\'s details and photograph.

PubMed Central

2024-06-05T03:55:47.344604

2004-8-19

PMC515308

Background ========== Throughout the world, there is an increasing awareness of mental illness as a significant cause of morbidity \[[@B1]\]. This awareness has increased with the steady decline of morbidity due to nutritional disorders, communicable diseases and other forms of physical illness, especially in countries undergoing epidemiological transition \[[@B2]\]. Mental and behavioral disorders are common, affecting more than 25% of all people at some time during their lives \[[@B3]\]. They are also universal, affecting people of all countries and societies, regardless of age, gender and income. The point prevalence of mental illness in the adult population at any given time is about 10% \[[@B3]\]. Similarly, around 20% of all patients seen by primary health care providers have one or more mental health disorders \[[@B3]\]. The role of the community in the prevention and care of the mentally handicapped has now been widely acknowledged and is regarded as the most appropriate basis for the development of mental health programs. Several studies have shown that knowledge of public attitude to mental illness and its treatment is a vitally important prerequisite to the realization of successful community-based programs \[[@B4]-[@B6]\]. The recognition of mental disorder also depends on a careful evaluation of the norms, beliefs and customs within the individual\'s cultural environment. Furthermore, community attitude and beliefs play a role in determining help-seeking behavior and successful treatment of the mentally ill. Unarguably, ignorance and stigma prevent the mentally ill from seeking appropriate help. People tend to have strong beliefs about the mentally ill, and many of these concepts are based on prevailing local systems of belief \[[@B7]\]. In developing any mental health education program, the basis of such beliefs must be taken into consideration. People\'s beliefs regarding mental illness should not only be known, but the purpose of their beliefs should be understood. Such attitudes and beliefs about mental illness can only be studied within a cultural context. Although the knowledge and perception of mentally ill patients and their relatives regarding mental illness has been reported from southwest Nigeria \[[@B8]\], to date there is little research on public attitudes towards mental illness from northern Nigeria, a culturally distinct part of the country. This paper is therefore one of the first to report findings related to attitudinal research on mental illness from northern Nigeria. The 1995 Nigeria National Mental Health Policy advocates the integration of mental health promotion, treatment and rehabilitation into primary health care services (PHC). However, this goal cannot be successfully achieved without an understanding of community attitudes towards mental illness. We therefore set out to ascertain the perceptions, attitudes and beliefs of adults regarding the causes, manifestations and treatment options of mental illness in a traditional Hausa community near Kano, Nigeria. Methods ======= Study area ---------- The study population included 250 adults residing at Karfi village, about 15 kilometers from Kano city. This typical Hausa community has 11,314 inhabitants \[[@B9]\]. Majority are Muslims preoccupied with farming and petty trading. About 32.0% of the populace are literate. The village has one primary health centre and several traditional healers. Referrals from the health centers are sent to Kura General Hospital and occasionally to Aminu Kano Teaching Hospital. Informed consent was obtained from respondents prior to commencement of the interviews. Study design ------------ A cross sectional descriptive study Sample size and sampling technique ---------------------------------- A multistage sampling technique was adopted. In the first stage, four wards out of the seven in the village were randomly selected. After house numbering, 250 houses were selected from the four wards using a table of random numbers. Where more than one household was found in a house, one household was selected by balloting. Finally, one adult was selected at random from each household for the interview. A final sample size of 250 adults was therefore obtained. Instrument description/ Data collection --------------------------------------- We adopted and modified a pre-existing semi-structured questionnaire \[[@B7]\] to evaluate the perceptions and beliefs of adults in Karfi village. The questionnaire was in three parts; the first section inquired about personal data including age, sex, ethnicity, religion, marital status, educational level and occupation; the second part elicited awareness of existence of mental illness in the community, knowledge of causal factors, manifestation of the disorder and awareness and preference of treatment options, while the third part explored the attitudes, beliefs and perception of the respondent towards the mentally ill. Attitudes such as fear, avoidance, anger, suspicion, mistrust, hostility were considered negative, whereas sympathetic attitude, willingness to care for a mentally ill relative or friend and tolerance were considered positive attitudes. The study instrument was validated using a pilot study of 10 randomly selected households in a nearby village with similar demographic characteristics (Kumbotso). Results of the pilot study were used to modify content and wording of the questionnaire. Previously trained medical undergraduates fluent in Hausa language administered the questionnaires to the sample population. Data analysis ------------- The data was analyzed using the Epi-Info^®^6.0 statistical software package (CDC Atlanta, Georgia, USA). Descriptive statistics were depicted using absolute numbers, simple percentages, range, and measures of central tendency (mean, median) as appropriate. The Chi-square test was used to test the significance of associations between categorical groups. All tests of hypothesis were two-tailed with a type 1 error rate fixed at 5%. Results ======= In all, 250 respondents were interviewed with 167 males and 83 females giving a sex ratio of 2:1 in favor of males. Their ages ranged from 18 to 74 years. Majority (78.0%) of the respondents were aged between 25 and 60 years. The median age (± SD) for the respondents was 34.5 ± 4.6 years; 36.0 ± 2.3 years for males and 28.5 ± 1.6 years for females. The Hausa-Fulani ethnic group constituted 89% of respondents, and the rest were Yoruba 2.0%, Igbo 5.0% and other minority Nigerian tribes 4.0%. About 16.0% of respondents had primary education, 12.0% had secondary education, and 4.0% had tertiary (post-secondary school) education. Approximately 27.0% of the population sampled had no formal education. Nevertheless, 41.0% of all respondents had Quranic education. The majority of respondents (90.0%) were Muslims and the remaining 10.0% were Christians. About 10.0% were single, 80.0% were married, 6.0% were divorced and the remaining 4.0% were widowed. Forty eight per cent of the respondents were engaged in farming, 25.0% were full-time housewives, 20.0% were engaged in petty trading, 7.0% were students and the remaining 4.0% were civil servants. About 13 persons interviewed (5.2%) had a relative with mental illness, but because of the small sample size we did not separately analyze findings from this group of participants. The most common symptoms proffered by respondents as manifestations of mental illness (Table [1](#T1){ref-type="table"}) included aggression/destructiveness (22.0%), talkativeness (21.2%), eccentric behaviour (16.1%), and wandering (13.3%). Drug misuse in form of alcohol ingestion, cannabis and other psychoactive street drugs were identified as major causes of mental illness (34.3%), followed by effect of divine wrath or God\'s will (18.8%), magic or spirit possession (18.0%), and accidents/trauma (11.7%) (Table [2](#T2){ref-type="table"}). Heredity, family conflicts and financial distress/poverty were uncommon responses. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Respondents\' perceived manifestations of mental disorder ::: **Manifestation** **No.\* (%)** **Rank order** ---------------------------- --------------- ---------------- Aggression/destructiveness 173 (22.0) 1 Talkativeness 167 (21.2) 2 Eccentric behavior 127 (16.1) 3 Wandering 105 (13.3) 4 Self-neglect 86 (10.9) 5 Nudity 56 (7.1) 6 Restlessness/anxiety 50 (6.4) 7 Insomnia 15 (1.9) 8 Loss of consciousness 8 (1.0) 9 \* Multiple responses recorded. Percentages represent proportions of responses obtained. ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Perceived causes of mental illness ::: **Perceived cause** **No.\* (%)** **Rank** ------------------------------------- --------------- ---------- Misuse of drugs^†^ 88 (34.3) 1 Divine punishment, God\'s will 48 (18.8) 2 Magic, spirit possession 46 (18.0) 3 Accidents/trauma 30 (11.7) 4 Heredity 27 (10.5) 5 Family conflicts/marital disharmony 14 (5.5) 6 Financial distress/poverty 3 (1.2) 7 \*Multiple responses recorded. ^†^Include street drugs and alcohol. ::: The majority of respondents (46.0%) opted for orthodox medical care when asked about preferred source of treatment for the mentally ill. This was followed by spiritual healing (exorcism) (34.0%) and the use of traditional herbal medicines (18.0%) (Table [3](#T3){ref-type="table"}). ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Respondents\' preferred treatment for mental illness ::: **Response** **No. (%)** ---------------------- ------------- Orthodox Medicine 116 (46.0) Traditional Medicine 46 (18.0) Spiritual Healing 85 (34.0) Others 3 (2.0) Total 250 (100.0) ::: Table [4](#T4){ref-type="table"} shows that majority of the respondents harbored negative feelings towards the mentally ill, mainly in the form of fear (n = 113) and avoidance (n = 81). A total of 117 respondents (46.8%) were sympathetic towards the plight of the mentally sick with female respondents showing more inclination for sympathy compared to their male counterparts. The female respondents, however, tend to be fearful and avoid the mentally sick more than their male counterparts. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Distribution of attitude towards the mentally ill by gender ::: **Attitude** **Male** **Female** **Total** -------------- ----------- ------------ ----------- No. (%) No. (%) Fear 24 (20.8) 89 (79.2) 113 Avoidance 12 (14.4) 69 (85.6) 81 Anger 48 (96.8) 2 (3.2) 50 Suspicion 36 (90.4) 4 (9.6) 40 Hostility 32 (93.6) 3 (6.4) 35 Mistrust 28 (89.6) 4 (10.4) 32 Indifference 16 (96.0) 1 (4.0) 17 Sympathy 39 (33.6) 78 (66.4) 117 Kindness 17 (20.8) 62 (79.2) 79 ::: Literacy status was significantly associated with the type of feeling exhibited by the participants. Literate respondents were seven times more likely to exhibit positive feelings towards the mentally ill as compared to non-literate subjects (OR = 7.6, 95% confidence interval = 3.8--15.1) (Table [5](#T5){ref-type="table"}). ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Influence of literacy level of respondents on attitude towards the mentally ill ::: **Positive attitude** **Negative attitude** **Total** **Odds ratio (95%CI)\*** -------------- ----------------------- ----------------------- ----------- -------------------------- Literate 92 15 107 7.6 (3.8--15.1) Not literate 64 79 143 Total 156 94 250 \*CI = confidence interval, p \< 0.0001 ::: Discussion ========== Aggression/destructiveness, talkativeness, and eccentric behaviors were the most frequently mentioned perceived symptoms of mental illness by respondents. This finding suggests that one has to display behaviour that attracts public attention and is therefore socially disruptive, to be recognized as having a mental disorder. This finding is similar to that documented by White in Tanzania \[[@B10]\] and Asuni *et al*. \[[@B7]\] among Yoruba patients in Western Nigeria. It is notable that hallucinations and delusions that are frequently mentioned in the literature as prototypes of gross psychotic states were not mentioned by the respondents as features of mental illness, probably because such features are not as tangible as aggressive attitudes. Misuse of drugs ranked highest among the respondents as a perceived cause of mental disorders than most of the other traditional etiologies. This finding may not be unconnected with increasing use of illicit drugs among the youth in developing countries. Although drug abuse was acknowledged by Iliyasu and Last \[[@B11]\] in their work on mental illness in Kano, northern Nigeria as a leading cause of drug dependent psychosis, Holzinger and colleagues \[[@B12]\] reported that drugs and alcohol was not considered by schizophrenia patients or their relatives to be a common cause of mental illness. Divine punishment ranked second as a perceived causative factor. This response may not be unconnected with the leading response (drug misuse), as many individuals are of the belief that one evokes supernatural wrath by taking intoxicants thus leading to the development of mental illness \[[@B3]\]. Belief in demons as the cause of mental health problems is a well-known phenomenon in many cultures of the world \[[@B13]\] but in our study this factor was ranked 3^rd^place by respondents (18% of the responses). Our finding is also in contrast to Adebowale and Ogunlesi \[[@B8]\] who found that \"supernatural causes\" were the most acceptable etiological factor among both mentally ill patients and their relatives in southwest Nigeria. Only 1.0% of respondents admitted that financial distress or poverty was a possible cause of mental disorder. Such a low score in the face of the present adverse socio-economic conditions prevailing in Nigeria may be explained by the Hausa cultural cum religious belief in providence and patience in the face of adversity. Other factors have also been reported by investigators as being associated with perceived causes of mental illness among patients and their relatives. Srinivasan and Thara \[[@B14]\] found that patient gender and education, duration of illness, the key relative\'s education, and the nature of relationship were associated with family beliefs about the cause of mental illness. There was a higher score on the preference for modern medical care in treating psychiatric illness. Similar changes in attitude towards the modern scientific approach regarding mental disorders was documented by Alem et al \[[@B15]\] in their work on mental illness in Ethiopia and by Iliyasu and Last \[[@B11]\]. Fear and avoidance of the mentally sick was frequent among female respondents. Traditionally, men are expected to be outwardly brave and less submissive towards aggression. The high frequency of negative attitudes among our respondents may be due to the fact that the concept of the mentally sick has an unfavourable public image. It has been shown that people may evaluate mental illness negatively, reject and discriminate against mental patients, and base their views on traditional stereotypes \[[@B4],[@B10],[@B16]\]. Literacy was found to be significantly associated with positive attitude towards the mentally sick. Similar findings were reported by Madianos *et al*. \[[@B17]\] in Greece, and by Alem *et al*. \[[@B15]\] in Ethiopia. A study on community attitudes towards the mentally ill in New Zealand also reported that those who had previous contact with the mentally ill held informed and enlightened views \[[@B18]\]. Interestingly, a recent Hong Kong study reported a generally negative attitude towards the mentally ill despite a fairly good knowledge of mental illness among the respondents \[[@B19]\]. Our results support the hypothesis by Wolff *et al*. \[[@B19]\] that negative attitudes towards the mentally ill are fuelled by a lack of knowledge. Conclusions =========== This study demonstrates the need for educational programs aimed at demystifying mental illness. A better understanding of mental disorders among the public would allay fear and mistrust about mentally ill persons in the community as well as lessen stigmatization towards such persons. Our findings may be of utility to health policy makers in the design of community mental health education programs and community mental health services in existing primary health centers in Nigeria. Competing interests =================== None declared. Authors\' Contributions ======================= KM and IZ initiated the study, and participated in the field work. IZ and AIS did the preliminary analysis and wrote the draft manuscript. KM, AIS and IZ participated in the design of the study and performed the statistical analysis. KM, IZ, and AIS participated in its design and coordination. AMH, KM and IZ wrote the final version of the draft manuscript. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1472-698X/4/3/prepub> Acknowledgements ================ The authors are indebted to the Chairman of Kura Local government and the village head of Karfi village for granting permission for this study to be conducted. We also wish to thank the field workers and medical students of Bayero University, Kano, Nigeria, who diligently collected the data.

PubMed Central

2024-06-05T03:55:47.346256

2004-8-20

PMC515309

Background ========== The emergence of complete genomes for many organisms, including humans, has created the need for hypotheses concerning the \"function\" of specific genes that encode specific proteins. While \"function\" is interpreted by different workers in different ways \[[@B1]\], Darwinian theory (by axiom) requires that the term be connected to fitness; natural selection is the only mechanism admitted by theory to generate functional behavior in a living system, macro or molecular. This, in turn, implies that the hypotheses about function have a \"systems\" component, including the interaction of the protein with other proteins, their impact on the physiology (defined broadly) of the cell and organism, and the consequences of physiology in a changing ecosystem in a planetary context \[[@B2]\]. Systems hypotheses can be supported by information from many areas. Geology, paleontology, and genomics, for example, provide three records that capture the natural history of past life on Earth. At the same time, structural biology, genetics, and organic chemistry describe the structures, behaviors and reactivities of proteins that allow them to support present life. It has been appreciated that a combination of these six types of analysis provides insights into functional behavior of proteins that cannot be provided by any of these alone \[[@B2]\]. Over the long term, we expect that the histories of the geosphere, the biosphere, and the genosphere will converge to give a coherent picture showing the relationship between life and the planet that supports it. This picture will be based, however, on individual cases that serve as paradigms for making the connection. The aromatase family of proteins offers an interesting system to illustrate the power of this combination as a way to create hypotheses regarding protein function within a system \[[@B3]\]. These hypotheses are not \"proof\", of course, but are limiting in genomics-inspired biological experimentation, now that genomic data themselves are so abundant. Aromatases are cytochrome P450-dependent enzymes that use dioxygen to catalyze a multistep transformation of an androgenic steroid (such as testosterone) to an estrogenic steroid (such as estradiol) (Figure [1](#F1){ref-type="fig"}). The protein plays a key role in normal vertebrate reproductive biology--a role that appears to have arisen before fish and tetrapods (land vertebrates, including mammals) diverged some 375 million years ago \[[@B4]\]. Aromatase is important in modern medicine as well, especially in breast and other hormone-dependent cancers \[[@B5]\]. Different numbers of aromatase genes are found in different vertebrates. Two aromatase genes are known in teleost fish \[[@B6],[@B7]\]. Only a single gene is known in the horse \[[@B8]\], rat \[[@B9]\], and mouse \[[@B10]\]. Cattle have both a functional gene and a pseudogene built from homologs of exons 2, 3, 5, 8, and 9 of their functional gene; these are interspersed with a bovine repeat element \[[@B11],[@B12]\]. In several mammalian species, including humans and rabbits, a single gene yields multiple forms of the mRNA for aromatase in different tissues via alternative splicing \[[@B13]-[@B16]\]. A still different phenomenology is observed in the pig (*Sus scrofa*). Three different mRNA molecules had been reported in different tissues from pig \[[@B17]-[@B21]\]. Compelling evidence then emerged that the three variants of mRNA identified in cDNA studies arose from three paralogous genes \[[@B22]\], rather than from a single gene differentially spliced \[[@B23]\]. This implies that the three aromatase paralogs in pigs arose via gene duplications relatively recent in geologic time. Hypotheses relating to the function of the three aromatase paralogs depend in part on when those duplications took place. If they were very recent, the three genes might have helped pigs adapt to domestication. If they pre-dated the divergence of pigs and fish \[[@B6]\], they may have different roles that are very fundamental to reproductive endocrinology in vertebrates. We apply here a series of tools to generate better hypotheses concerning the aromatase family of paralogs in swine. Results ======= One strategy useful for understanding the function of genes correlates events in their molecular evolution with events occurring in the history of other genes in the same and/or neighboring lineages, and with events recorded in the geological and paleontological records \[[@B2]\]. We incorporated a tool to date the divergence of two or more genes through an analysis of transitions at synonymous sites of two-fold redundant coding systems, where the encoded amino acid has been conserved \[[@B24]\]. This analysis exploits the approach-to-equilibrium kinetic behavior displayed by these sites. The analysis yields a transition redundant exchange (TREx) distance for any gene pair where the synonymous sites have not equilibrated. To calibrate the silent TREx clock, inter-taxa histograms relating pig (*Sus scrofa*) and ox (*Bos taurus*) were constructed for transitions at the silent sites of two-fold redundant codon systems where the encoded amino acid was conserved between the species \[[@B24]\]. The major peaks associated with the separation of these two lineages was observed at *f*~2~= 0.87, corresponding to a TREx distance of *kt*= 0.332. As the fossil record constrains the date of divergence of these two lineages to be 60 ± 5 Ma \[[@B25]-[@B27]\], and the codon biases in modern *Sus scrofa*and *Bos taurus*project an equilibrium value for *f*~2~= 0.54 \[[@B24]\], the rate constants for transitions at the TREx silent sites were estimated to be ca. 2.8 × 10^-9^transitions/silent site/year during the time interval that separates these lineages. Analogous *f*~2~values were then obtained for other vertebrate aromatase pairs, including fish *vs.*tetrapods (*f*~2~= 0.56), birds versus mammals (*f*~2~= 0.612), primates versus ungulates (*f*~2~= 0.823), and horses versus artiodactyls (*f*~2~= 0.828). Assuming a time-invariant single lineage first order rate constant of 3.6 × 10^-9^changes/site/year and an equilibrium *f*~2~of 0.54, the corresponding dates of divergence are calculated to be 435, 258, 67, and 65 Ma respectively, with the oldest dates being the least precise. The last three of these dates of divergence are similar to those suggested by the paleontological record \[[@B28]\], within the error of the calculation, which reflects the modest number of characters used to calculate the *f*~2~values. A tree for the artiodactyl lineage was constructed from the corresponding TREx distances (Figure [2](#F2){ref-type="fig"}). This was found to be consistent with the tree constructed from other metrics. The TREx clock is not widely used. It may, however, provide more accurate dates in regions where synonymous transitions have not equilibrated than conventional clocks that combine data from synonymous transitions and synonymous transversions, or from non-synonymous changes. A comparison of different clocks will be provided in detail elsewhere (Benner *et al.*, in preparation). Briefly, the rate constants for transitions and transversions are more different than the two rate constants for purine-purine and pyrimidine-pyrimidine transitions. Further, nucleotide frequencies can be used to calibrate the end equilibrium points for two-fold redundant codon systems directly, and this permits an \"approach to equilibrium\" formalism, well known in chemical kinetics, to be applied \[[@B24],[@B29]-[@B31]\]. From the tree, the TREx distances from the ancestor of fetal and placental aromatase to the modern enzymes are 0.113-0.079 (using an endpoint of 0.54 to reflect equilibration at the silent sites), corresponding to a range in the time of divergence of 26--38 Ma. The TREx distances from the divergence of all of the porcine aromatases and the modern forms ranges from 0.082--0.116, corresponding to dates of divergence in the range of 27--39 Ma. This suggests that the three aromatase paralogs diverged in the late Eocene to mid Oligocene. To further correlate the duplication of the genes with the fossil record, genomic DNA was analyzed from relatives of *Sus scrofa*. Both peccary and babirusa seminal plasma (*Tayassu pecari*, from the Center for Reproduction of Endangered Species, Zoological Society of San Diego; *Babyrousa babyrussa*, from the Bronx Zoo, New York) was probed by PCR (Polymerase Chain Reaction) amplification using exon 4-specific primers \[[@B32]\]. Bands having the sizes expected for the corresponding aromatases were observed by agarose gel electrophoresis. Based on sequence similarity, two isoforms of aromatase were obtained from both peccary and babirusa as clones derived from the PCR products (Figure [3](#F3){ref-type="fig"}). This establishes that at least one of the duplications occurred before the Tayassuidae (the peccaries) diverged from the Suidae (the true pigs) ca. 35 Ma \[[@B33],[@B34]\]. These data are consistent with an evolutionary model that holds that the ancestor of pig and oxen (approximated in the fossil record by *Diacodexis*, from the early Eocene ca. 55 Ma) \[[@B35]\] contained a single aromatase gene, and that the paralogous genes in pig arose some 20 million years later. This suggests that the paralogs in pig can be explained neither in terms of the fundamentals of vertebrate reproductive endocrinology, nor as a consequence of swine domestication. This does, however, suggest that the emergence of the aromatase paralogs was approximately contemporaneous with the emergence of a litter in the Suoidea larger than that found in the ancestral artiodactyl condition. While ruminant and camelid artiodactyls have only one-two young per litter, suoids in general have at least two young per litter (as seen in peccaries) and most suines (true pigs) routinely have three-four young (up to 12 in the domestic pig, *Sus*). Note that there has long been the tacit assumption that large litters in suoids represent the primitive artiodactyl condition. Large litters are primitive for mammals in general, and because suoids are plesiomorphic in some anatomical conditions relative to other artiodactyls (e.g., short legs, retention of four digits, bunodont cheek teeth), they have been assumed to be plesiomorphic in other respects. Other data suggest that small litters are in fact the primitive artiodactyl condition. Tragulids (mouse deer or chevrotains) are surviving small, primitive ruminants that are not too dissimilar from *Diacodexis*in body form, but only have one-two young per litter. Additionally, fossil record data on pregnant oreodonts (an extinct group probably related to the ruminant/camelid artiodactyl lineage, but with a suoid-like plesiomorphic postcranial morphology) shows that they also only had one-two young \[[@B36],[@B37]\]. A cladogram of the Artiodactyla (Figure [4](#F4){ref-type="fig"}) illustrates the probable acquisition of multiparous versus uniparous reproductive strategies, and places the character of litters with typically more than two members emerging just before the divergence of Tayassuidae and Suidae. The approximate correlation in time of the aromatase divergence in Suoidea with the enlargement of litters in Suoidea suggests, as a hypothesis, that the two are functionally related. To expand on this hypothesis, we sought genomic signatures of functional change within the aromatase paralogs. The number of non-synonymous changes in the gene divided by the number of the synonymous changes, normalized for the number of non-synonymous and synonymous sites (the K~A~/K~S~value), strongly suggests functional change when the value is significantly greater than unity \[[@B38],[@B39]\], and is also an indicator of hypothetical functional change when the value is high on a branch of a tree relative to other branches of the same tree \[[@B40]-[@B43]\]. K~A~/K~S~values were reconstructed for individual branches of the evolutionary tree derived from the Darwin bioinformatics workbench (see Methods) using a distance matrix and ancestral states constructed by the method of Messier and Stewart \[[@B39]\]. The typical branch in the aromatase evolutionary tree has a K~A~/K~S~value of 0.35. A higher K~A~/K~S~value of 0.85 is found in the episodes of evolution near when the pig aromatases diverged. While a K~A~/K~S~value of 0.85 does not require the conclusion that positive selection occurred during the emergence of these aromatase paralogs, an inference based on the magnitude of K~A~/K~S~in one branch, relative to the K~A~/K~S~value for typical branches \[[@B40]-[@B43]\], suggests that adaptive changes occurred during the duplications of the aromatase genes in pigs. A complete maximum likelihood analysis of the aromatase gene family was performed using the PAUP and PAML programs. The resulting tree, generated in PAUP, is shown in Figure [5](#F5){ref-type="fig"}, with parameters estimated using PAML. Once more, the generation of paralogs in the pig was found to have occurred after the divergence of pigs from oxen. A high K~A~/K~S~value (0.93) was again found in the divergence of the swine isoforms on the branch leading to the ancestor of the placental and embryonic enzymes following their divergence from the pig ovarian enzyme. The distribution of substitutions along this branch is consistent with altered functional constraints for the placental and embryonic enzymes compared with their extinct and extant counterparts (Tables [1](#T1){ref-type="table"} and [2](#T2){ref-type="table"}) \[[@B44]\]. We correlated the episode of rapid sequence change during the emergence of the embryonic and placental paralogs with the structural biology of aromatase. A homology model of aromatase was built from progesterone 21-hydroxylase from rabbit liver (coordinates from PDB file 1DT6) \[[@B45]\], a homologous cytochrome P450-dependent monooxygenase. Residues undergoing replacement during the episodes represented by branches in Figure [5](#F5){ref-type="fig"} (branches 1--3) are highlighted in color on the 3D model using a program in prototype with HyperChem (Figure [6](#F6){ref-type="fig"}). Multiple features within the pattern of amino acid replacement were apparent. First, the sites accepting amino acid replacements in the branches with low K~A~/K~S~values (as represented by branch 2 in Figure [5](#F5){ref-type="fig"}) were typically scattered without any obvious pattern over the surface of the protein. This is expected for neutral drift, although an adaptive role for these replacements is not excluded by this analysis. In contrast, the distribution of sites accepting amino acid replacements during the episode of rapid sequence evolution of branch 1 (as indicated by a relatively high K~A~/K~S~value) involving pig paralogs was not random over the protein surface. Rather, the sites are clustered near the substrate binding pocket, and in a region of the surface believed to contact the co-reductant protein, as identified by mutagenesis experiments in the homolog \[[@B46],[@B47]\]. The clustering of amino acid replacements near a substrate binding site during an episode of rapid sequence evolution suggests that the substrate specificity of the protein might be changing in correlation with a change in the detailed physiological role of the protein. Recent reports suggest that the substrate and product specificities of the placental and embryonic enzymes are indeed different from those of the ovarian enzyme \[[@B23],[@B48]-[@B50]\]. Further, synthesis of estrogen by the ovarian enzyme is more dependent on the structure of the co-reductant than is the placental enzyme \[[@B51]\]. Our *in silico*analyses rationalize these experimental observations from a structural perspective. The coupling of an evolutionary analysis to a crystallographic analysis suggests that the amino acid changes are functionally significant. Discussion ========== Today, natural history holds some of the most intellectually challenging conundrums to ever fascinate the human mind. Further, natural history offers biological chemists the opportunity to place broad biological meaning on the detailed analysis of the structure reactivity of isolated biological molecules studied in a reductionist setting. To do so, however, natural history must be connected to the physical and molecular sciences, both in subject matter and in culture. In part to make this connection, natural historians have sought to change the research paradigm in their field to favor quantitative data directed towards the \"proof\" of hypotheses over \"story telling\". Proving hypotheses is difficult in natural history (*pace*the philosophical reality that no significant statement in empirical science can ever be said to be \"proven\"). The events of interest (such as the extinction of dinosaurs) are frequently distant in time, or require a passing of time (as for speciation), making them difficult to reproduce in a laboratory. The scale of the concepts involved (species, environments, planets) also does not lend these concepts to laboratory models and laboratory-controlled tests. Further, a reductionist approach, even when available, will not necessarily generate data that are relevant to the big issue that concerns the natural historian. The emphasis on data and proof has ameliorated the worst excesses of storytelling in natural history, with enormous positive impact. Just as natural historians were purifying their field of storytelling, however, whole genome sequences began to emerge. By dramatically increasing the quantity of chemical data concerning the molecular structures of proteins, genomics changed the limiting steps in biochemical and biomedical research. No longer was the typical researcher attempting to solve an organic chemical or biotechnological question (What is the sequence of my protein? How do I express it at high levels to get the sequence?) for a protein that had been selected for functional reasons. Today, the typical researcher knows the structure of many proteins, and wishes to select one for expression and study based on a hypothesis about its potential function. Here, the fact that any definition of function, which must make reference to fitness, requires some systems, ecological, or planetary context, makes the natural historian a natural source of hypotheses. Their full reductionist armamentarium is available in the laboratory to test and explore any hypothesis that the natural historian might provide. The biomedical researchers may like some guidance from the natural historian to narrow the broad selection, or to shorten the random walk, if only slightly. For this purpose, the forswearing by natural historians of storytelling has come at a most inopportune time. To the modern natural historian, creating hypothesis can easily be regarded as \"storytelling\". They are reluctant to do so, and may criticize as atavistic colleagues who do. This has created a vacuum in the scientific community. Very few laboratories exist that can draw upon an expertise in natural history to generate stories that create hypotheses for the researcher working in experimental biochemistry and molecular biology. This article is designed in part to illustrate how this vacuum might be filled. Here, we do not just tell a story based on natural history, or even a story based on natural history supplemented with physiology and molecular sequence data. Rather, we show how the addition of other data, including data from X-ray crystallography, can make a story sufficiently rich that it can be viewed as being internally consistent with a wide range of independent data drawn from independent sources. This creates a hypothesis that is more than a story, even if it is less than proven. With aromatase, the congruence of our different analyses makes a compelling suggestion that the three aromatase paralogs in pigs arose by two duplication events in the late Eocene or early Oligocene. The emergence of the aromatase paralogs corresponded approximately in time to the emergence of larger litter size in suines. This implies that the two duplication events are functionally related to the larger litter sizes. This inference is consistent with the physiological impact of estrogen synthesis by these paralogs in *Sus*. Steroid production by the porcine embryo is tightly controlled by the transient expression of aromatase and 17-hydroxylase (P450C17) between days 10 and 13 \[[@B20],[@B21],[@B52]\]. In contrast, estrogen synthesis by the equine embryo begins as early as day 6 and increases with embryo age and diameter \[[@B52]\]. The estrogen produced by the pig embryonic aromatase is believed to have an impact on the mobility, spacing, and implantation of the concepti \[[@B52]-[@B56]\]. Adequate spacing would appear to be required to manage a larger litter. This is consistent with a structural biological analysis that correlates specific amino acid replacements with specific changes in the substrate and product specificity of the protein \[[@B57]\]. Interestingly, the substrate specificity of human aromatase is reported to be more similar to that displayed by the pig placental enzyme than the ovarian form \[[@B48],[@B49]\]. This is an unexpected similarity given that our evolutionary analysis suggests a change in biochemical function along the fetal/placental branch in the Suidae. It should be noted that the hypothesis is supported by the combination of data that individually would not have strength past storytelling. Thus, the K~A~/K~S~ratio of 0.93 would not, by itself, compel any particular interpretation. Its implications are greater given the relatively low K~A~/K~S~ratios of other branches of the tree. But the addition of crystallographic information, itself not compelling, makes a combination that is more compelling. Further, this hypothesis generation itself generates discoveries that might lead to their own hypotheses. An analysis of the evolutionary branches separating pigs and humans suggests an additional episode of adaptive change. The branch leading to the ancestor of human aromatase (branch 3) has a remarkably high K~A~/K~S~ratio (13 non-synonymous and no synonymous changes; Figure [5](#F5){ref-type="fig"}). This *is*a K~A~/K~S~ratio greater than unity, and does (pending evaluation of its statistical significance) compel the inference of an episode of adaptive change. Intriguingly, these changes are also clustered in the same regions of the structure as those changing along branch 1 leading to the stem fetal/placental enzyme, near the substrate and co-reductant binding sites. This implies that the substrate/product specificity of the ancestral aromatase protein was not like that of either the human or the pig placental forms, but rather reflects features that arose convergently in these two species \[[@B58]\]. Notably, four of the sites (positions 47, 153, 219, 269) that undergo replacement during the emergence of pig placental aromatase from the last common ancestor are the same as four that arose in the emergence of the human aromatase from its last common ancestor. Of these, the amino acid replacements are identical at two sites (Thr → Met at site 153; His → Arg at site 269). The probability associated with randomly observing this pattern is extremely low (0.000021) \[[@B59]\]. An additional site is displaced by a single position in the sequence alignment (259/260). We hypothesize that these represent an example of adaptive parallel evolution. It is important to point out that even an analysis this broad is likely to cover only a small part of a complicated reproductive endocrinology that must be associated with larger litter sizes. For example, the exact nature of the products produced by individual aromatases remains controversial, and may be different in laboratory studies depending on the conditions where they are studied \[[@B50],[@B60]-[@B62]\]. This is especially the case with the 19-nortestosterone derivatives in Figure [1](#F1){ref-type="fig"}. Further, an elegant recent study by Corbin *et al.*\[[@B23]\] identified 1β-hydroxytestosterone as a novel product produced by recombinant pig ovarian aromatase that is absent from the products produced by the porcine placental paralog, or by either human or bovine aromatase. This testosterone derivative binds to an androgen receptor, consistent with physiological activity. This was unknown before just this year, suggesting that more endocrine novelties remain to be discovered. Any of these may be relevant to a test of this system. For example, these hypotheses make predictions about the product specificities of the two peccary aromatases reported here. In fact, some data suggest that uterine exposure to androgens severely decreases litter size and embryonic survival during the time of maternal recognition of pregnancy \[[@B63]\]. This is consistent with the hypothesis of Corbin *et al.*\[[@B50]\] that the evolution of the placental paralog is associated with increased efficiency of testosterone aromatization. This is also consistent with the current data, and the argument presented here. It goes without saying that still more factors might be associated with an increase in litter size from one-two (presumed in *Diacodexis*, see Figure [4](#F4){ref-type="fig"}) to 12 or more in domestic swine. Most trivially, this increase might be associated with an increase in ovulation rate, and/or an adjustment in the structures and binding specificities of estrogen receptors \[[@B64]\]. Nevertheless, the first aromatase duplication, shared by pigs and peccaries, appears to have happened in the late Eocene (recognizing the error associated with these dates), around 35 Ma (Figure [4](#F4){ref-type="fig"}). This was a time of great global change, with dramatic cooling in the higher latitudes. More archaic kinds of mammals (e.g., some earlier families of perissodactyls and artiodactyls) became extinct, while many modern families (including the Suidae and Tayassuidae) became established at this time \[[@B65]\]. Suoids differed from other contemporaneous ungulates in their commitment to omnivory, even though a few forms, such as the modern warthog *Phacochoerus aethiopicus*, are more specialized herbivores. Perhaps the ability to bear a slightly larger litter than other artiodactyls was advantageous to them in this time of global ecological transition. However, it should be noted that larger litters usually mean altricial (i.e., relatively underdeveloped) young, a reproductive strategy apparently not available to larger, cursorial (running-adapted) ungulates, which give birth to precocial (i.e., well developed) young that are fully locomotory at birth \[[@B66]\]. The second aromatase duplication, with the ensuing capacity to produce multiple young, probably occurred within the family Suidae, some time during the Oligocene. The molecular data suggest dates of divergence between porcine fetal and placental aromatases as between 27--38 Ma, and the earliest known suid is of early Oligocene age \[[@B67]\], around 33 Ma (Figure [4](#F4){ref-type="fig"}). Large litters may have characterized the entire suid family. While the extant subfamily Suinae is primarily a Plio-Pleistocene radiation, during the Oligocene to Pliocene suids were exceedingly diverse taxonomically (with six other subfamilies known) as well as individually abundant as fossils \[[@B32],[@B33],[@B67]\]. In contrast, the predominantly North American tayassuids were never as diverse. It is possible that this tremendous Old-World diversity of suids, which continues to this day, is related to their capacity for the production of large litters. This type of speculation opens questions. For example, the babirusa (an Indonesian pig) is reported to have average litters of one-two individuals \[[@B68],[@B69]\]. While it is possible that litters contain three-four individuals, the occurrence is low \[[@B70]\]. If the common ancestor of babirusa with the African/Eurasian Suinae had a larger litter, then the babirusa must be hypothesized to represent a reversion to the more primitive condition. At present, however, relatively little is known of either the molecular biology or the natural history of babirusa. The date of divergence from modern swine is placed between 12--26 million years \[[@B71],[@B72]\], while our TREx analysis using cytochrome *b*places this data at ca. 18 Ma (data not shown). Clearly, further study is warranted. Conclusions =========== The aromatase family offers an example where a combination of phylogenetic analysis, molecular evolutionary analysis, and chemical analysis set within the context of the paleontological and geological records, and supported by contemporary bioinformatics and molecular modeling tools, permits a higher order level of hypothesis generation concerning the function of proteins. Rather than simply an Enzyme Commission number (E.C. 1.14.14.1 for aromatase), a description of catalytic activity (the enzyme oxidizes testosterone), or a description of the regulatory pattern (the protein expressed between day 10 and 13), this type of analysis can generate a truly functional hypothesis: that the embryonic enzyme oxidizes testosterone as a way of managing the larger litter sizes that emerged in the Suoidea during a time of dramatic planetary cooling (ca. 35 Ma). Such hypotheses set a higher bar, and a more useful standard, for the field of systems biology. Evolutionary theory holds that the only mechanism for obtaining functional behavior in a biological system is natural selection. Selection, based on a frequently poorly defined concept of \"fitness\", is determined by a context that not only includes the cell and tissue, but also the organism, the ecosystem, and a changing planet \[[@B73]\]. One cannot expect a collection of expression data with a mathematical model, by themselves, to provide this type of functional information unless it is set in the organismic, ecosystem, and planetary context. The historical view, of the type outlined here, becomes a critical tool for constructing this setting (Supplementary Figure \[see [Additional File 1](#S1){ref-type="supplementary-material"}\]). Humans have evidently exploited the molecular biology of larger litters to select for pigs that have truly large litters (as many as 14) following their domestication. Evidence for ancient domestication of pigs comes, *inter alia*, from a study of Indo-European languages. Proto-Indo-European (PIE) language had words for \"pig\" (PIE *su*-, compared with Tocharian B *suwo*, Latin *sus*, Greek *us*, Sanskrit *sukara*, Church Slavic *svinija*, Old High German *swin*, and English *sow*; also compare PIE *porko*-, with Latin *porcus*, Church Slavic *prase*, Old High German *farah*, etc. \[[@B74]\]), indicating that the pig has been under human domestication for at least 6000 years, enough time to have suffered a significant impact on its genotype through husbandry. We are unable, at this time, to exploit complete genome sequences of pigs or other closely related taxa to discuss the impact of domestication on aromatase, steroid receptors, amphiregulins, or other proteins that appear to be associated with uterine capacity and large litter sizes in the domesticated pig \[[@B75]\]. With the anticipated complete genome sequences of representatives of various mammal orders, including artiodactyls, it should be possible to extend this planetary biology approach. Methods ======= Calculations were done under the RedHat Linux 6.3 operating system on an Intel-Pentium III instrument using Blackdown\'s Java-SDK 1.1.8. PAML calculations were done on an IBM PC using the Unix operating system. Sequence analyses were aided by the DARWIN bioinformatics package \[[@B76]\]. The DARWIN package can be obtained by emailing a request to <cbrg@inf.ethz.ch>. Initially, pairwise alignments were constructed for the aromatase protein sequences available in the database. An evolutionary distance in PAM units was calculated for each pair by applying the PamEstimator-package from DARWIN using an empirical log-odds matrix. From this, a preliminary evolutionary tree was built for the mammalian sequences, with branch lengths along internal nodes calculated to minimize a least-squares distance. The sequences of the ancestral genes and proteins at branch points in the tree were then reconstructed. From there, mutations (including fractional mutations) at both the DNA level and protein level were assigned to individual branches in the tree using the method of Fitch \[[@B77]\]. The evolutionary history of the aromatase family was then analyzed using the transition redundant exchange (TREx) metric based on an analysis of two-fold redundant codon systems \[[@B24],[@B78]\]. These were obtained for each pairwise comparison of aligned aromatase genes. The number (*n*) of two-fold redundant amino acids (Cys, Asp, Glu, Phe, His, Lys, Asn, Gln, and Tyr) that are conserved in the aligned pairs was determined. The number of those amino acids that are encoded by the same codon (*c*) was determined, and the fraction (*f*~2~= *c*/*n*) of the codons that are the same were then tabulated (Supplementary Table \[see [Additional File 2](#S2){ref-type="supplementary-material"}\]). The TREx distances were calculated from *f*~2~values using the expression *kt*= -ln((*f*~2~-E~quil~)/(1-E~quil~)), where E~quil~is the *f*~2~value expected after a large number of nucleotide substitutions have occurred at the synonymous sites \[[@B24]\]. The DNA sequences for aromatase were phylogenetically analyzed using a maximum likelihood framework in PAUP 4.0\* (beta 10) \[[@B79]\], with the following parameters: alpha value representing the gamma distribution (2.1), the transition-transversion ratio (1.6), proportion of invariable sites (0.24), and empirical base frequencies. The resulting topology of the tree mirrors those based on other molecular studies \[[@B80]\]. For inter-taxon analyses, families in the MasterCatalog (EraGen Biosciences, Madison WI) were identified that contained at least one representative protein from both of the taxa of interest. For these families, all inter-taxa pairs of genes were extracted, together with the pairwise protein sequence alignment. A pairwise alignment of the DNA sequences was then generated to follow the protein sequence alignment. If a family contained more than one sequence of a species belonging to one of the taxa analyzed, then those sequences were checked to determine whether they were duplicate entries into the database. If this was the case, only one of the duplicate sequences was retained in the analysis. A histogram of inter-taxa pairs was constructed, and the *f*~2~value characteristic of orthologs determined \[[@B24]\]. This was used to calibrate the TREx clock using the divergence of pigs and oxen, and pigs and humans. Codon biases were obtained from the CUTG (Codon Usage Tabulated from GenBank) made available by the Kazusa DNA Research Institute Foundation, Japan \[[@B81]\]. Pairwise TREx distances were used to generate lengths for the branches connecting the swine paralogs using the minimum evolution criterion in PAUP. This preliminary analysis was followed by a maximum likelihood analysis for the complete dataset using the PAML program \[[@B82]\]. This includes the assignment of K~A~/K~S~values to individual branches. Tests of parallel evolution were conducted using Converge \[[@B59]\], implementing the JTT model. Secondary structural data based on homology modeling for aromatases were generated using the DARWIN bioinformatics package, and in agreement with previous studies \[[@B83],[@B84]\]. Renderings of the three dimensional structure of the proteins were obtained using a beta version of the HyperProtein package (HyperCube, Gainesville FL, USA 32601). Authors\' contributions ======================= EAG performed the evolutionary, statistical and structural analyses, and prepared the manuscript. LGG cloned genes as part of his Masters work, and called the evolutionary problem to the attention of SAB. TL provided computational infrastructure. RCMS and FAS initiated the work with suid reproductive endocrinology, and supervised LGG. DRS and DAL did the initial bioinformatics analysis. CMJ provided paleontological expertise, constructed the cladogram, and helped prepare the manuscript. SAB has developed planetary biological analysis as a paradigm for generating hypotheses about the biological function of proteins, and prepared the manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 Illustration of planetary biology. This figure illustrates the concepts of planetary biology as they relate to combining genomic, paleontological, chemical and ecological records to understand the history of the biosphere. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 An analysis of silent nucleotide substitutions in vertebrate aromatases. The first five columns from the left indicate the index number of sequence 1 compared with sequence 2, the fraction of sites at conserved two-fold redundant coding systems that are identical (f2), the number of such sites that are conserved (c2), and the number of such sites overall (n2). The remaining columns report analogous data: for silent sites in codon systems where a change at the third nucleotide is silent only if the change is a pyrimidine-pyrimidine transition (f2y, c2y, n2y); in silent sites where a change at the third nucleotide is silent only if the change is a purine-purine transition (f2r, c2r, n2r); for the silent sites at three-fold redundant codon systems (f3, c3, n3); and for the silent sites at four-fold redundant codon systems (f4, c4, n4). ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We thank three anonymous reviewers for their invaluable comments. We also thank Alaric Falcon, Andres A. Kowalski and Ge Zhao for their assistance. This work was supported in part by N.I.H. grants GM 54075 and GM 067439-01 (S.A.B.), N.I.H. grant HD 21961 (R.C.M.S., F.A.S.) and USDA-NRICGP grant 98-35205-6739 (F.A.S., R.C.M.S.). Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Reactions catalyzed by aromatases on multiple androgenic substrates. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Dating the pig duplication events. An evolutionary tree, following the topology of Figure 5, showing estimated TREx distances for individual branches calculated from reconstructed ancestral sequences. The scale corresponds to evolutionary time (in million years) estimated from the TREx\'s using a first order rate constant for transitions of 3 × 10^-9^changes per base per year. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### The amino acid alignment of exon 4 of two aromatase isoforms from both peccary and babirusa sequences with exon 4 of pig aromatase isoforms ovarian, fetal, and placental. Asterisks represent conserved sites. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Cladogram of the order Artiodactyla showing the extant families and some selected extinct ones. Ruminantia includes the modern families Tragulidae, Giraffidae, Bovidae, Moschidae, and Cervidae, plus a number of extinct families. \"Dichobunidae\" is a paraphyletic assemblage of primitive taxa considered broadly ancestral to the later families. The interrelationships of the families reflect the \"traditional\" relationship based on morphology \[85\]. Different arrangements based on molecular information \[86, 87\] would alter the placement of the Camelidae and Hippopotamidae but would make no difference to the arguments presented here concerning the Suoidea. The interrelationships within the Suidae are based on information in several studies \[32, 67, 88, 89\]. Note that only a couple of extinct suid subfamilies are shown, and that only extant genera of Suinae are shown. Thick, medium-thick and thin lines represent family or above, subfamily and genera, respectively. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Phylogenetic tree for the 18 vertebrate aromatase genes. Numbers above the branches represent the K~A~/K~s~ratios, while numbers below indicate branches highlighted in Figure 6. Single and double asterisks represent bootstrap values of 95--99% and 100%, respectively. The following sequences were used: *Tilapia nilotica*(rainbow trout), gi:1613859, *Oryzias latipes*(medaka), gi:1786171, *Danio rerio*(zebrafish), gi:2306966, *Carassius auratus*(goldfish, ovary), gi:2662330, *Ictalurus punctatus*(catfish), gi:912802, *Carassius auratus*(goldfish, brain), gi:2662328, *Sus scrofa*(pig) placental, isoform 2, gi:1762232, *Sus scrofa*(pig) embryo, isoform 3, gi:1244543, *Sus scrofa*(pig) ovary, isoform 1, gi:1928957, *Bos taurus*(ox), gi:665546, *Equus caballus*(horse), gi:2921277, *Mus musculus*(mouse), gi:3046857, *Rattus norvegicus*(rat), gi:203804, *Oryctolagus cuniculus*(rabbit), gi:2493381, *Homo sapiens*(human), gi:28846, *Gallus gallus*(chicken), gi:211703, *Poephila guttata*(zebra finch, ovary), gi:926845, *Ovis aries*(sheep), gi:7673985. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### The distribution of amino acid replacements on the tertiary structure of cytochrome P450 homolog. Amino acid replacements occurring along branches highlighted in Figure 5 are shown in red. The substrate binding pocket and nicotinamide co-factor are colored yellow and purple, respectively. The sites that bind the co-reductant are highlighted in green for reference. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Frequency distributions of stem pig duplication substitutions versus substitutions on all other terrestrial vertebrate branches ::: Terrestrial vertebrates Non-synonymous substitutions Synonymous substitutions Totals ------------------------------------------------ ------------------------------ -------------------------- -------- Stem pig duplicates (Branch \'1\' in Figure 5) 23 9 32 Remaining branches 598 1449 2047 Totals 621 1458 2079 Fisher\'s exact test, *P*= 0.00000094784 \[44\]. ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Frequency distributions of stem pig duplication substitutions versus substitutions within the Laurasiatheria subtree ::: Laurasiatheria subtree Non-synonymous substitutions Synonymous substitutions Totals ------------------------------------------------ ------------------------------ -------------------------- -------- Stem pig duplicates (Branch \'1\' in Figure 5) 23 9 32 Remaining branches 232 258 490 Totals 255 267 522 Fisher\'s exact test, *P*= 0.0056688 \[44\]. :::

PubMed Central

2024-06-05T03:55:47.348505

2004-8-17

PMC515310

Background ========== Cerebrovascular stroke is a highly prevalent condition and the major cause of language impairment in adults. Immediately following a stroke about 38% of the affected population experience aphasia \[[@B1]\]. Spontaneous recovery is reported within the first six months after the event, while only minimal spontaneous improvements of language functions are expected after more than one year post-stroke \[[@B2]\]. Additional rehabilitation efforts have produced beneficial effects, as reported for speech and language therapy on the basis of different performance indices \[[@B3],[@B4]\]. In accordance with recent progress in neurorehabilitation, which takes into account evidence of the brain\'s capacity for reorganization \[[@B5]-[@B7]\], intensive language training (several hours per week) seems to be the premise for substantial improvement of language functions in the chronic stage \[[@B8]\]. To date, the evaluation of impairment and recovery of function, including training-induced improvement in aphasia, has been based mainly on performance in neuropsychological tests. This is now being increasingly complemented by measures of brain function. Different mechanisms and time courses of recovery of language function after brain damage have been discussed. Hemodynamic imaging suggests the involvement of two mechanisms: (1) regression of diaschisis (reduced metabolism and function in areas connected with the damaged brain tissue, which have been cut off from essential input) and (2) functional reorganization of the neuronal networks involved in language processing. Regression of diachisis in perilesional and more distant regions have been shown to contribute to recovery of function particularly in early phases of the recovery process \[[@B9]\]. In contrast, \"re\"-recruitment (that is, reorganization) of perilesional areas of the left hemisphere \[[@B10]\] or reactiviation of left hemisphere network components \[[@B11],[@B12]\] predict long-term recovery of language function. Moreover, recruitment of homotopic right-hemispheric areas may contribute to language recovery when the left-hemisphere language network components are permanently impaired \[[@B13]\]. However, it has been debated whether the recruitment of right hemispheric networks constitutes an additional potential for language processing or whether it is just a by-product of increased general activation. Others suggest this recruitment may even impair the recovery of left hemispheric areas, leading to a persistence of deficits \[[@B14]\]. Brain structures in the vicinity of structural lesions produce a larger amount of slow wave activity. This might be due to a loss of afferent input (e.g. from the lesion) or to a primary metabolic change within these perilesional areas \[[@B15]\]. These abnormal slow waves can be detected in the electroencephalogram (EEG) and, due to their focal generators, they can be localized using magnetic source imaging, a magnetencephalogram (MEG) based technique. In Abnormal Slow Wave Activity Mapping (ASWAM) \[[@B16]\], generators of abnormal slow waves are localized and mapped on to brain structures in order to identify areas that are active but incapable of normal function. A number of studies have demonstrated that focal slow waves indicate abnormality resulting from neurological damage such as contusions, tumors, or cerebrovascular stroke. In particular, abnormal slow wave activity in the delta-frequency range (1--4 Hz) has been found in areas adjacent to the structural lesion \[[@B17]-[@B19]\]. Since focal slow wave activity varies with changes in metabolism and blood flow due to the insult \[[@B19],[@B20]\], it has been described as characteristic of a \'dysfunctional state\' \[[@B21]\] of the neuronal tissue or a dysfunctional border zone with little ongoing information processing. In patients with brain tumors, this relationship between slow wave activity and metabolic changes was further elucidated by combining MEG and proton MR spectroscopic imaging \[[@B22]\]. A mild reduction of N-acetyl aspartate (NAA) and slight accumulation of lactate (Lac) was found in association with sources of focal slow wave activity in the border zones of the tumors, suggesting a border zone between seriously damaged and normal tissue with potential for re-recruitment in the course of the disease. The mapping of abnormal slow wave activity can be used not only to identify dysfunctional neuronal networks, but also to track changes in the course of recovery or treatment. For instance, de Jongh et al. \[[@B17]\] reported increased focal delta activity in the MEG before and a reduction after resection of brain tumors. The utility of \'abnormal slow wave mapping\' (ASWAM) in diagnostics, recovery, or treatment evaluation may be validated by covariation with neuropsychological measures. Lewine et al. \[[@B23]\] found a correlation between symptom resolution and MEG-slow wave reduction in patients with minor traumatic brain injury (TBI) and Hensel et al. \[[@B24]\] reported a decrease of EEG-delta amplitude and dipole strength parallel to spontaneous recovery of language functions across the first year post stroke in aphasia patients. The present study employed ASWAM before and after intensive language training in aphasic patients. If ASWAM qualifies for the evaluation of treatment or training-supported rehabilitation in chronic aphasics, changes in the intensity and distribution of focally generated abnormal slow wave activity should vary with improvement of language function after a specific intervention. Aphasics were recruited from an ongoing project evaluating the effectiveness of an intensive language training program. This program combines the learning principles of shaping and the efficacy of concentrated training \[[@B25],[@B6]\] while considering the principles of cortical reorganization \[[@B5]\]. In order to minimize any influence of spontaneous recovery on changes in the brain-function measure, only chronic aphasics were selected to participate either in 30 hours of Constrained-Induced Aphasia Therapy (CIAT) \[[@B25]\] or in 30 hours of massed model-based (MB) aphasia therapy \[[@B26]\]. All training sessions were scheduled within a two-week period. It was hypothesized that (a) aphasics would display an increased density of slow wave generators in the damaged (left) hemisphere before training, (b) this density would be reduced in the perilesional zone following language training and (c) there would be an improvement of language functions as evaluated by a standardized language test (Aachen Aphasia Test Battery, AAT) \[[@B27]\]. Results ======= Language functions ------------------ The average test performance of the entire patient group increased after language training, as indicated by the AAT profile (t(27) = 9.85, p \< 0.0001, paired t-test, two tailed). Similar improvements were found for the Token Test (t(27) = 6.10, p \< 0.0001). The average improvement of the profile score was 2.9 ± 1.3 points and 6.1 ± 5.3 points on the Token Test (T-scores). Twenty-five of the 28 patients improved on at least one subtest (N = 19) or subscale (N = 6) of the AAT. Maximum delta activity ---------------------- In 26 subjects the maximum activity of delta dipoles was found in the left hemisphere and in the vicinity of the structurally obvious lesion (as verified by structural MRT; see Figure [1](#F1){ref-type="fig"} for three representative subjects). In one patient, the maximum delta activity was located in the right hemisphere anterior to the homologue of the lesion, a finding consistent across measurements. (The patient had a very mild amnesic aphasia, displayed the highest AAT profile score of the entire group \[63.15\] and showed the least amount of delta activity.) In another subject, the maximum delta activity was located at the posterior border of a large left fronto-temporal lesion due to an ischemic infarct of the middle cerebral artery in the first measurement. After training, the maximum delta activity was found in the right-hemispheric area anterior to the homologue of the lesion. Notably, both measurements showed that this patient had clusters of delta activity next to the lesion and its right hemispheric homologue. Left hemispheric Delta Dipole Density (DDD) decreased after training and increased in the right hemisphere, which might explain the shift of peak activity to the right. The location of this delta focus remained stable across the two measurements (Rho: x-axis: .69, p \< 0.0001, y-axis: .85, p \< 0.0001, z-axis: .73, p \< 0.0001). The coordinates of maximum delta dipole density were exactly the same in eleven patients, while maximum delta activity shifted by one voxel in one of the three cardinal planes in eight patients, and by more than one voxel in nine patients. (As emphasized above, one patient displayed a reversal in hemispheric lateralization after training). Hemisphere-specific average delta activity ------------------------------------------ Thresholds were significantly higher in the left hemisphere (F(1,54) = 49.03, p \< 0.0001). Clusters of voxels with delta activity \> 2 SD above the average DDD in a group of 25 healthy controls were found in 26 of the 28 patients in the left hemisphere before training. Such clusters were found in the right hemisphere in only 7 patients. In two patients only, delta activity in the right hemisphere exceeded left hemisphere activity. Average delta activity was significantly more pronounced in the left hemisphere before and after training (for the pre-measurement the main effect HEMISPHERE was F(1,54) = 55.35, p \< 0.0001; for the post-measurement, F(1,54) = 46.55, p \<0 .0001: Figure [2](#F2){ref-type="fig"}). Twelve patients showed an increase in left hemisphere delta activity after training, while a decrease occurred in sixteen patients (Figure [3](#F3){ref-type="fig"}). This diverging pattern became evident in the non-significant interaction TIME\*HEMISPHERE (F(1,54) \< 1). An increase in delta activity of the left hemisphere tended to covary with a longer amount of time since the lesion (F(1,26) = 3.69, p = 0.06). Changes in DDD relative to improvement of langauge functions ------------------------------------------------------------ \"Magnitude of change\" in the left hemisphere was more pronounced in those patients who displayed significant improvement in at least one subtest of the AAT (N = 19) compared to patients with minor improvements (in at least one subscale) or no improvements (N = 9; F(1,26) = 4.95, p \< 0.05). Magnitude of change of left-hemispheric delta activity varied significantly with improvements in language functions (AAT profile: r = .60, p \< 0.002; Token Test: r = .46, p \<0 .02, Figure [4](#F4){ref-type="fig"}), while there was no correlation between right hemisphere magnitude of change and language measures (AAT profile: r = .-0.07; Token Test: r = 0.01). Discussion ========== The present results provide further evidence that Abnormal Slow Wave Activity Mapping (ASWAM) discloses generators of abnormal slow waves. The mapping of abnormal slow wave activity on to brain structures allows for the identification of areas that are active, but not capable of normal function. This is, to the best of our knowledge, the first report of a re-test after controlled neuropsychological/-linguistic intervention within the same subjects. The comparison between the two measurements indicates a high reliability of peak locations in left hemispheric perilesional areas, even though successful training modified this activity in magnitude and spatial distribution. In almost all patients, the region surrounding the structural lesion continuously and reliably produced abnormal slow waves, whereas only very few patients presented slow wave activity distant from the structurally confirmed lesion. The amount of perilesional slow wave activity was markedly altered in patients who had improved after training, and the magnitude of this change was related to the changes in language functions. Substantial functional improvements were achieved even in chronic aphasic states by shaping procedures, constraint of non-verbal communication and massed practice. This replicates and extends the findings of Pulvermüller et al. \[[@B25]\]. The present results further suggest that similar improvements can be achieved irrespective of the particular training procedures (a comparable improvement occurred in the model-based group). Both strategies might produce their effects -- at least in part -- by reorganizing brain regions next to a lesion. Following Liepert et al. \[[@B28]\], who demonstrated with transcranial magnetic stimulation that constraint-induced (CI) movement training of the arm expanded the area of the brain involved in generating activity in the muscles of the hand, we might assume a similar mechanism for the presently observed language improvements after intensive speech and language training in aphasics, namely an increased number of increasingly functional areas. In contrast, some patients, though displaying language improvement after training, exhibited an increase of delta activity in the vicinity of the lesion. One explanation for this might be that the functional capabilities of the affected brain area remain disturbed, with no further potential to be restored or re-integrated in the language \'network\', and this might inhibit or impair functionally intact regions. Further segregation of these continuously dysfunctional areas from the remaining network might then lead to improved language functions and consequently to increased delta activity. This hypothesis is supported by the correlation between language improvements and either decrease *or*increase of delta activity in perilesional areas. Moreover, the increase of delta activity was related to longer duration of disease. In most of the patients, re-integration of \'spared\' brain areas into the language network should be completed in time (in patients exhibiting an increase of slow wave activity, the time-since-lesion averaged 55.4 ± 39.3 months, compared to 35 ± 13.9 months in patients exhibiting a decrease). Therefore, \"dysfunctional\" delta activity might be related, at least in a subgroup of chronic patients, to functionally more favorable outcomes. The increase of delta activity might be explained by reduced reciprocal exchange of information within functionally intact and permanently impaired network components. Conclusions =========== Compared to the more conventional procedures for aphasia treatment in the chronic stage, the present training involved an intense use of language capabilities and a restraining of alternative, non-verbal methods of communication. In our opinion, any training that encourages speech production several hours a day over several days has the potential to be efficacious. Massed practice is likely to produce activity-dependent cortical reorganization, found to result from CI-movement therapy \[[@B28]-[@B31]\]. It is also presumed to be the basis for a long-term increase in the amount of use of the more-affected extremity and of improved language functions following short-term intensive training. Methods ======= Subjects -------- Twenty-eight patients suffering from chronic aphasia participated in the training (14 females, mean age 55 years, range 35--80 years; see Table [1](#T1){ref-type="table"} for clinical data). All patients were right-handed before brain injury, as assessed with the Edinburgh inventory \[[@B32]\]. In 20 patients, aphasia resulted from left-hemispheric ischemic stroke; in 8 patients it resulted from a hemorrhage affecting left-hemispheric areas. All patients were in a chronic state as defined by a time-since-lesion \> 12 months. The average duration of the time-since-lesion was 43.78 months (range 12--156 months). Structural whole-head MRI was available in 26 patients and the scans were performed within the two week training period. For the other 2 patients, a left hemisphere lesion was verified by inspection of earlier MRI examination. Prior to training, aphasia was diagnosed according to guidelines of the Aachen Aphasia Test \[[@B27]\], and aphasic syndromes were classified as Wernicke (N = 4), Broca (N = 13), amnesic (N = 2) and global aphasia (N = 3). Six patients could not be classified according to the 4 syndromes given on the basis of the AAT. Aphasia was evaluated as mild (N = 11), moderate (N = 16), or severe (N = 1). Patients were recruited from the local rehabilitation centre (Kliniken Schmieder Allensbach & Konstanz) or from self-help groups, or were referred by neurologists and speech therapists. Design and procedure -------------------- Since this report focuses on changes in slow wave activity, principles of speech and language training and results will only be summarized (a detailed description will be provided elsewhere). Training took place 3 hours/day for 10 consecutive days and included (for 18 patients) language exercises with increasing levels of difficulty \[[@B25]\] or (for 10 patients) model-based intervention (training based on the patients\' functional deficit, with the main aim of gradually improving spoken word production). The patients received only language therapy during the two-week training period to ensure that changes in slow wave activity were not induced by improvement of potential comorbid neurological impairment (e.g. hemiplegia). Language function was evaluated by two sensitive measures of change of aphasia severity: the profile score and the Token Test of the Aachen Aphasia Test \[[@B27]\]. Tests were administered by trained psychologists or speech therapists one day before the onset of training and one day after the completion of training. Language function improved significantly after training in both groups (see results) regardless of the type of training (TREATMENT\*TIME interaction for AAT profile: F(1,26) = .72, p \> 0.3; for Token Test: F(1,26) = 3.59, p= 0.07), therefore data from the two groups were pooled for ASWAM. Training groups did not differ significantly with respect to age or time-since-lesion. Data acquisition and analysis ----------------------------- Using a 148-channel whole-head neuromagnetometer (MAGNES™ 2500 WH, 4D Neuroimaging, San Diego, USA), MEG-measurements were collected twice: once on the day before training and once on the day after training. MEG was measured in a 5-minute resting period, during which subjects were asked to relax while staying awake, and to not engage in any specific mental activity. MEG recordings were obtained in a supine position. Subjects were asked to fixate a colored mark on the ceiling of the magnetically shielded room throughout the recording in order to avoid eye- and head-movement. A video camera installed inside the magnetically shielded room allowed for a monitoring of the subject\'s behavior and compliance throughout the experiment. Written informed consent was obtained from subjects prior to each MEG-session and the study was approved by the ethics committee of the University of Konstanz. The fiducial points, coils, and head shape were digitized with a Polhemus 3Space^®^Fasttrack prior to each measurement. The subject\'s head position relative to the pickup coils of the sensor was estimated before and after each measurement. MEG was recorded with a sampling rate of 678.17 Hz, using a 0.1--200 Hz band-pass filter. For artifact control, eye movements (EOG) were recorded from four electrodes attached to the left and right outer canthus and above and below the right eye. The electrocardiogram (ECG) was monitored via electrodes attached to the right collarbone and the lowest left rib using a Synamps amplifier (NEUROSCAN^®^). Data reduction and analysis --------------------------- Data were reduced by a factor of 16 and digitally filtered for the delta (1.5--4.0 Hz) frequency band using a digital band pass filter (Butterworth filter of the order 6). Artifact-free time segments were determined by visual inspection. Single equivalent current dipoles were fitted for each time point in the selected artifact free segments (distance of time points 24 ms.). Five non-overlapping channel groups over left, right, center, anterior, posterior regions were chosen for dipole modeling. A homogeneous sphere, which gives the best least-squares fit to the digitized patient\'s headshape below the selected sensors, served as a model for the volume conductor. Dipole fit solutions at time points satisfying the following requirements were accepted: (1) a dipole moment (q) of 10 nAm \< q \< 100 nAm; (2) a goodness of fit (GOF) greater than 0.90. These restrictions should ensure that neither artifacts nor small amplitude biological noise would affect the results, and that only dipolar fields that were generated by focal sources were analyzed. Each data-set was divided into 1000 voxels, each of 20 mm^3^, using the AFNI-to3d-software (AFNI-Analysis of functional neuroimages \[[@B33]\]). For each patient, the percentage of dipoles in the delta frequency band per second in each voxel was z-transformed and statistically compared to the dipole density distribution of a group of 25 healthy controls, which were considered a \'norm\' group \[[@B16]\]. Whole brain magnetic resonance images (TR = 19, TE = 5,6, Flip angle = 30°, FOV = 256 mm, 1 mm isotropic resolution) were acquired within the 2-week training period across a 256 mm slab from each subject using a Philips Gyroscan 1.5 Tesla scanner (Philips Medical Systems, Gyroscan ACS-T). MRIs were aligned to the coordinates of the MEG according to anatomical landmarks, coil positions and head shape information using the AFNI software. Focal abnormal slow wave activity and its changes after training were evaluated by the following measures --------------------------------------------------------------------------------------------------------- ### Maximum delta activity A spatial clustering algorithm (FWHM, Filter Width Half Maximum, 60 mm) was applied to smooth the data. Maximum activity, i.e. the voxel with the highest percentage of delta dipoles, was determined using the AFNI subroutine 3dExtrema. The localization of this maximum was determined on the x- (medial-lateral), y- (anterior-posterior) and z- (inferior-superior) plane for the two measurements (pre- & post-training). ### Hemisphere-specific areas of high delta dipole density The localization of areas with high dipole densities (adjacent clusters of voxels) within each hemisphere was analyzed by applying a narrow filter of FWHM 20 mm to the original data. A narrow filter was used to minimize the influence of more distant voxels with low dipole density on the areas of higher density. First, the voxel with maximum activity was determined for each hemisphere in each patient. By setting thresholds according to the following criteria, voxels with high dipole density were extracted and averaged for each hemisphere: a\. Only voxels with z-values within one standard deviation below the maximum of each patient were considered for averaging within each hemisphere. b\. If the z-value was smaller than 2 standard deviations, voxels were not considered. c\. If the peak density was below 2 SD, an iterative process was initiated. Thresholds were lowered until at least one voxel became apparent where the dipole densities were different before and after training. ### DDD changes relative to improvement of language functions A measure of the \"Magnitude of change\" was determined to evaluate changes in dipole density relative to changes in language functions pre- and post-treatment. The averaged (absolute) intensity of delta dipoles in voxels above threshold (in each hemisphere and each patient) was scaled by dividing the magnitude of change in each hemisphere (\|T~2~-T~1~\|) by the mean of both hemispheres before and after training ((T~1left~+T~1right~+T~2left~+T~2right~)/4). This number was then log transformed (to ensure a Gaussian distribution with respect to statistical tests) and submitted to statistical analyses. Two patients were excluded from this final analysis (12: predominantly right hemispheric delta, and 20: shift in lateralization). ### Statistics Changes in AAT test scores across the two assessments were verified by two-tailed t-tests. Stability of the location of the maximum delta activity was verified by correlation coefficients (Spearman\'s Rho) for the posterior-anterior, medial-lateral, and inferior-superior axes. Differences in thresholds and average DDD between hemispheres (for both measurements) were verified by means of analysis of variance (ANOVA), as were differences between patients that exhibited an increase or decrease of DDD concerning time-since-onset. Changes in language functions (Token Test, profile score AAT) relative to changes in DDD-magnitude were evaluated by Pearson correlation. Authors\' contributions ======================= MM, TE and BR participated in the design of the study. MM was responsible for conducting the study, performed data analysis and drafted the manuscript. TE and BR participated in the discussion and general conclusions. CW provided knowledge of data analysis and wrote most of the scripts used for data analysis. DD and GB conducted therapy, assisted in collecting the data, and provided experience of therapeutic issues. Acknowledgements ================ The work was supported by a grant from the Deutsche Forschungsgemeinschaft (For 348) and the Kuratorium ZNS (Schmieder Kliniken Allensbach). We thank Dr. Schmidt, Mr. Koebbel (MA) and Mr. Greitemann (MA) for their cooperation with patient recruitment and for helpful comments on the draft. We thank C.J. Robert for editorial assistance and helpful comments on the manuscript. Special thanks to N. Weisz for his helpful comments on the final version of the draft. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Maximum delta activity in three representative subjects.**MEG superimposed on individual structural MRI of each patient: Localization of the voxel with maximum delta dipole density (red square) in three representative subjects (no.15, 25 & 28) located at the border zone of the structural lesion (l = left; r = right). :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Hemisphere averages before (pre) and after (post) intensive language training.**Average of voxels with high delta dipole density (DDD) in each hemisphere before and after therapy (all patients). It is notable that, at both measurements, clusters of high DDD are predominantly located in the left hemisphere. ANOVA revealed significantly higher DDD averages in the left hemisphere for both measurements. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Direction of change of DDD after therapy in the left hemisphere.**Bidirectional change in delta dipole density (DDD) after therapy. As hypothesized, we observed a decrease in DDD in 16 patients after therapy, while there was an increase in 12 patients (Note: pre-DDD scaled to 1). :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Magnitude of change in DDD relative to changes in language measures.**Positive correlation between the \"Magnitude of change\" ((\|T~2~-T~1~\|)) in left hemispheric delta dipole density (DDD) and measures of language functions (Difference T~2~-T~1~, t-scores of AAT profile and Token Test). y-axis: data scaled by average delta activity of each patient and log-transformed. CIAT refers to patients treated according to the principles of Constrained-Induced aphasia therapy, MB refers to patients treated according to the principles of model-based aphasia therapy. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Relevant clinical and demographic parameters of the training groups (Constraint-Induced aphasia therapy, CIAT; Model-based aphasia therapy, MB). ::: ID Group Sex Age Onset (months)^1^ Etiology Syndrome Severity -------- --------------- ----------- ------ ------------------- ------------- ------------ ---------- 1 CIAT female 35 33 hemorrhagic Not-class. mild 2 CIAT female 53 32 hemorrhagic Broca mild 3 CIAT male 51 13 ischemic Wernicke moderate 4 CIAT male 69 33 ischemic Wernicke mild 5^2^ CIAT female 70 38 hemorrhagic Wernicke moderate 6 CIAT male 51 29 hemorrhagic Broca moderate 7 CIAT female 47 54 hemorrhagic Not-class. moderate 8 CIAT female 67 42 ischemic Amnesic mild 9 CIAT male 49 92 ischemic Broca moderate 10 CIAT male 41 46 hemorrhagic Not-class. mild 11 CIAT male 66 26 ischemic Not-class. moderate 12 CIAT male 39 56 ischemic Amnesic mild 13 CIAT male 36 12 ischemic Broca mild 14 CIAT female 47 87 ischemic Broca moderate 15 CIAT female 53 50 ischemic Broca moderate 16 CIAT female 80 23 ischemic Wernicke mild 17 CIAT male 47 29 ischemic Broca moderate 18 CIAT female 36 32 ischemic Global moderate 19 MB male 59 39 hemorrhagic Not-class. mild 20 MB male 37 36 ischemic Broca moderate 21 MB female 49 28 ischemic Global severe 22 MB male 57 40 ischemic Broca mild 23 MB male 62 156 ischemic Broca moderate 24^2^ MB female 76 13 ischemic Global moderate 25 MB male 66 64 ischemic Broca moderate 26 MB female 65 29 hemorrhagic Broca moderate 27 MB female 47 53 ischemic Broca mild 28 MB female 75 40 ischemic Not-class. moderate N = 28 18 CIAT 10 MB 14 ♀/14 ♂ 54,6 43,8 20 I, 8 H N = 26 ^1^MRI was acquired within the two-week training period. ^2^For patients 5 and 24, MRI could not be acquired within the training period. Lesions of the left hemisphere were verified by clinical case report forms obtained from earlier clinical presentation (for both patients, MRI was performed immediately after the insult). :::

PubMed Central

2024-06-05T03:55:47.352861

2004-8-25

PMC515311

Background ========== Aggressive primary prevention of coronary heart disease (CHD) is most appropriate in patients at relatively high risk of CHD events \[[@B1],[@B2]\]. The coronary artery calcium (CAC) score is an independent predictor of coronary heart disease risk \[[@B3]-[@B7]\], and therefore may help in deciding how aggressively to pursue cholesterol-lowering, anti-platelet therapy and other primary prevention strategies. To use a given CAC score result, however, one must know how that score compares with the score of an average person of the same sex, age and CHD risk factor profile. A CAC score of 50, for example, may be unusually high for a 40-year-old woman without other CHD risk factors, but unusually low for a 70-year-old man with hypertension. The same score, therefore, affects risk assessment in opposite directions for these two patients. How should a clinician use this CAC score (or any other) when assessing the CHD risk of a more typical patient, say a 60-year-old woman with hypertension and high cholesterol? To answer this question, we need to know the effects of age, sex and other CHD risk factors on the expected distribution of CAC scores. Several large cross-sectional studies have described the prevalence and extent of CAC among different age/sex groups \[[@B6],[@B8]-[@B10]\] without accounting for conventional CHD risk factors that may strongly influence predicted CAC scores. Five previous studies examined how CAC relates to conventional CHD risk factors \[[@B11]-[@B15]\]. Only one of these was adequately powered \[[@B15]\], none adequately accounted for the abnormal distribution of CAC scores, and none yielded estimates usable for clinical decision-making. We identified a large sample of men and women without clinical CHD who presented for electron beam computed tomography scanning. Using questionnaire data collected from these patients about smoking habits and medical history (hypertension, high cholesterol and diabetes), we determined how conventional CHD risk factors, along with age and sex, affect CAC scores. We then developed a method for combining information from conventional risk factors and the CAC score (easy spreadsheet calculator attached), and we present several examples illustrating how that method may be applied in common clinical situations. Methods ======= Study sample ------------ All persons referred by their physician to an electron beam computed tomography (EBCT) scanning center in Nashville, Tennessee for measurement of coronary artery calcification between May 15, 1995 and December 31, 1997 were eligible for inclusion. Subjects with a history of CHD or complaining currently of any chest pain were excluded, as were subjects for whom CHD risk factor data were incomplete or missing. Only the first CAC score was included for those who received more than one EBCT scan. Measurement of coronary heart disease risk factors -------------------------------------------------- Current age, sex and presence of CHD risk factors were elicited by questionnaire from subjects and referring physicians. Each subject was labeled with hypertension, high cholesterol and/or diabetes mellitus if they answered affirmatively to the question, \"Has your physician ever told you that you needed medicine for X?\", or if their physician confirmed that such a condition was documented in their medical records. Patients were labeled as smokers if they currently smoked or had quit smoking within the preceding 3 months. No direct measurements of blood pressure, lipids or glucose were taken for the purposes of this study. Estimation of the 10-year risk of coronary heart disease events --------------------------------------------------------------- We estimated the 10-year risk of a first CHD event using published mathematical models based on the Framingham study \[[@B16]\]. For this purpose, we assumed that subjects reporting hypertension had systolic blood pressures of 140--160 mmHg and/or diastolic blood pressures of 90--100 mmHg (Stage I hypertension), and that subjects without hypertension had systolic pressures of 120--130 and diastolic pressures of 80--85 mmHg. We also assumed that patients with high cholesterol had low-density lipoprotein (LDL) cholesterol levels of 130--159 mg/dl and high density lipoprotein (HDL) cholesterol levels of 35--44 mg/dl, whereas patients without high cholesterol had LDL cholesterol levels of 100--129 mg/dl and HDL cholesterol levels of 45--49 mg/dl (for men) or 50--59 mg/dl (for women). Smoking and diabetes mellitus were dichotomous variables in both Framingham models \[[@B16]\] and our data set. We then used published model coefficients \[[@B16]\] to estimate the 10-year risk for each patient in our study. Measurement of the CAC score ---------------------------- Each subject underwent electron beam computed tomography scanning with an Imatron C-100 or C-150 scanner (Imatron, South San Francisco, California) after giving written informed consent. During a single breath hold, 40 consecutive slices of 3 mm thickness were obtained starting at the level of the carina and proceeding to the level of the diaphragm. Scans were obtained within 100 ms and were electrocardiographically triggered at 60--80% of the R-R interval. Coronary calcification was defined as a plaque of at least 3 consecutive pixels (area = 1.03 mm^2^) with density ≥ 130 Hounsfield units. The CAC score was calculated according to the method described by Agatston \[[@B17]\]. Statistical analysis -------------------- We categorized patients according to age and sex, and examined histograms, quantile plots and box plots in each category to evaluate distributional normality. The CAC score is fundamentally not normally distributed because of the large percentage of zero measurements, and hence is not amenable to a normalizing transformation, as noted by others \[[@B13]\]. We also considered a censored normal distribution, which would have allowed a one-step Tobit regression analysis. However, even after square- and cube-root transformations, the zero scores were distributed in a manner inconsistent with the Tobit regression model. After exclusion of zero values, however, the log-transformed CAC score was approximately normally distributed (Figure [1](#F1){ref-type="fig"}). This led us naturally to a two-stage modeling approach. We first applied logistic regression to model the probability of a non-zero score, and then used linear regression to model the actual CAC score, log-transformed, for the subset of patients with non-zero values. Using this methodology, we assessed the independent effects of CHD risk factors on both the presence and extent of CAC. We considered three sets of predictors: 1) age and sex, 2) age, sex, hypertension, high cholesterol, smoking, and diabetes, and 3) the Framingham 10-year CHD risk estimate. We examined whether the effects of age were linear (as opposed to J-shaped, for example) by testing a quadratic term in the model containing only age and sex. We evaluated the ability of each logistic model to discriminate subjects at high and low risk for CAC using the C-statistic, and estimated the proportion of variability in the extent of CAC explained in each linear regression model using the adjusted-R^2^statistic. Finally, we used coefficients, intercepts and residual variance from logistic and linear models to estimate the probability that the CAC score of an individual with known risk factors would fall into each of four standard CAC score categories (0, 1--100, 101--400, and \>400). We estimated these probabilities, using models containing the 10-year risk estimate as the only predictor, for a range of 10-year risk estimates. We also estimated these probabilities, using models with all CHD risk factor predictors, for the specific clinical scenario described in the Introduction (a 60-year-old woman with hypertension and high cholesterol) and for several other scenarios. We compared the actual distribution of CAC scores among 58--62-year-old women with hypertension and high cholesterol in our sample (n = 130) with predictions from 1) our two-stage model, 2) a one-stage model using Ln(CAC score + 1) as a continuous outcome in a linear regression model, and 3) a one-stage model using a censored normal distribution of cube-root transformed CAC scores (a Tobit regression model). This comparison was made both graphically and statistically, using X^2^tests with 3 degrees of freedom to compare the expected frequencies based on each model with the observed frequencies. Lower p values, in this case, indicate a poorer fit of the model to the observed data. All statistical analyses were performed with Stata 7.0 (College Station, Texas). Combining information from conventional risk factors and the CAC score ---------------------------------------------------------------------- First, we calculated the Framingham 10-year CHD risk estimate (and corresponding 1-year risk estimate assuming an equal event rate each year) according to published models \[[@B16]\]. Next, we calculated the probability, as described above, that that individual\'s CAC score would fall into each one of four standard CAC score categories \[[@B15],[@B18],[@B19]\] (0, 1--100, 101--400, and \>400). We obtained risk factor-adjusted relative risk (RR) estimates from a meta-analysis \[[@B7]\] comparing the risk of a CHD event among persons with CAC scores of 1--100 (RR = 2.1), 101--400 (RR = 5.4) and \<400 (RR = 10) to the risk in a person with a CAC score of zero. The analysis was repeated using more conservative estimates from the same paper: RR = 1.7 (for CAC 1--100), RR = 3.0 (for CAC 101--400), and RR = 4.3 (for CAC\>400). The post-test CHD risk estimates for each CAC score category were then calculated algebraically by assuming that the overall 1-year CHD risk estimate represents an average of the 1-year risk estimates from the four CAC score categories, weighted by the probabilities that an individual\'s score would fall into each category. A spreadsheet that automates these calculations is attached. Results ======= Study population ---------------- We identified 9341 persons without chest pain or a history of CHD presenting for their first EBCT scan between 4/15/95 and 12/31/97. Our sample was mostly middle-aged, but included persons as young as 35 years and as old as 88 years of age. Forty percent were women. The proportion with cardiac risk factors was high, though only 9% were diabetic (Table [1](#T1){ref-type="table"}). Framingham 10-year CHD risk estimates ranged widely, mostly dependent on age, but most were between 7% and 15%. Coronary artery calcium score distributions ------------------------------------------- Coronary artery calcium scores ranged from 0 to 4058. The mean score (± standard deviation) was 135 (± 377), and the median was 4 (25^th^--75^th^percentile: 0 -- 87). The prevalence of zero scores ranged from 80% among women younger than 50 years to 5% among men 70 years old or older. After excluding zero scores, log-transformed CAC scores were approximately normally distributed, and appeared to be strongly associated with age and sex (Figure [1](#F1){ref-type="fig"}). Predictors of the presence and extent of coronary artery calcification ---------------------------------------------------------------------- Age and sex were strong predictors of the presence of CAC in logistic regression models (Table [2](#T2){ref-type="table"}). There was no evidence that the effects of age were non-linear (i.e. J- or U-shaped) (p-value = 0.32 for a quadratic age term). Conventional CHD risk factors were also independent predictors of the presence of CAC (p \< 0.001 in all cases). The logistic model with age, sex and all CHD risk factors produced the most accurate model (C-statistic = 0.78). The Framingham 10-year CHD risk estimate was also a very strong predictor of coronary artery calcification, though the model containing the 10-year risk estimate as the only predictor was slightly less accurate (C-statistic = 0.74). Among patients with non-zero CAC scores, age and sex remained strong predictors of the extent of coronary artery calcification, as measured by the Ln(CAC score) (Table [3](#T3){ref-type="table"}). Again, the effects of age appeared to be linear (p = 0.16 for the quadratic age term). All conventional CHD risk factors remained statistically significant predictors of the extent of coronary artery calcification (p \< 0.001 for all predictors except high cholesterol at p = 0.004). Again, the Framingham 10-year CHD risk estimate was a very strong predictor of the extent of calcification, though when used alone in a model, it explained somewhat less of the variance (R^2^= 0.11) than the full model (R^2^= 0.17). Coronary artery calcium distribution predictions ------------------------------------------------ Using these models, we estimated the probability of measuring a CAC score in each of four standard CAC score categories (0, 1--100, 101--400, and \>400) using the Framingham 10-year CHD risk estimate, a value easily calculated from conventional CHD risk factors using accessible web- or handheld computer-based software. These probabilities ranged widely based on the value of the 10-year risk estimate, with the probability of measuring a zero CAC score going from 75% (at a 10-year risk of 2.5%) to 13% (at a 10-year risk of 25%) (Table [4](#T4){ref-type="table"}). Risk integration example ------------------------ Using the case example presented in the Background section, we calculated that a 60-year-old woman with Stage I hypertension (140/90 mmHg) and high cholesterol (LDL cholesterol = 155 mg/dl, HDL cholesterol = 40 mg/dl) will have a 15% risk of suffering a CHD event in 10 years, according to the Framingham equation. If this women undergoes EBCT scanning, our models predict a 47% chance that her CAC score will be zero, a 36% chance that it will be between 1--100, a 12% chance that it will be between 101--400, and a 5% chance that it will be greater than 400. By integrating this information with previously published relative risk estimates (see [Additional File 1](#S1){ref-type="supplementary-material"}), we estimate her 10-year CHD risk to be as low as 6% (if her CAC score is 0), or as high as 51% (if her CAC score is \>400). These estimates are only moderately sensitive to variation in the relative risk assumptions (Table [5](#T5){ref-type="table"}), and may be easily calculated in any clinical scenario in which CHD risk factor data is available; see Table [5](#T5){ref-type="table"} for several other examples. Comparing predictions from different modeling strategies -------------------------------------------------------- Our strategy outperformed two other modeling strategies in predicting the actual CAC distribution among the 58--62-year-old non-smoking non-diabetic women with hypertension and high cholesterol in our study sample (n = 127) (Figure [2](#F2){ref-type="fig"}). The one-stage regression model using Ln(CAC score +1) as the outcome, which has been utilized extensively in previous research \[[@B11],[@B12],[@B14],[@B20]\], performed particularly poorly. Discussion ========== In this article, we present a clinically useful method of combining information from the CAC score with pre-test coronary risk estimates. To fully appreciate the utility of this analysis, it may be worthwhile to discuss the example from the Background section further. According to current guidelines, this 60-year-old woman, whose 10-year CHD risk estimate is about 15%, should receive both aspirin and cholesterol-lowering drug therapy, aiming for a goal LDL cholesterol of 130 mg/dl \[[@B1],[@B2]\]. After measuring her CAC score, however, there is a good chance (64%) that our recommendations would change. If her CAC score were zero (47% chance), our estimate of her 10-year CHD risk would be approximately halved (6--9%). Given this information, we would continue to recommend a healthy diet and exercise, but might decide that cholesterol-lowering medication is unnecessary \[[@B1]\], and that the benefits of aspirin in terms of CHD prevention do not outweigh the risk of hemorrhagic stroke associated with aspirin use \[[@B2]\]. On the other hand, if her CAC score were over 100 (17% chance), our estimate of her CHD risk would be approximately doubled (25--31% if CAC score = 101--400) or tripled (34--51% if CAC score \> 400). In such a case, we would certainly recommend both aspirin \[[@B2]\] and cholesterol-lowering medication \[[@B1]\] and would probably aim for a more aggressive LDL cholesterol goal of \< 100 mg/dl \[[@B1]\]. The probability that her treatment plan would be altered by measurement of her CAC score, therefore, is approximately 64% (the probability that her score is either 0 or \>100 = 47% + 17%), indicating likely usefulness of the test in this situation. The third and fourth clinical scenarios presented in Table [5](#T5){ref-type="table"}, on the other hand, provide examples where the test is unlikely to change management. The 40-year-old woman who smokes, for example, has a very low pre-test 10-year CHD risk (3%). It is very likely her CAC score will be zero (89%) or less than 100 (10%), in which case her post-test 10-year CHD risk will still be low (≤ 5%) and her management would not change. The 80-year-old man with high cholesterol has a high pre-test 10-year CHD risk (26%) and a high probability of having a high CAC score (70% will have a score \> 100), in which case his post-test 10-year CHD risk would remain over 20% and his management would have to remain aggressive. In these cases, and others in which the risk factor profile indicates very low or very high pre-test risk, the test is not likely to provide useful results, and the clinician might decide not to order the test. We have provided a simple spreadsheet (see [Additional File 1](#S1){ref-type="supplementary-material"}) that may be used by readers of this article to replicate these analyses and apply our models to other clinical scenarios. While others have proposed similar Bayesian approaches to use of the CAC score for coronary risk prediction \[[@B6],[@B21]-[@B24]\], ours has advantages. Previous approaches do generally take into account the pre-test probability of coronary heart disease, but none consider the expected distribution of CAC scores in the tested population after adjustment for conventional CHD risk factors. Raggi et al advocate use of an age- and sex-adjusted calcium score percentile, but this ignores both persons with zero scores and the strong effects of other risk factors such as hypertension and hypercholesterolemia \[[@B6]\]. Some approaches use only sensitivity and specificity from dichotomized CAC score cutoffs \[[@B21],[@B23]\], and others use CAC score-specific relative risks generated from a single study population \[[@B6],[@B24]\]. Only two provide actual post-test risk estimates for specific clinical situations \[[@B23],[@B24]\]. Our approach takes into account the pre-test coronary risk, the expected distribution of CAC scores adjusted for all conventional CHD risk factors, and summary adjusted relative risks from a recent meta-analysis, and provides clinically relevant post-test risk estimates that may be directly useful to primary care physicians, cardiologists and patients as they decide whether or not to take medications for primary prevention of CHD. This analysis confirms that conventional risk factors for CHD (hypertension, diabetes, smoking and high cholesterol, as well as increasing age and male sex) are independent predictors of coronary artery calcification. This finding is consistent with previous studies \[[@B11]-[@B15]\]. We also present expected CAC score distributions for a variety of clinical situations, which are not easily calculated from other studies, via Tables [4](#T4){ref-type="table"} and [5](#T5){ref-type="table"} and the attached spreadsheet calculator. Our finding that high cholesterol was less strongly associated with the extent of CAC than other CHD risk factors is consistent with the other large study addressing this issue \[[@B15]\], and perhaps reflects effective medical treatment for hypercholesterolemia. Male sex was a very strong predictor of the presence and extent of CAC -- women with the same CHD risk factor profile would be expected to develop CAC approximately 12 years later than men, and remain approximately 11 years behind men in the extent of their calcification. Finally, our analysis provides a guide for how to use the CAC score as a surrogate outcome when studying causes of coronary artery disease (a widely used study design \[[@B25]-[@B27]\]). The central problem with this approach is the fundamentally non-normal distribution of CAC scores, which makes parametric statistic testing (including both simple t-tests and multivariable linear regression) invalid. In dealing with this issue, some researchers have used the Ln(CAC score +1) as an outcome in linear regression analyses \[[@B11],[@B12],[@B14],[@B20]\]. This approach is not ideal, as the Ln(CAC score +1) is still grossly non-normal -- there are too many zero scores. Adding 1 to the CAC score makes the log-transformation possible (yielding zeroes instead of negative infinity), but it does not solve the distributional problem, and leads to predictions that misrepresent actual CAC score distributions (Figure [2](#F2){ref-type="fig"}). This observation has led others to present only non-parametric percentile data without multivariable modeling \[[@B6],[@B8]-[@B10]\], but this approach does not allow adjustment for conventional CHD risk factors that we have shown are strong predictors of the CAC score. One other group used ordinal logistic regression analysis to analyze CAC scores categorized into four ordinal categories (quartiles in their study sample) \[[@B13]\]. While such an approach does allow multivariable modeling with ordinal logistic regression, it does not take full advantage of the continuous nature of the CAC score and may blur the important distinction between zero and non-zero scores. Our analysis suggests that a two-step approach (using first logistic regression to model the risk of having a non-zero score, then linear regression of log-transformed non-zero CAC scores to model the extent of coronary calcification) will allow multivariable analysis of the interval data provided by the CAC score without violating the basic assumptions of parametric statistics. Our analysis has a number of limitations, perhaps the most important being a lack of clinical detail about participants. While we had information about conventional risk factors (hypertension, high cholesterol, diabetes mellitus and tobacco use), the data were only available from a questionnaire, and were not confirmed by direct measurement. Only dichotomous indicators of such conditions were used. Furthermore, other conditions and indicators of high CHD risk such as family history of CHD, obesity, physical activity, income, education, and levels of C-reactive protein, triglycerides and Lp(a), for example, were unavailable. Whether such factors are important predictors of the presence and extent of coronary artery calcification is unknown. On the other hand, CHD risk assessment is often based on the same type of limited information we had available on each of our patients, so the models we present are perhaps more easily applicable to common clinical situations than models based on more detailed clinical data. Furthermore, a historical indicator of past exposure to high blood pressure or high cholesterol, as we had access to in this study, may actually be more useful as a predictor of CAC than treated blood pressure measured at one point in time. Another important limitation of this study is our lack of data on race/ethnicity -- our results may not apply to all ethnic groups. Finally, our data are limited in application to CAC scores measured by electron beam computed tomography with 3 mm slice thickness and the described protocol. While CAC scores measured by the latest spiral computed tomography scanners appear to be similar to those generated by electron beam computed tomography \[[@B28]\], we cannot guarantee that our results apply to such scores. Our models should be applied to other similar cohorts for validation, and also applied in cohorts that include different racial/ethnic groups and different ways of measuring the CAC score before being used in these clinical situations. Conclusions =========== The Clinical Research Roundtable at the Institute of Medicine has identified translation of clinical research findings into improvements in medical care as the \"next scientific frontier\" \[[@B29]\]. While our analysis has some limitations, it provides methodology that will directly assist in the translation of research into practice. Our models, once validated, can be used directly by patients and clinicians to decide when it might be useful to order this potentially expensive test, and what to do with the results. Competing interests =================== MP has received speaking and consulting fees from Bayer. Authors\' contributions ======================= MJP conceived the idea for the study, performed the analysis and drafted the manuscript. JAT and MP helped design and interpret the analysis. CM provided statistical guidance and interpretation. TQC recruited the patients and collected the data. WSB provided senior guidance in all aspects. All authors reviewed and commented on multiple drafts of the manuscript and approved the final draft. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1741-7015/2/31/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional File 1 This spreadsheet is used for combining information from conventional risk factors and the coronary artery calcium score to estimate coronary heart disease risk in an individual patient. Step 1: Enter your patient\'s clinical information (the red numbers). Step 2: Choose an assumption about the coronary artery calcium score relative risks (optimistic or conservative). Step 3: Find the following results: 1) \"Pre-test\" 10-year risk of coronary heart disease (CHD) based on Framingham equations; 2) The probability of having a coronary artery calcium (CAC) score that falls within 4 standard CAC score categories; and 3) The \"post-test\" 10-year risk of CHD for each CAC score category. Step 4: Use the results to interpret a CAC score, or to decide whether or not to order a coronary artery calcium scan. If a score that would change your management is unlikely to occur, it may not be worth the money. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ Dr. Pletcher was supported by funds from the Health Resources and Services Administration, Grant D14 HP00178. The authors would like to thank Paolo Raggi and Joseph Schwartz for their helpful input. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Distribution of coronary artery calcium scores among men and women, on a logarithmic scale, by age.**Categories chosen for histograms are evenly spaced on a logarithmic scale, corresponding to Ln(CAC) scores of \<1, 1--2, 2--3, 3--4, 4--5, 5--6, 6--7, 7--8, and \>8. The first bar represents subjects with no detectable CAC, which corresponds to an undefined Ln(CAC) value. CAC -- Coronary artery calcium. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Comparison between actual and predicted CAC score distributions among a subset of the study population using three different modeling strategies.**Actual prevalence measurements were from the 58- to 62-year-old non-smoking women in our study sample with hypertension, high cholesterol level, and no diabetes (n = 127). The \"two-stage model predictions\" use the coefficients presented in Tables 2 and 3 (the full model). The Ln(CAC+1) model predictions are from a linear regression model including all conventional CHD risk factors using Ln(CAC score +1) as a continuous outcome in a one-step modeling process (coefficients not presented). The Tobit model uses the cube-root of the CAC score as a continuous outcome for linear regression analysis, but assumes that scores at or below zero have been censored (coefficients not presented). P-values refer to a X^2^test with 3 degrees of freedom comparing the expected frequencies based on each model with the observed frequencies. Lower p-values indicate a poorer model fit. CAC -- Coronary artery calcium; CHD -- Coronary heart disease; Ln -- Natural logarithm. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Characteristics of 9341 patients meeting inclusion criteria ::: Characteristic N (%) or mean ± SD ---------------------------------- -------------------- Age (years) 54 ± 10 years Women 3782 (40%) Hypertension\* 4069 (44%) High cholesterol\* 5847 (63%) Diabetes mellitus\* 807 (9%) Smoking† 3679 (39%) Framingham 10-year risk estimate  - range 1.0% -- 74%  - median, 25%--75% 11%, 7.0% -- 15% \* -- Per self report. Patients were asked whether they were under medical treatment for \"X\". † -- Current, or quit within the past six months SD -- Standard deviation ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Predictors of the presence of coronary artery calcium, in three logistic regression models ::: Modeling approach Model results ------------------------------------------------------------------------------ --------------------- --------- ------------- Predictors Odds ratio (95% CI) p-value C-statistic Age and sex only\*  - Age, per 10 years 2.83 (2.67 -- 2.99) \<0.001 0.76  - Male sex 3.60 (3.26 -- 3.96) \<0.001 All CHD risk factors\*  - Age, per 10 years 2.78 (2.62 -- 2.94) \<0.001 0.78  - Male sex 3.67 (3.31 -- 4.06) \<0.001  - Hypertension 1.51 (1.37 -- 1.66) \<0.001  - Diabetes mellitus 1.85 (1.55 -- 2.21) \<0.001  - High cholesterol 1.40 (1.27 -- 1.54) \<0.001  - Smoking 1.71 (1.56 -- 1.89) \<0.001 Estimated 10-year risk of CHD†, only\*  - 10-year risk, per 5% increase in the Framingham 10-year CHD risk estimate 1.96 (1.88 -- 2.04) \<0.001 0.74 \* -- Intercepts (on a log-odds scale) were -6.20, -6.76, and -1.44 for each model, respectively. † -- 10-year risk of CHD estimated according to Framingham equations \[16\]; for assumptions used, see Methods. CHD -- Coronary heart disease; CI -- Confidence interval. ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Predictors of the extent of coronary artery calcium, as measured by log-transformed non-zero coronary artery calcium scores, in three linear regression models. ::: Model approach Model results --------------------------------------- ----------------------- -------------------------------------------------------- ---------- --------------- Predictors Coefficients (95% CI) Corresponding percent increase in natural CAC scores\* p-values Adjusted R^2^ Age and sex only†  - Age, per 10 years 0.68 (0.63 -- 0.73) 97% (88 -- 107%) \<0.001 0.14  - Male sex 0.72 (0.61 -- 0.82) 105% (85 -- 127%) \<0.001 All CHD risk factors†  - Age, per 10 years 0.69 (0.64 -- 0.73) 99% (89 -- 109%) \<0.001 0.17  - Male sex 0.73 (0.63 -- 0.83) 108% (88 -- 130%) \<0.001  - Hypertension 0.23 (0.14 -- 0.32) 26% (14 -- 38%) \<0.001  - Diabetes mellitus 0.48 (0.33 -- 0.62) 61% (40 -- 88%) \<0.001  - High cholesterol 0.15 (0.05 -- 0.24) 16% (4.8 -- 28%) 0.004  - Smoking 0.45 (0.35 -- 0.54) 56% (42 -- 71%) \<0.001 Estimated 10-year risk of CHD‡, only†  - 10-year risk, per 5% increase 0.34 (0.31 -- 0.36) 40% (36 -- 44%) \<0.001 0.11 \* -- The percent increase in the natural (non-transformed) CAC score associated with each predictor is calculated by exponentiating the regression coefficient from the linear regression model (when the dependent variable is log-transformed), and subtracting 1. † -- Intercepts were -0.181, -0.705, and 3.17 for each model respectively. The standard deviations of the residuals were 1.682, 1.653, and 1.707. ‡ -- 10-year risk of CHD estimated according to Framingham equations \[16\]; for assumptions used, see Methods. CAC -- Coronary artery calcium; CI -- Confidence interval; CHD -- Coronary heart disease. ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Estimated prevalence of a coronary artery calcium score in each of four standard categories, depending on the Framingham estimated 10-year risk of coronary heart disease events. ::: Framingham 10-year CHD risk estimate\* Estimated prevalence of a CAC score in the given range†, % ---------------------------------------- ------------------------------------------------------------ ---- ---- ---- 2.5% 75 19 4 1 5.0% 68 23 6 2 7.5% 61 28 8 3 10.0% 52 32 11 5 12.5% 44 36 13 7 15.0% 36 38 17 9 17.5% 29 40 19 12 20.0% 22 41 22 15 22.5% 17 40 25 18 25.0% 13 39 27 22 \* -- 10-year risk of CHD events estimated according to equations derived from the Framingham study\[16\]. For assumptions, see Methods. † -- Proportions of subjects in each given CAC score category were estimated by a two step process: 1) Logistic regression to predict the presence of CAC according to estimated 10-year risk (see Table 2 for model coefficients), and 2) Linear regression to predict the extent of CAC, as measured by the natural log-transformed CAC score (see Table 3 for model coefficients). Other values required for this calculation were the logistic regression constant (-1.44), the linear regression constant (3.17), and the standard deviation of the residuals after linear regression (1.707). CAC -- Coronary artery calcium; CHD -- Coronary heart disease. ::: ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Examples of how to use the coronary artery calcium score to refine risk estimates. ::: Clinical scenario Pre-test 10-year CHD risk estimate\* CAC score category Proportion of CAC scores falling within the given category† Post-test 10-year risk estimate for each CAC score category‡ ---------------------------------------------------------- -------------------------------------- -------------------- ------------------------------------------------------------- -------------------------------------------------------------- ----- 60-year-old woman with hypertension and high cholesterol 15% 0: 0.47 9% 6% 1--100: 0.36 15% 13% 101--400: 0.12 25% 31% \>400: 0.05 34% 51% 50-year-old man without other CHD risk factors 6% 0: 0.59 4% 3% 1--100: 0.31 7% 6% 101--400: 0.07 11% 15% \>400: 0.03 16% 27% 40-year-old woman who smokes 3% 0: 0.89 2% 2% 1--100: 0.10 4% 5% 101--400: 0.01 7% 12% \>400: 0.00 10% 22% 80-year-old man with high cholesterol 26% 0: 0.05 9% 5% 1--100: 0.25 15% 10% 101--400: 0.30 26% 23% \>400: 0.40 35% 39% \* -- From published Framingham equations \[16\]. † -- These probabilities are calculated using regression equations presented in Tables 2 and 3 of this paper (full models with all predictors). See Methods for details. ‡ -- Post-test risk estimates are calculated by assuming that the pre-test 10-year CHD risk estimate represents an average of persons with different CAC scores, weighted by the probability of having a CAC score in each category. The risk in each category is calculated algebraically using relative risk estimates§ from a recent meta-analysis \[7\] (see Additional File 1). §-\"Conservative\" and \"Optimistic\" refer to assumptions made in a recent meta-analysis that attempted to quantify the value of CAC scores in predicting CHD events, independent of other CHD risk factors \[7\]. With conservative assumptions, relative risks associated with different CAC score categories were 1.7 (for CAC = 1--100), 3.0 (for CAC = 101--400) and 4.3 (for CAC\>400) compared with a CAC score of zero. With optimistic assumptions, the corresponding relative risks were 2.1, 5.4, and 10.3. CHD -- Coronary heart disease; CAC score -- Coronary artery calcification score. :::

PubMed Central

2024-06-05T03:55:47.356248

2004-8-24

PMC515365

Introduction {#s1} ============ Insertion of new sequences into nuclear DNA has a major impact on its architecture and is an important mechanism for the evolution of eukaryotic genomes. Moreover, when targeted to gene loci, these insertions can be mutagenic, and in humans this process contributes to a number of diseases ([@pbio-0020273-Deininger1]; [@pbio-0020273-Neil1]; [@pbio-0020273-Nelson1]). The frequency of the insertion events and the site of integration are therefore critical factors influencing genomic stability. In humans these two aspects have been investigated for mobile elements, including long and short interspersed elements and retroviruses ([@pbio-0020273-Li1]; [@pbio-0020273-Batzer1]), but much less is known about nuclear DNA sequences of mitochondrial origin (NUMTs), which have been found associated with diseases in humans ([@pbio-0020273-Willett-Brozick1]; [@pbio-0020273-Borensztajn1]; [@pbio-0020273-Turner1]). DNA fragments of mitochondrial origin, originating from both coding and noncoding regions, are found as sequence fossils in the nuclear genomes of various eukaryotes ([@pbio-0020273-Blanchard1]). However, de novo integrations have been recently detected in yeast and humans ([@pbio-0020273-Ricchetti1]; [@pbio-0020273-Yu1]; [@pbio-0020273-Turner1]), and insertion of NUMTs in the nuclear genome has been found to be an ongoing process in yeast ([@pbio-0020273-Ricchetti1]). We and others have previously shown that NUMTs integrate in the nuclear genome during the repair of double-strand breaks (DSBs) in yeast growing mitotically ([@pbio-0020273-Ricchetti1]; [@pbio-0020273-Yu1]). In these studies, sequences of mitochondrial origin were the main or the exclusive type of DNA able to integrate at an induced DSB. Before the sequencing of the human genome was completed, occasional reports described sequences of mitochondrial origin located in the nucleus ([@pbio-0020273-Tsuzuki1]; [@pbio-0020273-Perna1]), in one case in vivo ([@pbio-0020273-Zischler1]). More recently, sequence analysis performed on the first human genome draft revealed the presence of between 280 and 296 NUMTs ([@pbio-0020273-Mourier1]; [@pbio-0020273-Tourmen1]). According to one study, it appears that only one third of NUMTs were integrated as new sequences, whereas the remaining two thirds originated as duplications of preexisting NUMTs ([@pbio-0020273-Hazkani-Covo1]). Another report suggests that most NUMTs arose from independent insertion events ([@pbio-0020273-Bensasson1]), thereby raising questions regarding the real insertion rate of these sequences in the human genome. Moreover, it has been suggested that most NUMTs have been inserted in a primate ancestor ([@pbio-0020273-Tourmen1]; [@pbio-0020273-Bensasson1]). Thus, the rate and the effects of colonization of the human genome by DNA fragments of mitochondrial origin remain unclear, and the presence of such sequences has not been fully investigated in humans. With the availability of the human genome sequence coupled with significant discoveries on the evolution of *Homo sapiens,* experimental approaches that compare individuals within this species and its closest relative, the chimpanzee, can be undertaken ([@pbio-0020273-Chen2]). In the present study, we demonstrate the presence in vivo of NUMTs in the human and in the chimpanzee genomes, using genome-wide sequence analysis combined with direct evaluation of DNA samples of individuals. Moreover, we show a significant degree of insertion polymorphism of NUMTs in human populations. We also provide a comprehensive analysis of NUMTs that have specifically colonized the human genome, and we determine the fixation rate of these sequences in *H. sapiens.* Furthermore, we correlate these findings with observations of the insertion of NUMTs involved in human diseases. We show that human-specific NUMTs*,* unlike those in yeast, preferentially integrate in known or predicted genes and can therefore be mutagenic, thereby generating genetic alterations in humans. Results {#s2} ======= NUMTs Are Mitochondrial Sequences Residing in the Human Nuclear Genome {#s2a} ---------------------------------------------------------------------- From a blastn search on the database of H. sapiens published by the public consortium ([@pbio-0020273-Lander1]), using as query the human mitochondrial (mt) DNA sequence ([@pbio-0020273-Anderson1]), 211 NUMTs were found. We observed that approximately 93% of NUMTs represent insertions of a single DNA fragment, whereas 7% consist of multiple, unrelated mtDNA fragments, similar to the pattern frequently found in yeast ([@pbio-0020273-Ricchetti1]). We observed that NUMTs ranged in size from 47 to 14,654 bp with a sequence identity to the human mtDNA of 78%--100%. Previous analyses were based on less complete genome sequencing (84% complete for the most recent study; [@pbio-0020273-Bensasson1]), while our study was performed on a 99% complete sequencing of the euchromatic genome of *H. sapiens.* Our updated analysis (data not shown) reveals that the majority of these NUMTs correspond to those previously documented ([@pbio-0020273-Mourier1]; [@pbio-0020273-Tourmen1]). However, our study draws attention to NUMTs that are highly identical to mtDNA, and some of these sequences did not appear in earlier analyses. Indeed most of the NUMTs in our study are shorter than 100 bp (see [Table 1](#pbio-0020273-t001){ref-type="table"}), whereas former investigations focused more on longer NUMTs. Although the presence of shorter NUMTs has been reported, these have not been published ([@pbio-0020273-Tourmen1]). Combining the lowered size threshold and the more complete database, we were able to identify 23 new NUMTs (labeled with an asterisk in [Table 1](#pbio-0020273-t001){ref-type="table"}). Most of the new sequences are of the highest interest for our studies of the acquisition of NUMTs by H. sapiens and of insertion polymorphism in humans (see below). ::: {#pbio-0020273-t001 .table-wrap} Table 1 ::: {.caption} ###### PCR Amplification and Sequence Analysis of NUMTs from Humans and Chimpanzees :::  Upper part, scheme of the PCR strategies; thin line, chromosomal DNA; thick line, NUMT; arrowed lines, PCR primers PCR amplifications were done either with primers A + B or A + C. In NUMT code names, the first number represents the chromosome number, and the second the NUMT size; an asterisk to the right indicates NUMTs described in this paper. "Percent" indicates the percentage of identity of the NUMT to the human mtDNA as scored by blastn. Missing values in the "no NUMT" column indicate that the PCR was done with the strategy A + B, rather than A + C (see upper part of table). "Amplified in Chimp" indicates whether the NUMT did amplify (+) or not (−) in the chimpanzee genome. Where BLAST output search results from a database in April 2003 did not fit with our PCR and sequence analysis, we chose as indicated in footnotes "a" and "c." ^a^January 2003 ^b^Strategy A + B ^c^July 2001 BLAST results, which were consistent with our sequencing ^d^These two lines indicate the 5′ (up) and the 3′ (dw) portion of the same NUMT ^e^This NUMT is present also on Chromosome 9; separate analysis of these two NUMTs was not possible by PCR and sequencing (mixed products) ^f^NUMT is specific to chromosome Y; it is absent from the eight-female sample ^g^This is the same as "f," but an additional locus without NUMT is present on the X chromosome ^h^This NUMT, described previously (Zischler et al. 1995), was renamed here ND, no amplification was detected More information on these NUMTs is available in Table S1 ::: To determine whether sequences of mitochondrial origin were actually integrated in the human nuclear genome, and were not a result of contamination of DNA library preparations, we selected 42 NUMTs for analysis in human samples. Our choice included the 36 NUMTs with the highest identity (91% to 100%) to the mtDNA, one NUMT having the longest stretch of DNA with high identity (88%), one NUMT corresponding to the highly variable region of the mtDNA (D-loop) ([@pbio-0020273-Cann1]), and four NUMTs randomly chosen with identity from 79% to 90% ([Table 1](#pbio-0020273-t001){ref-type="table"}). Each of these NUMTs was amplified by PCR from DNA obtained from 21 human donors (eight females, 13 males) representing different ethnical groups. Our pool consisted of ten Caucasians, seven Africans (including four Pygmies), two Japanese, and two Chinese ([Table 2](#pbio-0020273-t002){ref-type="table"}). To amplify chromosomal NUMTs and avoid amplification of the mt chromosome, we used a primer located in the upstream flanking region, in combination with a primer located either in the 3′ region of the NUMT or in the downstream flanking region (see upper part of [Table 1](#pbio-0020273-t001){ref-type="table"}, primers A + B or A + C, respectively). In the former case, PCR amplification served as a supplementary control for bona fide mtDNA integration at the locus, while in the latter, PCR amplification was followed by sequencing to assay for the presence of the NUMT. Forty-one out of 42 loci tested amplified a fragment of the expected length ([Table 1](#pbio-0020273-t001){ref-type="table"}), while one locus (14--1023 \[Chromosome 14; size = 1023 bp\]) amplified a fragment not containing the NUMT in all individuals tested. Eighteen loci were analyzed further in one or more individuals by sequencing the amplified fragment to verify whether these DNA fragments included the sequence of mt origin (see [Table 1](#pbio-0020273-t001){ref-type="table"}). The expected sequence was indeed present in all cases tested (except at polymorphic loci, described later, and at insertions in the Y chromosome, present only in males). ::: {#pbio-0020273-t002 .table-wrap} Table 2 ::: {.caption} ###### Insertion Polymorphism of NUMTs Displaying Distinct Lineage Characteristics in Humans :::  For each NUMT is indicated the PCR amplification containing (+/+) or not containing (−/−) NUMT; (+/−) indicates it is heterozygous. NUMT 11-541 was identified previously (Zischler et al. 1995) ^a^Locus has been sequenced ^b^Both allelic forms have been sequenced ::: In summary, results from two amplification strategies and from sequencing demonstrated that these NUMTs were indeed present at the expected chromosomal location and that they are bona fide mt sequences residing in the human nuclear genome. Insertion Polymorphism of NUMTs in Human Populations {#s2b} ---------------------------------------------------- The colonization of human populations by various NUMTs revealed striking disparities. Thirty-five NUMTs were present in homozygous form in all individuals tested (see [Table 1](#pbio-0020273-t001){ref-type="table"}). Interestingly, NUMTs 1-74, 2-53, 12-89, and 18-192 were present in only a few individuals either as homozygous or heterozygous loci (see [Figure 1](#pbio-0020273-g001){ref-type="fig"}; [Tables 1](#pbio-0020273-t001){ref-type="table"} and [2](#pbio-0020273-t002){ref-type="table"}). A more limited heterogeneity was observed for NUMTs 13-75 and 2-132, where only two and one individual, respectively, were heterozygous. In total, six out of 41 NUMTs showed insertion polymorphism. Integration of NUMTs was further confirmed by sequencing both positive and negative samples (see [Table 2](#pbio-0020273-t002){ref-type="table"} for the samples tested). The sequences of these six NUMTs are shown in [Figure 2](#pbio-0020273-g002){ref-type="fig"}. NUMT 11-541, whose insertion polymorphism was previously described ([@pbio-0020273-Zischler1]; [@pbio-0020273-Thomas1]), was reanalyzed here ([Tables 1](#pbio-0020273-t001){ref-type="table"} and [2](#pbio-0020273-t002){ref-type="table"}). By comparing the flanking sequences of individuals carrying a NUMT with those of individuals who have no NUMT, it is possible to identify the junction sites and to also clarify the mechanism by which NUMTs were inserted. This analysis was done for NUMTs 2-132 and 18-192, in which the junction sites ([Figure 2](#pbio-0020273-g002){ref-type="fig"}) show microhomology between the invading NUMT and the chromosomal end, and sometimes addition of a few nucleotides. Both the presence of microhomology and the addition of short sequences also accompanied the insertion of NUMTs in the yeast genome ([@pbio-0020273-Ricchetti1]), and they are hallmarks of the DSB repair mechanism non-homologous end-joining (NHEJ) ([@pbio-0020273-Critchlow1]). This suggests that NHEJ may also account for the insertion of NUMTs in humans. ::: {#pbio-0020273-g001 .fig} Figure 1 ::: {.caption} ###### Polymorphism of NUMTs 18-192, 1-74, and 2-53 The polymorphism of NUMTs 18-192, 1-74, and 2-53 as revealed by PCR amplification and electrophoresis of the products on 2% agarose gels. For each locus, the upper arrow indicates the fragment that contains the NUMT, and the lower arrow indicates the fragment that does not contain the NUMT. The individual tested is indicated above. The (+/+) are homozygous positive, (+/−) are heterozygotes, and (−/−) are homozygous negative. :::  ::: ::: {#pbio-0020273-g002 .fig} Figure 2 ::: {.caption} ###### Sequence Insertion Polymorphism of Six NUMTs Sequence of NUMTs 1-74, 2-53, 2-132, 12-89,13-75 and 18-192 are indicated in lower case and the flanking sequences in capital letters. Underlined letters represent nucleotides homologous to both the mt and the chromosomal sequences (microhomology). Bold and italicized letters correspond to nucleotide additions, following the NUMTs insertion, which are absent from the −/− individuals. The individuals sequenced are indicated in [Table 2](#pbio-0020273-t002){ref-type="table"}. In all cases the sequence corresponded to the one available on the human genome public Web sites. Boxes represent exon sequences. In 12-89, the exon sequence would extend till the stop codon (taa). :::  ::: Interestingly, one or more of these six NUMTs were detected among individuals within each ethnic group, indicating that their insertion in the nuclear genome occurred soon after the origin of modern humans and that they represent the most recent integrations of our studied cases. Despite the limited sampling size (42 alleles), the frequency of alleles carrying the insertion varies greatly according to the NUMT (calculated from [Table 2](#pbio-0020273-t002){ref-type="table"}): 98%, 95%, 48%, 29%, and 21% for NUMTs 2-132, 13-75, 2-53, 18-192, and both 12-89 and 1-74, respectively. Moreover, allele frequencies among different ethnic groups are not equal. For example, NUMTs 1-74 and 18-192 are poorly represented among Caucasians and Asians and are more frequent in non-pygmy Africans. NUMT 12-89, unlike other NUMTs, is poorly represented in non-pygmy Africans. As a result, each NUMT presents a unique population fingerprint. Acquisition of NUMTs by H. sapiens {#s2c} ---------------------------------- To evaluate which NUMTs are specific to humans, we amplified by PCR the 42 loci described above on chimpanzee DNA *(Pan troglodytes).* For each locus, one to three chimpanzee individuals were analyzed. Forty-two out of 42 primer pairs successfully amplified the target site also in chimpanzees because of the high sequence identity of the two genomes (average 98.7%) ([@pbio-0020273-Chen2]). Only the regions flanking the previously described NUMT 11-541, which is considered separately in our investigation, did not amplify in chimpanzees. Locus 14-1023, where no NUMT was identified in humans, also showed no insertion in the chimpanzee. Surprisingly, only 14 loci contained the NUMT (see [Table 1](#pbio-0020273-t001){ref-type="table"}). All of these NUMTs were also found in all human individuals tested, indicating that they were present in the common ancestor of human and chimpanzee. On the contrary, 27 NUMTs absent from the chimpanzee genome represent recent acquisitions in *H. sapiens.* The distribution of these NUMTs in the human chromosomes is shown in [Figure 3](#pbio-0020273-g003){ref-type="fig"}. All NUMTs whose presence was not found in all human individuals fell in this category. From our data, 24 out of 27 NUMTs specific to humans had greater than 94% of sequence identity to the human mtDNA, and three out of 27 NUMTs had sequence identity of 91%--92%. This higher level of identity is consistent with the idea that NUMTs specific to humans are recent insertions (for the calculation of the insertion time of NUMTs, see [Materials and Methods](#s4){ref-type="sec"}). Similar values of identity to the mtDNA were also found for the recent insertions of mt sequences in the yeast genome ([@pbio-0020273-Ricchetti1]). Seven out of 14 NUMTs present both in humans and in chimpanzees have lower levels of identity to the mtDNA (between 79% and 90%), as expected; however, the remaining seven NUMTs have 94%--96% identity to the mtDNA (see [Table 1](#pbio-0020273-t001){ref-type="table"}), indicating that the level of identity per se is not a rigorous criterion for human specificity. Interestingly, most of NUMTs present only in humans are short sequences, and about half of them are less than or equal to 100 bp. In summary, out of 211 NUMTs recognizable in the human genome, at least 27 (or 28, if we also include NUMT 11-541) were specific to humans, and we do not expect this value to increase significantly because 99% of the euchromatic genome of H. sapiens was analyzed, and we assume that most NUMTs with low identity to the mtDNA (less than or equal to 90%) are unlikely to be human-specific. This results in an average of one NUMT integration in the germline for each 180,000 y, in the last 4--6 million years (Myr). Interestingly, one fourth of these NUMTs (6/27, or 7/28 if we include NUMT 11-541) show insertion polymorphism (see [Table 1](#pbio-0020273-t001){ref-type="table"} and above), indicating that they have occurred in more recent times. ::: {#pbio-0020273-g003 .fig} Figure 3 ::: {.caption} ###### Distribution of Human-Specific NUMTs in Chromosomes A scale representation of the human chromosomes. The location of human-specific NUMTs is indicated with a red arrow. A green arrow indicates the position of NUMTs showing insertion polymorphism in humans, and a blue arrow indicates a previously described NUMT (11-541). :::  ::: High Frequency of Human-Specific NUMTs in Chromosomes 18 and Y {#s2d} -------------------------------------------------------------- The distribution of human-specific NUMTs in human chromosomes is not proportional either to the chromosome size or to the total number of NUMTs present in the chromosome ([Figure 4](#pbio-0020273-g004){ref-type="fig"}). In Chromosomes 13 and 20, human-specific NUMTs represent 37% and 50%, respectively, of the NUMT insertions detected in the chromosome. More surprisingly, in Chromosomes 18 and Y there is a proportionally higher number of human-specific NUMTs (2/3 present in each chromosome; see [Figure 4](#pbio-0020273-g004){ref-type="fig"}), whereas at the genome-wide level about 13% (27/211) of NUMTs are specific to humans. The high number of human-specific NUMTs on the Y chromosome is particularly intriguing since this chromosome is 4-fold less present in the human population than the other chromosomes (it is the only haploid chromosome, present only in males). Additionally, the NUMT value for the Y chromosome may be an underestimate, since its large heterochromatic portion has not yet been sequenced. Since no more NUMTs are available to increase sampling size, we cannot formally distinguish between a founder effect and an increased insertion rate in Chromosomes 18 and Y above that in other chromosomes during recent human evolution. ::: {#pbio-0020273-g004 .fig} Figure 4 ::: {.caption} ###### Human-Specific NUMTs in Human Chromosomes For each human chromosome, indicated on the x-axis, the number of NUMTs (y-axis, on the left) common to human and chimpanzee (white columns) and specific to humans (black columns) are shown. An open circle indicates the chromosome size in millions of base pairs (Mbp; y-axis on the right). :::  ::: NUMTs Mainly Integrate in Known or Predicted Genes {#s2e} -------------------------------------------------- The integration of NUMTs in the human genome takes place, surprisingly, mainly in known or predicted coding or regulatory regions. As indicated in [Table 3](#pbio-0020273-t003){ref-type="table"} and [Figure 5](#pbio-0020273-g005){ref-type="fig"}, out of 28 human-specific NUMTs, 22 integrate in a known or predicted intron, one in an exon, and one in a promoter region. Only 4/28 NUMTs are in intergenic regions. This is also the case for older NUMTs, common to humans and chimpanzees, where 10/14 NUMTs are inserted in intron regions, and 4/14 in intergenic regions. All seven of the most recent integrations, those displaying insertion polymorphism, were found in exons or introns. In summary, about 80% of NUMTs are inserted in known or predicted introns/exons, which together should cover about 25% of the human genome ([@pbio-0020273-Venter1]). Analysis of the position of NUMTs inside introns reveals that one NUMT, 12-89, was inserted exactly at the splice-donor site of the last predicted intron ([Figures 2](#pbio-0020273-g002){ref-type="fig"} and [6](#pbio-0020273-g006){ref-type="fig"}). The other 31 NUMTs appear to be randomly integrated within the introns, although in two cases the insertion generates one or two new exons in the predicted proteins (NUMT 17-653 and NUMT 5-8781, respectively); see [Figure 6](#pbio-0020273-g006){ref-type="fig"}. Moreover, NUMT 1-74, which is the only NUMT found inserted within an exon, splits the last exon of the gene *Q8N7L5* into two, and the NUMT itself becomes a new intron ([Figures 2](#pbio-0020273-g002){ref-type="fig"} and [6](#pbio-0020273-g006){ref-type="fig"}). Thus, for at least four NUMTs, out of the 33 that are inserted in genes, the exon/intron pattern looked modified after integration of the sequence of mt origin, essentially by a change in the number of exons. We would expect these to be the most likely candidates to perturb gene function. ::: {#pbio-0020273-g005 .fig} Figure 5 ::: {.caption} ###### Insertion Sites of NUMTs in the Human Genome Histogram of the insertion sites of NUMTs in the human genome. Only NUMTs tested in human and in chimpanzee samples are shown. This includes the 27 NUMTs specific to humans and absent from chimpanzees (21 present in all individuals tested and 6 with insertion polymorphism in humans), one additional NUMT with insertion polymorphism, previously described, see text, and 14 NUMTs common to human and chimpanzee, out of 183 found by BLAST search. Colors of the blocks indicate the different target sites. For details see [Table 3](#pbio-0020273-t003){ref-type="table"}. :::  ::: ::: {#pbio-0020273-g006 .fig} Figure 6 ::: {.caption} ###### Scheme Representing Some NUMT Insertions in Genes Four known or predicted genes, found in loci with NUMT insertion in humans, have been schematically represented either in the absence (A) or in the presence (B) of the insertion. Boxes represent exons, and thick lines represent introns. Red boxes and lines indicate the sequence corresponding to the NUMT, which has been identified for each case. A dotted line in (A) indicates that, in the absence of insertion, the exon/intron pattern was not identified by gene identification programs. Representation not to scale. :::  ::: ::: {#pbio-0020273-t003 .table-wrap} Table 3 ::: {.caption} ###### Insertion Sites of NUMTs in the Human Genome :::  "Gene Reference" indicates the targeted gene, or the transcript code---hypothetical protein when based on prediction programs. Swiss-Prot indicates Swiss-Prot/TrEMBL. Data were obtained using <http://www.ensembl.org/Homo_sapiens>, <http://us.expasy.org/sprot>, <http://genome.ucsc.edu/cgi-bin/hgBlat>, <http://genes.mit.edu/GENSCAN.html>, and related sites. Detailed coordinates of the predicted genes are shown in Table S2. The last two columns indicate the organ(s) where the corresponding transcript was found and the phenotype associated with mutations in the gene (references in <http://us.expasy.org/sprot> and in Table S2) CL, colon; FB, fetal brain; PL, placenta; SM, skeletal muscle ::: Twenty-one out of 33 NUMTs are inserted in predicted genes, and the other 12 in known genes, including a heart-specific serine protease and a thiamine transporter ([Table 3](#pbio-0020273-t003){ref-type="table"}). Interestingly, three of the genes targeted by NUMTs with insertion polymorphism in humans are *MADH2,* a tumor-suppressor gene, mutated in colorectal carcinoma ([@pbio-0020273-Eppert1]; NUMT 18-192); a gene coding for a homologue of the thrombospondin gene (an angiogenesis inhibitor that retards tumor growth; [@pbio-0020273-Bogdanov1]; NUMT 1-74); and a mt ribosomal precursor (NUMT 13-75; [Table 3](#pbio-0020273-t001){ref-type="table"}). For NUMT 1-74, which is inserted in an exon, and for NUMTs 18-192 and 13-75, both inserted in an intron, it is not known if individuals carrying the insertion are mutated for these genes. These findings suggest that in humans, the insertion of NUMTs is elevated in gene-containing regions of the genome. Insertions in such regions are potentially mutagenic. Interestingly, at least two cases of NUMT insertions---one in an exon, the other in an intron region---associated with diseases have been recently reported in humans ([@pbio-0020273-Borensztajn1]; [@pbio-0020273-Turner1]). Discussion {#s3} ========== Integration of mt genes into the nuclear genome is a physiologically important process that contributes to the origin and evolution of the eukaryotic cell ([@pbio-0020273-Margulis1]), and the transfer of entire genes from mitochondria to the nucleus appears to be continually active in some plants ([@pbio-0020273-Knoop1]). Although the transfer of entire genes seems to have ended in animals, DNA fragments of mitochondrial origin continue to integrate in the nuclear genome. In the present study, we examined the extent and the consequences of this process in humans and in chimpanzees. NUMTs Are Present in the Human Genome and Display Insertion Polymorphism {#s3a} ------------------------------------------------------------------------ The direct investigation of samples of different individuals provided in this study clearly demonstrates the presence of DNA fragments of mitochondrial origin in the nuclear genome of humans, as previously suggested by the analysis of the databases ([@pbio-0020273-Mourier1]; [@pbio-0020273-Woischnik1]) and by a few tests in cells ([@pbio-0020273-Tourmen1]). Although the presence of a single NUMT was previously shown in vivo ([@pbio-0020273-Zischler1]), our study provides direct evidence that the large colonization of the human genome by NUMTs detected in silico, an outcome of the sequencing of the entire human genome, corresponds to the in vivo situation. We can thus exclude that, at least for the tested loci, NUMTs result from contaminations during the sequencing process, a situation that could not be previously ruled out formally (see comments in [@pbio-0020273-Venter1], and in [@pbio-0020273-Mourier1]). Our investigation of the distribution of NUMTs in human populations, which includes some of the less divergent among the 211 NUMTs, reveals that in most cases NUMTs are present in all the individuals tested, and therefore these sequences have colonized the nuclear genome of all major human populations. However, six NUMTs described here and one described previously ([@pbio-0020273-Zischler1]; [@pbio-0020273-Thomas1]) are present only in some individuals. These seven NUMTs, not fixed within the human population, must have been recently acquired. Since they are present in individuals within each ethnic group, their insertion most probably occurred after the origin of modern humans and before the emergence of distinct ethnic groups (see also below). We did not detect NUMTs restricted to only one or a few ethnic groups. Furthermore, these seven NUMTs appear to have colonized the genome of human populations at different rates. Indeed, the frequency of alleles carrying the insertion varies greatly according to the NUMT (from 21% to 98% in our samples). This suggests that each sequence exhibits different colonization dynamics, involving the time of insertion and/or the expansion rate of the founder individual(s). Moreover, the distribution of each NUMT is unequal between ethnic groups, and a larger analysis of human populations will be necessary to reveal distinct population patterns and to perform phylogenetic studies. Nevertheless, most individuals tested had a unique combination of these seven NUMTs, suggesting that the individual pattern of NUMT insertion polymorphism can be useful as genetic fingerprints for familial pedigree studies. We expect that other NUMTs displaying such polymorphism will be discovered when larger population samples are examined. The locus 14-1023, which contains a NUMT according to the genome sequence, does not amplify a NUMT-containing fragment in all of 21 individuals. If this does not represent a sequencing artifact, it may be a further example of insertion polymorphism. Furthermore, we propose that the number of NUMT insertion polymorphisms is currently underestimated, since sequencing of the human genome was done only on a limited number of individuals ([@pbio-0020273-Lander1]). Several independent markers are needed to accurately retrace the phylogeny of human populations ([@pbio-0020273-Rosenberg1]), and insertion polymorphism is particularly interesting because of the low fixation rate and lack of reversion, unlike markers such as single nucleotide polymorphisms. Each of the six insertion polymorphisms described here is a rare event, and if a neutral genetic marker, it provides an important tool for tracing human dispersal. Insertion Rate of NUMTs in H. sapiens {#s3b} ------------------------------------- An important question concerning the integration of DNA sequences in the nuclear genome is their rate of colonization. For exogenous sequences like NUMTs, this has not been investigated in vivo. To determine the extent of colonization of a given genome, it is necessary to compare the insertions within this genome with those of a closely related species. To date, a comprehensive analysis of the presence of NUMTs has been done for several complete nuclear genomes ([@pbio-0020273-Ricchetti1]; [@pbio-0020273-Mourier1]; [@pbio-0020273-Tourmen1]; [@pbio-0020273-Richly1]), but the colonization rate of these sequences was not investigated, because of the absence of data in closely related species. In the case of *H. sapiens,* although a proportion of NUMTs present in its genome appears as ancient insertions (see Results; [@pbio-0020273-Tourmen1]; [@pbio-0020273-Bensasson1]), it is not clear how many NUMTs were inserted in primate ancestors and how many specifically colonized the human genome. Chimpanzee *(P. troglodytes),* a species closely related to humans and whose evolutionary relationship with humans has been widely investigated, is an ideal candidate for a comparative analysis. Moreover, the high level of identity (more than 98%) between the two species ([@pbio-0020273-Chen2]; [@pbio-0020273-Fujiyama1]) allows the investigation of the respective genomes using molecular tools. No previous analysis in vivo showed the presence of NUMTs in the chimpanzee genome. By direct PCR amplification and sequencing of chimpanzee samples, we found that out of 41 NUMTs, 14 are also integrated in the chimpanzee genome (see above) and were therefore present in the common ancestor of humans and chimpanzees. However 27 NUMTs are absent from the chimpanzee genome, and are therefore recent acquisitions in *H. sapiens.* For NUMTs fixed in the human genome (not displaying insertion polymorphism), we do not expect this value to increase significantly, since our analysis was made on essentially the entire human genome. Among all NUMTs detected in the human genome, only about 13% (27 out of 211, or 28 if we include NUMT 11-541) are specific to H. sapiens and have integrated in the human genome in the last 4--6 Myr, after the split of the two species from their common ancestor ([@pbio-0020273-Chen1]). This corresponds to an average of one integration in the germline each 180,000 y, or 5.4 insertions per Myr, a value remarkably close to that estimated by a phylogenetic analysis (5.1 insertions per Myr), which assumed a uniform insertion rate over time ([@pbio-0020273-Bensasson1]). However, the rate of integration of the more recent NUMTs may not be consistent with a constant insertion rate. Indeed, out of 28 human-specific NUMTs, seven display insertion polymorphism and are present in all populations; thus they must have appeared early after the origin of modern humans. The date of this origin is still uncertain. If we assume that NUMTs with insertion polymorphism have been inserted at the same rate as the older NUMTs (fixed in the population), then they must have integrated in the genome of the human ancestor not earlier than 1.4 Myr ago, long before the origin of modern humans, and after the spread of Homo erectus out of Africa (1.7 Myr ago). Living humans would still be polymorphic for these NUMTs, as a result of interbreeding of the nonmodern human populations with modern humans ([@pbio-0020273-Templeton1]). On the contrary, if we assume that these insertions are more recent, as suggested by the poor allelic presence of most of them in present populations, then they must have appeared shortly before the expansion of modern humans, estimated at about 100,000 y ago ([@pbio-0020273-Templeton1]). In this case, their integration rate would be significantly higher than that of NUMTs inserted in the human genome in the previous 4--6 Myr. If the latter is the case, this strikingly high difference in the rate of colonization of the human genome may be due to a founder effect (i.e., the sporadic expansion of individuals carrying specific NUMTs) or, alternatively, to a genuine increase in the integration rate in modern humans. A third possibility is that there is no increase in the insertion rate in modern humans and that the number of "recent" NUMTs is overestimated because they include unfixed NUMTs that are destined to be lost eventually. In this latter case, we would need to assume that at least some of the NUMTs with insertion polymorphism are not neutral and are associated with a selectable phenotype. Although this may not be true for NUMTs 2-132 and 13-75, present in more than 95% of alleles tested, we cannot exclude that the low allelic presence (21%) of NUMTs 1-74 and 12-89 (both inserted in the context of an exon) is the result of the progressive counterselection of a defective phenotype; this would have implications for the mutagenic potential of NUMTs (see below). Compared to 28 NUMT insertions in the human nuclear genome in the last 4--6 Myr, it has been calculated that about 5,000 new insertion events of Alu repeats have occurred in the human genome in the same timescale (reviewed in [@pbio-0020273-Batzer1]). This large difference may be due to the fact that Alu elements are endogenous sequences that can be amplified by reverse transcriptase provided by long interspersed elements and inserted in the genome using L1 endonuclease ([@pbio-0020273-Batzer1]), whereas the integration of NUMTs depends only on the availability of DSBs and of the repair machinery ([@pbio-0020273-Ricchetti1]; [@pbio-0020273-Yu1]). This suggests that retrotranscription/integration mechanisms increase the insertion efficiency of DNA sequences by two orders of magnitude. Alternatively, the limited number of NUMTs in the human genome may result from the selection process, if NUMTs preferentially integrate in coding regions (see below). Consequences of the Preferential Integration of NUMTs in Genes {#s3c} -------------------------------------------------------------- Contrary to previous findings, which indicated that NUMTs were inserted mostly outside annotated genes ([@pbio-0020273-Woischnik1]), we find that NUMTs preferentially integrate in known or predicted genes. The availability of a more powerful database analysis on genome Web sites and the resulting increase in the number of potential new genes may explain this different evaluation. Unlike previous analyses, we investigated more recent insertions, frequently characterized by short sequences, which may have been missed in earlier studies. Moreover, we find that all of the most recent integrations, namely NUMTs with insertion polymorphisms, are integrated in genes. In cases where it was possible to identify the gene, its transcript was detected in one or more tissues ([Table 3](#pbio-0020273-t001){ref-type="table"}). Among the targeted genes we found *MADH2,* a transcriptional modulator with tumor-suppressor properties ([@pbio-0020273-Eppert1]), and a gene involved in microvessel development, and in both cases the NUMTwas present only in some individuals. In these cases it is not known whether the insertion has affected the function of the gene. Recent findings suggest that transcription promotes DNA breaks ([@pbio-0020273-Gonzalez-Barrera1]). Insertion of NUMTs is, at least in yeast, dependent on DSBs ([@pbio-0020273-Ricchetti1]), and in humans it is frequently associated with a hallmark of NHEJ, a DSB repair mechanism (our study). It is therefore possible that highly transcribed genes, perhaps carrying DSBs, are the preferential targets for the insertion of NUMTs. Only eight out of 41 NUMTs were found in intergenic regions, which should represent 75% of the human genome ([@pbio-0020273-Venter1]). The Genscan program, which detected several insertion targets in our analysis, identifies around 20% more genes than previous estimates ([@pbio-0020273-Das1]), but this does not significantly change the proportion of the genome that is noncoding. Approximately 80% of NUMTs are in coding regions, and we consider this to be a statistically significant event. Interestingly, this was not the case for yeast, where NUMTs integrated with 41-fold preference in intergenic regions ([@pbio-0020273-Ricchetti1]). The intronless structure of the yeast genome may explain this difference, since NUMTs inserted in genes would essentially target exons in yeast and would be selected against if deleterious. In humans, the high content of introns would buffer most of the mutagenic potential of these insertions. Nevertheless, we expect that a fraction of insertions would be harmful also in humans. Although most of the analyzed NUMTs are internal to introns, in at least three cases the insertion modified the exon/intron pattern, and this may be mutagenic. Our analysis indeed confirms the recent finding that two NUMTs*,* occasionally found as new insertions in the human genome, and associated with diseases in humans, are inserted in genes, either in an exon or at the junction between exons and introns ([@pbio-0020273-Borensztajn1]; [@pbio-0020273-Turner1]). Therefore, it is likely that future insertion events in the human genome would also preferentially target genes. An intriguing finding is that NUMT 12-89 is located exactly at the splice-donor site of the predicted intron. Insertion in a splice-related site was found also in human factor VII gene, where a 251-bp NUMT integrated a splice-acceptor site in a patient with severe plasma factor VII deficiency ([@pbio-0020273-Borensztajn1]). Taken together, these results account for two insertions at intron-splice sites out of 45 NUMT insertion sites analyzed (42 NUMTs in our study and present in human populations and three more NUMTs found in one or more individuals and correlated with a disease; [@pbio-0020273-Willett-Brozick1]; [@pbio-0020273-Borensztajn1]; [@pbio-0020273-Turner1]). The limited sampling size does not permit us to determine if these finding are significant, although it is tempting to speculate that splice sites can be favored targets for the insertion of NUMTs. Is the rate of de novo insertions in the human germline limited to one each 180,000 y, or even ten times higher? The number of NUMTs detected as very recent insertions in one or a few individuals suggest that the insertion rate of these sequences in humans is currently dramatically underestimated ([@pbio-0020273-Willett-Brozick1]; [@pbio-0020273-Borensztajn1]; [@pbio-0020273-Turner1]). Three new insertions of NUMTs have been found in living individuals, occasionally detected because of the search for the cause of a mutated phenotype. It seems reasonable to predict that a wider search would reveal many more NUMTs present in single or in small groups of individuals. Since NUMTs preferentially target genes, a fraction of these NUMTs could also be connected with diseases. Moreover, one expects that harmful insertions, whose probability increases as genes become preferential targets, would be subject to negative selection and thus removed from the gene pool. Thus the low fixation rate of NUMTs in the human genome may be a direct consequence of their preference for insertion in genes. NUMTs specific to the genome of H. sapiens and widespread in major human populations may represent only a small fraction of insertions that have occurred continually in the human genome. We propose therefore that the insertion of NUMTs, previously considered as functionless ([@pbio-0020273-Perna1]; [@pbio-0020273-Hazkani-Covo1]), at best an evolutionarily important but essentially harmless process, is a potentially mutagenic process, challenging the functional integrity of the human genome. Remarkably, the integration of NUMTs in the nuclear genome can be accelerated under increased induction of DSBs. In the yeast nuclear genome, where only 34 NUMTs were detected, new NUMTs are integrated at an induced DSB site with a high frequency (10^−3^−10^−4^ per repair event; [@pbio-0020273-Ricchetti1]). In keeping with this notion, a de novo insertion was reported on Chromosome 7 for a patient conceived during the Chernobyl nuclear meltdown ([@pbio-0020273-Turner1]). By analogy to our previous findings in yeast ([@pbio-0020273-Ricchetti1]), it is possible that this novel insertion is the consequence of a de novo DSB in the chromosome resulting from radiation exposure. Consistent with this view, a NUMT has been found inserted at the breakpoint junction of a familial constitutional reciprocal translocation, also associated with the occurrence of a DSB ([@pbio-0020273-Willett-Brozick1]). The fixation rate and the insertion strategy used by NUMTs are probably the prototype for the integration of exogenous sequences in the nuclear genome. Like NUMTs, sequences lacking specific amplification and integration mechanisms would rely on occasional DSBs to integrate in chromosomes. Coding or transcribed sequences could represent the preferred insertion target for these sequences as well. This strategy is in sharp contrast with the integration procedure of retrotranscribed elements, which have successfully colonized the human genome and only rarely target coding regions ([@pbio-0020273-Lander1]). In conclusion, we provide direct evidence that NUMTs are present in the human and in the chimpanzee genomes and that the insertion polymorphisms of six NUMTs reveal new markers for the study of human population genetics. Further, our in vivo analysis reveals that on average one new NUMT is fixed in the human genome each 180,000 y, although during the expansion of modern humans the fixation rate of NUMTs may have increased. The frequency of insertion of NUMTs may represent the genuine fixation rate of exogenous sequences colonizing the human nuclear genome. Strikingly, NUMTs preferentially integrate in introns and in exons, and they are thus potentially mutagenic, and novel NUMT integrations have been shown to be associated with diseases in humans. The recent case of de novo insertion of a NUMT following the Chernobyl accident, if not coincidental, provides a compelling example of how environmental insults can drive NUMTs to colonize the nuclear genome and induce genetic dysfunctions in humans. Materials and Methods {#s4} ===================== {#s4a} ### BLAST search {#s4a1} The human mtDNA sequence ([@pbio-0020273-Anderson1]) was compared to the "Homo sapiens genomic contig sequences" database version of April 11, 2003, using the National Center for Biotechnology Information (NCBI, Bethesda, Maryland, United States) "BLAST the Human Genome" server (<http://www.ncbi.nlm.nih.gov/genome/seq/page.cgi?F=HsBlast.html&&ORG=Hs>). The blastn program was used with default parameters on April 24, 2003. Only output parameters were changed to 1,000 descriptive lines and to 1,000 segment alignments. In a few cases (see [Table 1](#pbio-0020273-t001){ref-type="table"}) results of previous searches (July 2001 and January 2003) were also used. BLAST output results were saved locally in a text format and parsed using the readblastn script (see [@pbio-0020273-Tekaia1]), so that results were presented in a table format, including the query sequence, its size, the hit sequence, its size, the blastn E-value, the percent identity, the percent similarity, the matching segment size, and its coordinates on the query sequence as well as on the hit sequence. Only scores less than or equal to 10^−15^ have been selected. Each selected sequence was further aligned with the mtDNA sequence using <http://www.ncbi.nlm.nih.gov/blast/b12seq/b12.html>. ### Amplification and sequencing of NUMTs from human and chimpanzee cells {#s4a2} PCR amplification was performed on lysed cells originating from the buccal mucosa of healthy volunteers. Appropriate informed consent was obtained from human subjects. Pygmy samples (two Biakas and two Mbuti pygmies) were obtained as purified DNA from Coriell Institute (Camden, New Jersey, United States). Purified chimpanzee DNA, obtained either from tissues or from fecal material, was a kind gift from J.-P. Vartanian at the Pasteur Institute (Paris, France). For both PCR strategies described in the text, primer sequences are available upon request. Cell lysis was performed by incubating fresh cells overnight at 55 °C in a Tris-EDTA buffer (pH 8.5) in the presence of 200 μg/ml of proteinase K. PCR amplification was performed with 30 cycles of denaturation (1′ at 94 °C), annealing (1′ at 68 °C), and DNA synthesis (3′ at 72 °C) using Invitrogen (Carlsbad, California, United States) *Taq* polymerase. In heterozygous samples, a specific stochiometry of the two bands was found for each couple of primers used. Amplified NUMTs have been sequenced by specialized commercial services, using PCR amplification bands purified by gel extraction. ### Calculation of the insertion time of NUMTs {#s4a3} The age of insertion of NUMTs was estimated using, as reference, the sequence divergence of the NUMT from the mtDNA. We assumed that the NUMT, when inserted into the nuclear genome was identical to the corresponding mt sequence. We also assumed that, once inserted into the nuclear genome, the NUMT mutated at the same rate as the nuclear genome, μN, which corresponds, for noncoding sequences, to 2.5 × 10^−8^ mutations per nucleotide per generation, or 1.25 × 10^−9^ mutations per nucleotide per year, assuming a generation time of 20 y ([@pbio-0020273-Nachman1]). By comparison, the original sequence remaining in DNA is assumed to have undergone mutation at the rate, μM, of 1.7 × 10^−8^ substitutions per nucleotide per year, excluding the D-loop ([@pbio-0020273-Ingman1]). Thus, from the date of insertion (in Myr from the present) the sequence divergence between the NUMT and the cognate mitochondrial sequence is expected to be nearly the sum of mutations accumulated in each compartment (the possibility of compensation by two identical mutations is negligible given the limited divergence). It follows that the date of insertion, *i,* is given by *i* = *d*/(μM + μN), where *d* is the frequency of sequence divergence between the NUMT and present mtDNA sequence. As an example, for a sequence 300 bp long, 94% identity to mtDNA corresponds approximately to an insertion time of 3.3 Myr, and 96% to 2.2 Myr. Supporting Information {#s5} ====================== Table S1 ::: {.caption} ###### Sequence Analysis of NUMTs in the Human Genome (53 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S2 ::: {.caption} ###### Coordinates of the Genes Where NUMTs Are Inserted in the Human Genome (63 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a3} ----------------- The NCBI (<http://www.ncbi.nlm.nih.gov/genome/seq/page.cgi?F=HsBlast.html&&ORG=Hs>) accession number for the human mtDNA sequence is AB055387. We thank Shahragim Tajbakhsh, Marco Pontoglio, Simon Wain-Hobson, and Etienne Patin for stimulating discussions, suggestions, and for critical reading of the manuscript, and Lluis Quintana for critical advice. BD is Professor at the Université P. et M. Curie and member of Institut Universitaire de France. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. Academic Editor: Jonathan A. Eisen, Institute for Genomic Research Citation: Ricchetti M, Tekaia F, Dujon B (2004) Continued colonization of the human genome by mitochondrial DNA. PLoS Biol 2(9): e273. DSB : double-strand break mt : mitochondrial Myr : million years NHEJ : non-homologous end-joining NUMT : nuclear DNA sequence of mitochondrial origin

PubMed Central

2024-06-05T03:55:47.360272

2004-9-7

PMC515366

Introduction {#s1} ============ Activation of innate immunity in response to bacterial pathogens requires a group of molecules, known as the pattern recognition receptors, that recognize conserved motifs, present in bacteria but absent in higher eukaryotes, and trigger downstream signaling events. In *Drosophila,* two distinct signal transduction pathways are involved in the pathogen-specific innate immune response by inducing the expression of a panel of specific antimicrobial peptides ([@pbio-0020277-Tzou1]; [@pbio-0020277-Hoffmann1]). The Toll signaling pathway responds mainly to Gram-positive bacterial or fungal infections, which lead to the proteolytic processing of the cytokine-like polypeptide Spätzle. Binding of the cleaved Spätzle to the transmembrane receptor Toll activates an intracellular signaling cascade that results in the degradation of the IκB-like protein Cactus and the nuclear localization of the NF-κB--like proteins Dif and Dorsal, which induce the transcription of several antimicrobial peptide genes, such as *Drosomycin* ([@pbio-0020277-Lemaitre2], [@pbio-0020277-Lemaitre3]; [@pbio-0020277-Meng1]; [@pbio-0020277-Rutschmann2]; [@pbio-0020277-Tauszig-Delamasure1]; [@pbio-0020277-Weber1]). By contrast, the immune deficiency (Imd) pathway mediates defense reactions against primarily Gram-negative bacteria through different signaling components and regulates the cleavage and activation of another NF-κB--related nuclear factor, Relish, which activates a different set of antimicrobial peptide genes, including *Diptericin* ([@pbio-0020277-Lemaitre1]; [@pbio-0020277-Hedengren1]; [@pbio-0020277-Leulier1]; [@pbio-0020277-Rutschmann1]; [@pbio-0020277-Vidal1]). Several genetics studies have shown that the Toll pathway and the Imd pathway are activated specifically by two distinct peptidoglycan recognition proteins (PGRPs) in response to bacterial infections ([@pbio-0020277-Michel1]; [@pbio-0020277-Choe1]; [@pbio-0020277-Gottar1]; [@pbio-0020277-Ramet1]). PGRPs constitute a highly diversified family of proteins present in both insects and mammals. Members of the PGRP family are expressed as either secreted, cytosolic, or transmembrane forms, which all share a conserved 165-amino acid domain (the PGRP domain) with an evolutionary connection to bacteriophage T7 lysozyme ([@pbio-0020277-Yoshida1]; [@pbio-0020277-Kang1]; [@pbio-0020277-Ochiai1]; [@pbio-0020277-Werner1]; [@pbio-0020277-Liu1]). There are 13 PGRP genes in the genome of *Drosophila* ([@pbio-0020277-Werner1]). Remarkably, a gene knockout of PGRP-SA, an extracellular protein, is sufficient to eliminate Toll activation in response to the Gram-positive bacterium Micrococcus luteus in adult flies ([@pbio-0020277-Michel1]). Similar loss-of-function screenings have identified PGRP-LC as the surface transmembrane receptor for the Imd pathway, although another PGRP member, PGRP-LE, may also be involved in Imd activation ([@pbio-0020277-Choe1]; [@pbio-0020277-Gottar1]; [@pbio-0020277-Ramet1]; [@pbio-0020277-Takehana1]; [@pbio-0020277-Werner2]). Several PGRPs have been shown to bind peptidoglycan (PG) ([@pbio-0020277-Yoshida1]; [@pbio-0020277-Werner1]; [@pbio-0020277-Takehana1]; [@pbio-0020277-Kim1]), an essential and unique cell-wall polymer found in both Gram-positive and Gram-negative bacteria. PG is composed of long glycan chains made of two alternating sugars and cross-linked by short peptides. The subunits of PG, also known as muropeptides, are composed of *N*-acetyl glucosamine (GlcNAc) and *N*-acetyl muramic acid (MurNAc) plus a stem peptide chain consisting of [D]{.smallcaps}- and [L]{.smallcaps}- (or *meso*-) amino acids, with the third amino acid being most frequently lysine in Gram-positive bacteria and diaminopimelic acid (DAP) in Gram-negative bacteria. Recently, [@pbio-0020277-Leulier2] have shown that the Toll pathway is activated primarily by lysine-type PG found in most Gram-positive bacteria but responds weakly to DAP-type PG from Gram-negative bacteria. Not only did this finding reinforce the identification of PGRP-SA and PGRP-LCs as the putative receptors of the Toll and Imd pathways for bacterial molecular patterns, respectively, it also suggested that the signaling specificities of these two pathways might rely on the binding capability of the two activating PGRPs towards specific PG forms. Results/Discussion {#s2} ================== To facilitate molecular characterization of PG recognition and signal transduction mediated by PGRP-SA, we overexpressed and purified recombinant PGRP-SA (rPGRP-SA) in a baculovirus-insect cell expression system. PGRP-SA is a secreted protein circulating in the hemolymph (the insect blood) of *Drosophila.* We tested the activity of rPGRP-SA in vivo by injecting the protein into wild-type *(wt)* and PGRP-SA--deficient *(PGRP-SA^seml^)* flies. For this assay we used flies carrying a *Drosomycin-GFP* reporter transgene, which served as the target gene of the Toll signaling pathway. The *wt* flies injected with water produced *Drosomycin-GFP* after challenge by *M. luteus,* whereas *PGRP-SA^seml^* flies failed to express the reporter gene after the same treatment ([Figure 1](#pbio-0020277-g001){ref-type="fig"}A and [1](#pbio-0020277-g001){ref-type="fig"}B). When 112 ng of rPGRP-SA was injected into *PGRP-SA^seml^* flies, the recipient flies became capable of producing *Drosomycin-GFP* after challenge with M. luteus ([Figure 1](#pbio-0020277-g001){ref-type="fig"}C). As little as 11 ng of rPGRP-SA was sufficient to rescue *PGRP-SA^seml^* flies ([Figure 1](#pbio-0020277-g001){ref-type="fig"}D). Injection of 11 ng of rPGRP-SA in *wt* and *PGRP-SA^seml^* flies without any further microbial challenge could not activate *Drosomycin-GFP* expression (unpublished data). These results demonstrate that rPGRP-SA expressed in insect cell culture medium is active in vivo. ::: {#pbio-0020277-g001 .fig} Figure 1 ::: {.caption} ###### The In Vivo Rescuing and In Vitro PG-Binding Activities of Wild-Type rPGRP-SA (A--E) *Drosomycin-GFP* expression in (A) wild-type and (B to E) *PGRP-SA^seml^* flies after challenge by *M. luteus.* (A and B) Water or (C to E) rPGRP-SA at variable concentrations was injected into *Drosomycin-GFP* flies prior to the challenge with *M. luteus.* \(F) rPGRP-SA binds to both lysine-type *(M. luteus* and *E. faecalis)* and DAP-type *(E. coli* and *P. aeruginosa)* PGs but not to amidated DAP-type *(Bacillus subtilis* and *Bacillus thuringiensis)* PG. The left lane (Input) is loaded with the same amount (20 μg) of protein used for the binding assay. :::  ::: The selective activation of the Toll and Imd pathways by distinct classes of bacteria is mediated via recognition of specific forms of PGs ([@pbio-0020277-Leulier2]). We analyzed the PG binding of rPGRP-SA to test whether the differential activation of Toll by different PG forms reflects their different binding ability towards PGRP-SA. We found that rPGRP-SA binds to purified lysine-type PGs from M. luteus or Enterococcus faecalis and to DAP-type PGs from Escherichia coli or *Pseudomonas aeruginosa,* but not to amidated DAP-type PGs from *Bacillus* ([Figure 1](#pbio-0020277-g001){ref-type="fig"}F). Although the sensitivity of this assay is insufficient to compare the differential binding of rPGRP-SA to lysine- and DAP-type PGs, these results are overall in good agreement with previous in vivo challenge data ([@pbio-0020277-Leulier2]). Unlike lysine-type PGs, DAP-type PGs can only weakly activate the Toll pathway, which may be explained by the unexpected hydrolyzing activity of PGRP-SA for DAP-type PGs, as described below. We crystallized PGRP-SA using seeding methods, and collected complete data to 2.2-Å resolution from a single crystal plate at the SBC 19-ID beamline of the Advanced Photon Source (APS) at Argonne National Laboratory. The crystal structure was determined by molecular replacement, using the structure of PGRP-LB as a search model. PGRP-LB is a zinc amidase similar in structure to T7 lysozyme ([@pbio-0020277-Kim1]) and is 29% identical in amino acid sequence to PGRP-SA. Although the crystals were grown from full-length PGRP-SA (177 residues plus 6×His-tag; residues numbered from the N terminus of the purified polypeptide chain, as determined by N-terminal sequencing), clear electron density was visible only from Cys11 to Pro177. The structure of PGRP-SA reveals a single domain composed of a central seven-stranded mixed β sheet (B1, B3, B4, B5, B7, B8, and B9) flanked by three major helices (H2, H3, and H5), a small two-stranded parallel β sheet (B2 and B6), and two single-turn helices (H1 and H4) ([Figure 2](#pbio-0020277-g002){ref-type="fig"}A). The H2 helix contains one turn of rarely observed π helix at its C terminus (residues 64--70). This helix, together with the L1--L4 loops and the central β sheet, forms a prominent extended surface groove ([Figure 2](#pbio-0020277-g002){ref-type="fig"}A and [3](#pbio-0020277-g003){ref-type="fig"}), which in PGRP-LB includes a zinc cage ([@pbio-0020277-Kim1]). The overall structure of PGRP-SA strongly resembles that of PGRP-LB ([Figure 2](#pbio-0020277-g002){ref-type="fig"}B). The root-mean-square deviation (r.m.s.d.) of the 167 Cα positions after superposition is 1.22 Å. However, PGRP-SA has lost two of the four zinc-coordinating residues present in PGRP-LB ([Figure 2](#pbio-0020277-g002){ref-type="fig"}C); accordingly, the rPGRP-SA crystals exhibit no X-ray absorption at the zinc edge, and we found no electron density for possible metal ions around the groove. Other major differences are located in the N and C termini, the loop immediately preceding B3, the B4-B5 β-hairpin, and the L1 loop ([Figure 2](#pbio-0020277-g002){ref-type="fig"}B), where the sequences among PGRPs are highly diverse ([Figure 2](#pbio-0020277-g002){ref-type="fig"}C). PGRP-SA contains two disulfide bridges (Cys11-Cys134 and Cys48-Cys54), whereas PGRP-LB has only one. The highly conserved disulfide bridge Cys48-Cys54 is the target of the *PGRP-SA^seml^* mutation in which Cys54 is changed into a tyrosine ([@pbio-0020277-Michel1]). The Cys48-Cys54 bridge tethers the H2 helix to the central β sheet through the L1 loop ([Figure 2](#pbio-0020277-g002){ref-type="fig"}A). The other disulfide bond, between Cys11 and Cys134, is solvent exposed and anchors the N-terminal portion of PGRP-SA onto the H3 helix. As the C terminus of the protein is also tethered by insertion of the proline ring of the terminal residue Pro177 into a hydrophobic pocket formed by Ile148 of B8 and Val153 of H4, the structure of PGRP-SA appears to be more compact than that of PGRP-LB. The integral domain structure of PGRP-SA may be required for protein stability, considering that PGRP-SA is an extracellular protein secreted into the *Drosophila* hemolymph. The disulfide bridge Cys11-Cys134 may also be present in several mouse and human PGRPs ([Figure 2](#pbio-0020277-g002){ref-type="fig"}C). ::: {#pbio-0020277-g002 .fig} Figure 2 ::: {.caption} ###### PGRP-SA Structure and Sequence Comparisons \(A) Ribbon diagram showing the front view (left) and side view (right) of PGRP-SA. The ribbon is colored from N to C terminus in a progression from blue to red. Disulfide bridges are shown as sticks. The π helix turn at the end of the H2 helix is indicated. \(B) Comparison of PGRP-SA (blue coil, from N to C) and PGRP-LB (green coil, from N′ to C′). (A) and (B) were prepared with Bobscript ([@pbio-0020277-Esnouf1]), GL\_RENDER (E. Esser, personal communication), and POV-Ray (Persistence of Vision Ray Tracer v3.1g). \(C) Aligned sequences of selected PGRP domains, with a serine and a histidine at position 158 and position 42 of PGRP-SA (marked with asterisks), respectively, from *Drosophila* (d), mouse (m), and human (h). Secondary-structure elements in PGRP-SA are indicated above the alignment. Invariant residues are boxed in black and colored in white, conserved residues are shaded in yellow, and those lining the putative PG docking groove are in pink. The disulfide bond-forming Cys residues are boxed in gray. The residue number of PGRP-SA is shown above the alignment. The residues chosen for mutagenesis are marked with black circles. A structurally based alignment of the dPGRP-LB sequence is shown at the bottom with its amidase catalytic zinc-coordinating residues colored in red. :::  ::: ::: {#pbio-0020277-g003 .fig} Figure 3 ::: {.caption} ###### Structural Analysis of PGRP-SA (A and B) Molecular surfaces of (A) PGRP-SA and (B) PGRP-LB shown in similar orientations. Selected PGRP-SA residues on the putative PG docking groove are highlighted on the surface. Thr158 of PGRP-LB, the residue corresponding to Thr156 of PGRP-SA, is highlighted for reference. \(C) Stick model of the PGRP-SA residues chosen for mutational analysis. Residues are colored with the same rainbow-coloring scheme as in [Figure 2](#pbio-0020277-g002){ref-type="fig"}A. Figures were prepared with GRASP ([@pbio-0020277-Nicholls1]), Bobscript, GL\_RENDER, and POV-Ray. :::  ::: The most prominent feature of PGRP-SA is a long surface groove demarcated by residues of the H2 helix from one side and of the L1--L4 loops from the other side, with the residues from B3, B4, and B7 of the central β sheet forming the base of the groove ([Figure 2](#pbio-0020277-g002){ref-type="fig"}A). These residues are among the least conserved, and even the lengths of the L1 and L3 loops vary among members of the PGRP family ([Figure 2](#pbio-0020277-g002){ref-type="fig"}C). Therefore, the surface groove structure of PGRP-SA is distinct from that of PGRP-LB despite their overall structural similarity ([Figure 3](#pbio-0020277-g003){ref-type="fig"}A and [3](#pbio-0020277-g003){ref-type="fig"}B). The presence of a surface groove on PGRP-SA suggests that it may have a role in PG binding. We performed mutagenesis and functional analysis to test this hypothesis. In the following text, mutations are identified by the one-letter code for the residue in wild-type PGRP-SA, followed by the position of the residue in the amino acid sequence and the one-letter code for the residue to which it was mutated (e.g., S158A has serine in position 158 mutated to alanine). Residues on the surface groove whose side chains are solvent accessible were chosen for mutational analysis (see [Figure 2](#pbio-0020277-g002){ref-type="fig"}C, [3](#pbio-0020277-g003){ref-type="fig"}A, and [3](#pbio-0020277-g003){ref-type="fig"}C). These residues are located in three different subregions of the putative docking groove. The first group of residues constitutes the right-side wall of the groove, based on the front view shown in [Figure 3](#pbio-0020277-g003){ref-type="fig"} (Tyr64, His65, Asp70, Phe71, and Asn72). The second group is located on the left-side wall of the groove (Val44, Thr45, Tyr100, Ile154, and Ser158). The last group includes Ser75, which sits at the base of the groove. Based on the structure, we also made a Thr-to-Tyr mutation for residue 156; we reasoned that the introduced bulky side chain of Tyr would prevent access of PG to the putative docking groove. In addition, the single mutation I14A, located on the backside of the molecule, was made as a control. None of the residues chosen for mutagenesis is involved in extensive packing interactions. Hence, alterations of these residues are not expected to disrupt the tertiary structure of PGRP-SA. Our hypothesis was that, if the surface groove is indeed involved in PG recognition, the Ala mutations within the groove should exhibit reduced or altered PG-binding activities, whereas the T156Y mutation should completely abolish PG interaction. We analyzed the ability of these single- or multiple-Ala mutants to bind lysine-type PG from M. luteus by in vitro PG-binding assays ([Figure 4](#pbio-0020277-g004){ref-type="fig"}A). The in vivo activity of these rPGRP-SA mutants was examined by analyzing their capacity to rescue the *PGRP-SA^seml^* mutation in the assay described earlier; in addition, the *Drosomycin* expression was measured by quantitative real-time PCR analysis ([Figure 4](#pbio-0020277-g004){ref-type="fig"}B). These studies showed that mutations at almost every position tested on both walls of the groove region led to impaired PG binding and Toll signaling activity except the S75A mutant, which exhibited an enhanced PG-binding ability. The T156Y mutation, as expected, resulted in a complete loss of the interaction with PG ([Figure 4](#pbio-0020277-g004){ref-type="fig"}A); as a result, the mutant protein failed to activate the Toll pathway ([Figure 4](#pbio-0020277-g004){ref-type="fig"}B). Notably, the single mutations S158A and S158C also completely abolished the function of the protein both as a PG recognition receptor and as a Toll activator. The enhanced activity on both PG binding and Toll activation of the S75A mutants suggests that the removal of the hydroxyl group of Ser75 may create a better binding surface for the PG. In fact, Ala and Gly are commonly found at this position in the sequences of PGRPs (see [Figure 2](#pbio-0020277-g002){ref-type="fig"}C). As expected, the I14A mutation on the backside of the molecule did not affect PG binding or Toll activation (unpublished data). It is interesting that some mutants of PGRP-SA with apparent PG-binding deficiency, for example Y100A and V44A/T45A, could still induce *Drosomycin* expression upon injection in response to challenge with PG from *M. luteus.* This discrepancy may be the result of different sensitivities between the gel-based PG-binding assay, which examines plain physical interaction between PGRP-SA and PG, and the rescue assay, by which the amplified signaling outcome of PG interaction with PGRP-SA, namely *Drosomycin* expression, is observed. Nevertheless, these results together indicate that the surface groove of PGRP-SA mediates both interaction with PG and activation of the Toll pathway. These studies further underscore the role of PGRP-SA as a true pattern recognition receptor, as they demonstrate the correlation between the biochemical recognition of PG and Toll activation through PGRP-SA. ::: {#pbio-0020277-g004 .fig} Figure 4 ::: {.caption} ###### Mutational Analysis of PGRP-SA \(A) Upper panel shows the wild-type and mutant rPGRP-SA pulled down by lysine-type PG from *M. luteus.* Lower panel (Input) shows the corresponding protein samples (20 μg) without incubation with PG. \(B) The relative Toll signaling activities of the rPGRP-SA mutants. At least three repeats were performed for each experiment. Each bar represents the mean with the standard deviation. The values obtained for the wild type after M. luteus PG injection were arbitrarily set to 100 (upper dashed line). The background activity level is indicated by the lower dashed line. CTR, unchallenged control. :::  ::: The *PGRP-SA^seml^* mutation results in a Cys-to-Tyr mutation at position 54, which is engaged in a highly conserved Cys48-Cys54 disulfide linkage ([@pbio-0020277-Michel1]). We conducted mutational analysis to investigate whether the *PGRP-SA^seml^* mutation eliminates PGRP-SA function by disrupting the conserved disulfide bridge and thus affecting the protein stability, or by sterically blocking the surface groove with the bulky side chain of Tyr. We found that the C48A mutant failed to be expressed in the insect cell culture, in which all the other wild-type and mutant proteins were expressed (unpublished data). As a result, the culture medium from the C48A mutant failed to restore the *PGRP-SA^seml^* phenotype after injection in *Drosophila* ([Figure 4](#pbio-0020277-g004){ref-type="fig"}B). By contrast, the C11A mutation, which disrupts the Cys11-Cys134 disulfide bond on the backside of the molecule, had little effect on the function of the protein ([Figure 4](#pbio-0020277-g004){ref-type="fig"}B). These results suggest that the *PGRP-SA^seml^* mutation disrupts the proper folding of PGRP-SA rather than PG interaction via the docking groove. It is intriguing that both of the single mutants, S158A and S158C, fail to bind lysine-type PG and to activate Toll. The S158 residue is located on one wall of the docking groove. Seven of the *Drosophila* PGRPs, including PGRP-SC1B and PGRP-LB, have been suggested to possess amidase activity. These "amidase PGRPs" all have a Cys residue at this position, which appears to participate in the Zn coordination in the active site. Substitution of Cys to Ser or Ala in PGRP-SC1B eliminates its enzymatic activity but not its capacity to bind PG, suggesting that the Cys is required for amidase activity ([@pbio-0020277-Mellroth1]). Our data show that, in PGRP-SA, Ser158 is essential for its interaction with PG. Also, Ser158 is highly conserved among PGRPs that have lost the amidase catalytic residues (see [Figure 2](#pbio-0020277-g002){ref-type="fig"}C). The drastic loss-of-function effect of S158A and S158C mutations suggests that the chemical property of the Ser side chain at this position may be critical for function. Recently, muropeptides have been identified as the bacterial molecular patterns detected by Nod proteins ([@pbio-0020277-Chamaillard1]; [@pbio-0020277-Girardin1], [@pbio-0020277-Girardin2]; [@pbio-0020277-Inohara1]). Free muropeptides are found within bacterial cells as they are constantly synthesized de novo or hydrolyzed from PG and recycled during cell division ([@pbio-0020277-Goodell1]; [@pbio-0020277-Goodell2]); they could be released from bacterial cells during infection and exploited for bacterial sensing by pattern recognition molecules in the host. PGRPs are structural homologues of the *N*-acetylmuramoyl-[L]{.smallcaps}-alanine amidase superfamily of proteins, including AmpD and T7 lysozyme, which can hydrolyze monomeric muropeptides or larger PG fragments ([@pbio-0020277-Inouye1]; [@pbio-0020277-Kang1]; [@pbio-0020277-Liepinsh1]; [@pbio-0020277-Mellroth1]). Therefore, muropeptides or their peptidic moieties may serve similarly as specific ligands for PGRP-SA and PGRP-LC via interaction with the PG docking groove. Accordingly, our structural modeling has indicated that the structure of the long docking groove on PGRP-SA is able to fit a ligand with elongated conformation, which a muropeptide or its stem peptide could adopt (unpublished data). Previously, PGRP-SC1B and PGRP-LB have been demonstrated to display a T7 lysozyme-like amidase activity ([@pbio-0020277-Kim1]; [@pbio-0020277-Mellroth1]). We sought to determine if rPGRP-SA has any enzymatic activity, although it has been believed not to possess such an activity; PGRP-SA is missing a critical cysteine residue found in the active site of these amidase PGRPs ([@pbio-0020277-Mellroth1]) (see [Figure 2](#pbio-0020277-g002){ref-type="fig"}C). We incubated rPGRP-SA separately with either the lysine-type muropeptide, GlcNAc-MurNAc(anhydro)-[L]{.smallcaps}-Ala-γ-[D]{.smallcaps}-Glu-[L]{.smallcaps}-Lys-[D]{.smallcaps}-Ala, or the corresponding DAP-type muropeptide, GlcNAc-MurNAc(anhydro)-[L]{.smallcaps}-Ala-γ-[D]{.smallcaps}-Glu-*meso*-DAP-[D]{.smallcaps}-Ala, and analyzed the reaction mixtures afterwards by high performance liquid chromatography (HPLC) ([Figure 5](#pbio-0020277-g005){ref-type="fig"}). To our surprise, we observed within 40 h of incubation a near-complete cleavage of the DAP-type muropeptide, but not the lysine-type compound, at a specific peptide bond position, resulting in a product consisting of the tripeptide derivative GlcNAc-MurNAc(anhydro)-[L]{.smallcaps}-Ala-γ-[D]{.smallcaps}-Glu-*meso*-DAP, with the release of the terminal [D]{.smallcaps}-Ala ([Figure 5](#pbio-0020277-g005){ref-type="fig"}B to [5](#pbio-0020277-g005){ref-type="fig"}E). These results demonstrated that rPGRP-SA had cleaved between the *meso*-DAP at position 3 and the [D]{.smallcaps}-Ala at position 4 on the stem peptide and thus exhibited an [L]{.smallcaps},[D]{.smallcaps}-carboxypeptidase activity. Typical Michaelis-Menten kinetics were observed in the substrate concentration range considered (10--500 μM). The *K* ~m~ value of rPGRP-SA for its substrate GlcNAc-MurNAc(anhydro)-[L]{.smallcaps}-Ala-γ-[D]{.smallcaps}-Glu-*meso*-DAP-[D]{.smallcaps}-Ala was 21.4 ± 1.8 μM and the catalytic constant *k* ~cat~ was 0.48 ± 0.02 h^−1^. The small turnover number of PGRP-SA estimated for hydrolyzing DAP-PG substrates is comparable to the *k* ~cat~ of small G proteins such as Ras GTPases. We also tested different other PG-related tetrapeptide compounds as substrates and found that rPGRP-SA hydrolyzed the same peptide bond in DAP-containing muropeptides but had no detectable activity on all the lysine-type compounds tested (unpublished data). This enzymatic activity of rPGRP-SA was not inhibited by ethylenediaminetetraacetic acid (EDTA) or by phenylmethylsulfonyl fluoride (PMSF) (unpublished data). We observed that the two Ser158 mutants, S158C and S158A, did not exhibit any detectable activity ([Figure 5](#pbio-0020277-g005){ref-type="fig"}F and unpublished data), although both bind DAP-PG from E. coli as wild-type rPGRP-SA does ([Figure 6](#pbio-0020277-g006){ref-type="fig"}A). In the PGRP-SA structure the Oγ of the Ser158 residue is positioned within hydrogen-bonding distance (2.95 Å) of the Nδ1 of the highly conserved His42 residue ([Figure 6](#pbio-0020277-g006){ref-type="fig"}B). The fact that the enzymatic activity of rPGRP-SA can be completely eliminated by removing the hydroxyl group of Ser158 (S158A) or by replacing it with a thiol group (S158C) suggests that Ser158 is involved in catalysis rather than in the binding of the DAP-containing substrates. In fact, a Ser-His catalytic dyad of a catalytic antibody was found to be sufficient for catalyzing the hydrolysis of amino acid esters ([@pbio-0020277-Zhou1]). To test this hypothesis, we generated a H42A mutant and analyzed its enzymatic activity. Indeed, this mutant is incapable of hydrolyzing DAP-peptide substrate, although it preserves the ability to bind DAP-PG (see [Figure 5](#pbio-0020277-g005){ref-type="fig"}G and [6](#pbio-0020277-g006){ref-type="fig"}A). Therefore, this result supports the catalytic role of the S158-H42 dyad for the hydrolyzing activity of PGRP-SA. Our enzymatic analysis data show that rPGRP-SA is a carboxypeptidase with an apparent specificity for the [L,D]{.smallcaps}-configured DAP-peptide bond between the carboxyl group at the [L]{.smallcaps}-center of the *meso*-DAP and the amino group of the following [D]{.smallcaps}-Ala residue. Through showing that DAP-containing muropeptides are the substrates for the PGRP-SA enzyme, we provide biochemical evidence suggesting that PGRP-SA may recognize a specific monomeric PG fragment. In support of this finding, it has been demonstrated very recently that specific monomeric DAP-PG fragments can activate the Imd pathway via PGRP-LC in flies and in cell culture ([@pbio-0020277-Kaneko1]). ::: {#pbio-0020277-g005 .fig} Figure 5 ::: {.caption} ###### PGRP-SA Is an [L,D]{.smallcaps}-Carboxypeptidase \(A) Chemical structures of the DAP-type muropeptide, GlcNAc-MurNAc(anhydro)-[L]{.smallcaps}-Ala-γ-[D]{.smallcaps}-Glu-*meso*-**DAP**-[D]{.smallcaps}-Ala, and lysine-type muropeptide, GlcNAc-MurNAc(anhydro)-[L]{.smallcaps}-Ala-γ-[D]{.smallcaps}-Glu-[L]{.smallcaps}-**Lys**-[D]{.smallcaps}-Ala, used in the enzymatic assays (substrates S~1~ and S~2~, respectively). The arrow indicates the site of cleavage of the DAP-type substrate S~1~ by the [L,D]{.smallcaps}-carboxypeptidase activity. (B--G) Reverse-phase HPLC analysis. Cleavage of the DAP-type substrate S~1~ by wild-type rPGRP-SA results in the generation of GlcNAc-MurNAc(anhydro)-[L]{.smallcaps}-Ala-γ-[D]{.smallcaps}-Glu-*meso*-DAP (P~1~ product). The position of the peak corresponding to standard GlcNAc-MurNAc(anhydro)-[L]{.smallcaps}-Ala-γ-[D]{.smallcaps}-Glu-[L]{.smallcaps}-Lys (P~2~ product, not generated by rPGRP-SA) is indicated. \(B) Incubation of S~1~ without rPGRP-SA for 40 h. (C and D) Incubation of S~1~ with rPGRP-SA for (C) 24 h and (D) 40 h. \(E) Incubation of S~2~ with rPGRP-SA for 70 h. (F and G) Incubation of S~1~ with the (F) S158C and (G) H42A mutants for 40 h. :::  ::: ::: {#pbio-0020277-g006 .fig} Figure 6 ::: {.caption} ###### DAP-Type PG-Binding Activities of the S158A/C and H42A Mutants and the Structure of the S158-H42 Dyad in the Active Site \(A) Upper panel shows the wild-type and mutant rPGRP-SA pulled down by DAP-type PG from *E. coli.* Lower panel (Input) shows the corresponding protein samples (20 μg) without incubation with PG. \(B) Stereo diagram showing the putative active-site residues. Prepared with Bobscript, GL\_RENDER, and POV-Ray. :::  ::: So far, only one [L,D]{.smallcaps}-carboxypeptidase, from *E. coli,* has been identified and characterized ([@pbio-0020277-Ursinus1]; [@pbio-0020277-Templin1]). Hence, we report here the first eukaryotic protein exhibiting such an activity specific for peptide bonds existing only in prokaryotes. The DAP-PG hydrolyzing activity of PGRP-SA has a rather slow turnover number. It would be interesting to see if this low intrinsic DAP-PG hydrolyzing activity could be stimulated by another hemolymph protein(s) that could associate with PGRP-SA, such as Gram-negative bacteria-binding protein 1 (GNBP1) (see below). Our observation that a pattern recognition receptor has enzymatic activity is unexpected. However, as rPGRP-SA binds to both lysine-type and DAP-type PGs (see [Figure 1](#pbio-0020277-g001){ref-type="fig"}F), this finding suggests that the specific activation of Toll by lysine-type PG is achieved by the concomitant ability of PGRP-SA to recognize lysine-type PG and to hydrolyze DAP-type PG. Since the latter has been identified as a strong inducer of the Imd pathway, it will be interesting to see whether the [L,D]{.smallcaps}-carboxypeptidase activity of PGRP-SA can influence the Imd pathway positively by generating specific PG fragments that are recognized by PGRP-LC or negatively by scavenging DAP-PG to eliminate its immune-elicitor activity. As PGRP-LCx/a, the other known pathogen-sensing receptors, and several other PGRPs also possess a Ser residue at the position equivalent to S158 in PGRP-SA (see [Figure 2](#pbio-0020277-g002){ref-type="fig"}C), it will be interesting to see if they possess similar enzymatic activity. In the present study, we characterized the PG docking groove of PGRP-SA through a combined structural and functional analysis, and we showed that this surface groove mediates both PG sensing and Toll signaling and that Ser158 in the groove is involved in PG interaction, Toll activation, and the newly discovered [L,D]{.smallcaps}-carboxypeptidase activity. The S158C mutation has a dramatic negative effect on the ability to activate the Toll pathway. This suggests that the hydroxyl group of Ser158 may mediate critical interaction, perhaps via a hydrogen bond, with a specific lysine-type PG fragment and that this interaction may contribute the bulk of the binding energy. However, we found that the same mutation did not affect binding to DAP-type PG, suggesting that Ser158 is not critical for DAP-type PG interaction ([Figure 6](#pbio-0020277-g006){ref-type="fig"}A). Therefore, PGRP-SA appears to employ different binding modes for interactions with lysine-type PG versus DAP-type PG. *meso*-DAP differs from [L]{.smallcaps}-lysine only by the substitution of a carboxyl group on the Cɛ with [D]{.smallcaps}-chirality. As the carboxypeptidase activity of PGRP-SA can act only on DAP-type and not on lysine-type PG compounds, it is likely that the carboxyl group at the [D]{.smallcaps}-center of DAP provides the critical interaction(s) with the docking groove residue(s) to help orient the peptide bond between DAP and [D]{.smallcaps}-Ala. However, understanding the structural basis of the selectivity to DAP-PG over lysine-PG would require a cocrystal structure of PGRP-SA and a lysine-type PG ligand. As the PG docking groove is lined with residues that are highly diverse among different PGRPs, indicating that each PGRP protein may bind to a specific set of PG fragments, determining the structure of PG ligand-bound PGRP-SA will also provide important insights into PG recognition specificity of this family of proteins. However, so far, no cocrystal of a PGRP protein with a PG compound has been obtained. *Drosophila* possesses a high number of genes encoding serine proteases and serine protease inhibitors (serpins) ([@pbio-0020277-Rubin1]). Serine protease cascades, operating through sequential zymogen activation, have been implicated in dorsal-ventral fate determination and hemolymph clotting in arthropods ([@pbio-0020277-Krem1]). A hemolymph serine protease (Persephone) has been shown to mediate the cleavage of Spätzle in response to fungal infection ([@pbio-0020277-Ligoxygakis1]). As PGRP-SA is not involved in fungal-dependent cleavage of Spätzle and activation of Toll, it is believed that this hemolymph pattern recognition protein activates another unidentified proteolytic enzyme(s), resulting in the cleavage of Spätzle specifically in response to bacterial infection. Recently, another hemolymph protein, GNBP1, has been shown to critically participate in activating Toll, perhaps by associating with PGRP-SA, in response to Gram-positive bacterial infection ([@pbio-0020277-Gobert1]; [@pbio-0020277-Pili-Floury1]). Based on our result indicating that PGRP-SA may recognize monomeric PG ligands, it is likely that docking of the specific PG compound onto the surface groove of PGRP-SA may create a new molecular surface that would allow interaction with other Toll-activating factors such as GNBP1. Furthermore, perhaps a multiprotein complex involving PG ligand-bound PGRP-SA and GNBP1 is involved in direct proteolytic activation of the upstream protease of a Spätzle-processing protease cascade. Alternatively, a PG-dependent PGRP-SA/GNBP1 complex may be involved in binding and sequestering a serpin to release the inhibition of the Spätzle-processing enzyme cascade. A hemolymph serpin (Necrotic) has been implicated in inhibiting the proteolytic cleavage of Spätzle upon fungal infection ([@pbio-0020277-Levashina1]). Although a better understanding of PGRP-SA/GNBP1-activated cleavage of Spätzle will require identification of critical players that link the microbial recognition to the proteolytic activation of Spätzle, more detailed biochemical and structural studies on the minimal PG moiety recognized by PGRP-SA and the interaction between PGRP-SA, its specific PG ligand, and GNBP1 are necessary to help define the molecular mechanism of PG recognition mediated by these pattern recognition receptors. Materials and Methods {#s3} ===================== {#s3a} ### Protein expression, purification, and crystallization {#s3a1} Details on the cloning, expression, purification, and crystallization of recombinant *Drosophila* PGRP-SA will be presented elsewhere. Briefly, full-length PGRP-SA (including its N-terminal signal peptide) with a 6×His tag at the C terminus was overexpressed in insect Hi-5 cells using the Bac-to-Bac baculovirus expression system (Invitrogen, Carlsbad, California, United States) and purified with Talon metal affinity resins (Clontech, Palo Alto, California, United States) followed by size exclusion on a Superdex 75 column (Pharmacia, New York, New York, United States) pre-equilibrated in 20 mM Tris-HCl (pH 7.8) and 300 mM NaCl. The purified protein was analyzed by N-terminal sequencing and mass spectrometry to ensure its identity and purity. Crystallization was carried out at 21 °C by the hanging-drop vapor diffusion technique. The protein formed plate-like clusters over a reservoir containing 2.0 M NaKPO~4~ (pH 6.2). Single crystals were produced by two successive rounds of streak- and macroseeding and were cryoprotected in reservoir solution supplemented with 30% glycerol before data collection. ### Structure determination {#s3a2} We collected X-ray diffraction data using synchrotron radiation at the 19-ID beamline at APS. The diffraction images were processed and scaled with the HKL2000 package ([@pbio-0020277-Otwinowski1]). The positions of the two molecules in the asymmetric unit were determined by molecular replacement with the program AmoRe ([@pbio-0020277-Navaza1]) using the PGRP-LB structure as the search model (PDB code 1OHT). The two solutions were related by rotation and translation operations, generating a nonsymmetric dimer. The current model was refined after iterative cycles of manual rebuilding with the program O ([@pbio-0020277-Jones1]) and refinement with the program CNS ([@pbio-0020277-Brunger1]) ([Table 1](#pbio-0020277-t001){ref-type="table"}). The PGRP-SA dimer in the crystal is probably not biologically relevant, as it was not revealed by gel-filtration chromatography; moreover, the dimer interface was found to involve several phosphate ions from the crystallizing reagent and the first His residue of the affinity tag from one of the monomers. ::: {#pbio-0020277-t001 .table-wrap} Table 1 ::: {.caption} ###### Data Collection and Refinement Statistics :::  ^a^  *R* ~merge~ = Σ\|(*I* ~hkl~) − \<*I*\>Σ\|/(*I* ~hkl~), where *I* ~hkl~ is the integrated intensity of a given reflection ^b^  *R* = (Σ\|*F* ~o~ − *F* ~c~\|)/(Σ*F* ~o~), where *F* ~o~ and *F* ~c~ are observed and calculated structure factors, respectively ^c^ Calculated using the program PROCHECK ([@pbio-0020277-Laskowski1]) ::: ### Site-directed mutagenesis {#s3a3} Point mutations were generated by a PCR-based strategy using the QuikChange Kit (Stratagene, La Jolla, California, United States), and the identities of the mutagenized products were verified by sequencing. ### Fly stocks and protein microinjection {#s3a4} y, w, P(*ry^+^, Diptericin-lacZ*), P(*w^+^, Drosomycin-GFP*) flies were used as wild-type strains ([@pbio-0020277-Manfruelli1]). *Drosomycin-GFP, PGRP-SA^seml^* is a line carrying the *semmelweis* mutation in *PGRP-SA* (C54Y) ([@pbio-0020277-Michel1]). *Drosophila* stocks were maintained at 25 °C with standard medium. A quantity of 9.2 nl of water or rPGRP-SA protein was injected into the thorax of wild-type or *PGRP-SA^seml^* female adults (3--4 d old) using a Nanoject apparatus (Drummond, Broomall, Pennsylvania, United States). One hour later, flies were infected with a thin needle previously dipped into a concentrated culture of M. luteus or given an injection of 9.2 nl of M. luteus PG (5 mg/ml). Flies were then incubated for 24 h at 25 °C. A highly purified solution of M. luteus PG was produced and injected in flies as described by [@pbio-0020277-Leulier2]. ### Quantitative real-time PCR {#s3a5} For *Drosomycin* quantification from whole animals, RNA was extracted using RNA TRIzol (Invitrogen). cDNAs were synthesized using SuperScript II (Invitrogen) and PCR was performed using dsDNA dye SYBR Green I (Roche Diagnostics, Basel, Switzerland) on a Lightcycler (Roche). All samples were analyzed in duplicate and the amount of mRNA detected was normalized to control Rp49 mRNA values. We used normalized data to quantify the relative levels of a given mRNA according to cycling threshold analysis (ΔCt). ### PG-binding assay {#s3a6} The assay was performed at 4 °C by incubating 20 μg of purified wild-type or mutant rPGRP-SA with 300 μg of insoluble PGs, prepared as described previously ([@pbio-0020277-Leulier2]), in 300 μl of binding buffer containing 20 mM Tris-HCl (pH 7.8) and 300 mM NaCl on a shaking platform for 1 h. Bound protein, retained in the PG pellet after spinning the incubation mixture at 16,000 × *g* for 5 min, was washed with 1 ml of binding buffer followed by a 5-min spin and finally dissolved in 10 μl of SDS buffer for sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) analysis. The PG-bound rPGRP-SA was visualized by Coomassie Blue staining. ### Enzymatic assay and reverse-phase HPLC analysis {#s3a7} The activity was tested in 20 mM HEPES (pH 7.4) containing 2.5 mM EDTA, 50 μM substrate, and enzyme (25 μg of wild-type or mutant rPGRP-SA) in a total volume of 50 μl. After incubation for the indicated period at 37 °C, the mixture was injected on a Nucleosil 100 C~18~ 5μ reverse-phase HPLC column (4.6 mm × 250 mm, Alltech France, Templemars, France) and elution was performed at 0.6 ml/min with buffer A (50 mM sodium phosphate \[pH 4.45\]) for 10 min and then with a gradient of methanol in buffer A (from 0% to 25% in 50 min). Peaks were detected at 215 nm. In all cases, substrates and products were purified and desalted by HPLC, and their identity was confirmed by amino acid and mass spectrometry analyses. Supporting Information {#s4} ====================== The atomic coordinates and structure factors have been deposited in the Protein Data Bank (<http://www.rcsb.org/pdb/>) under accession number 1S2J. We thank Kirsten Fischer Lindahl for critical reading of the manuscript, Geneviève Auger and Didier Blanot (UMR 8619) for amino acid and mass spectrometry analyses, Peter Wu for technical assistance, and Zbyszek Otwinowski, Dominika Borek, and the staff of the SBC beamline 19-ID at the APS (supported by the United States Department of Energy) for advice in data collection. The original PGRP-SA cDNA plasmid was a gift from Thomas Werner. The laboratory of BL was funded by the Fondation Schlumberger and the Programme Microbiologie. The laboratory of DML was supported by the Centre National de la Recherche Scientifique. JD is an Investigator of the Howard Hughes Medical Institute. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** CIC, BL, and DML conceived and designed the experiments. CIC, SPF, BL, and DML performed the experiments. CIC, SPF, BL, DML, and JD analyzed the data. MH, CP, and YC contributed reagents/materials/analysis tools. CIC, BL, DML, and JD wrote the paper. Academic Editor: Pamela Bjorkman, Howard Hughes Medical Institute and California Institute of Technology Citation: Chang CI, Pili-Floury S, Hervé M, Parquet C, Chelliah Y, et al. (2004) A *Drosophila* pattern recognition receptor contains a peptidoglycan docking groove and unusual [L,D]{.smallcaps}-carboxypeptidase activity. PLoS Biol 2(9): e277. APS : Advanced Photon Source DAP : diaminopimelic acid EDTA : ethylenediaminetetraacetic acid GFP : green fluorescent protein GlcNAc : *N*-acetyl glucosamine GNBP1 : Gram-negative bacteria-binding protein 1 HPLC : high performance liquid chromatography Imd : immune deficiency MurNAc : *N*-acetyl muramic acid PG : peptidoglycan PGRP : peptidoglycan recognition protein *PGRP-SA^seml^* : PGRP-SA deficient PMSF : phenylmethylsulfonyl fluoride r.m.s.d. : root-mean-square deviation rPGRP-SA : recombinant PGRP-SA SDS-PAGE : sodium dodecyl sulfate polyacrylamide gel electrophoresis serpin : serine protease inhibitor *wt* : wild type

PubMed Central

2024-06-05T03:55:47.364832

2004-9-7

PMC515367

**Introduction** {#s1} ================ Despite intense study and interest, a detailed understanding of the evolutionary and demographic forces that have shaped extant patterns of human genomic variation remains elusive. An important goal in studies of DNA sequence variation is to identify loci that have been targets of natural selection and thus contribute to differences in fitness between individuals in a population. Identifying regions of the human genome that have been subject to natural selection will provide important insights into recent human history ([@pbio-0020286-Sabeti1]; [@pbio-0020286-Tishkoff1]), the function of genes ([@pbio-0020286-Akey1]), and the mechanisms of evolutionary change ([@pbio-0020286-Otto1]), and it may also facilitate the identification of complex disease genes ([@pbio-0020286-Jorde1]; [@pbio-0020286-Nielsen1]). The neutral theory of molecular evolution ([@pbio-0020286-Kimura1]; [@pbio-0020286-King1]), which posits that the majority of polymorphisms have no appreciable effects on fitness, has been integral to recent studies of natural selection. Specifically, the neutral theory makes explicit and quantitative predictions about the amount, structure, and patterns of sequence variation expected under neutrality, and serves as a null hypothesis by which to evaluate the evidence for or against selection in empirical data ([@pbio-0020286-Otto1]; [@pbio-0020286-Nielsen1]). Unfortunately, robust inferences of natural selection from DNA sequence data are difficult because of the confounding effects of population demographic history. For example, both positive selection and increases in population size lead to an excess of low-frequency alleles in a population relative to what is expected under a standard neutral model (i.e., a constant-size, randomly mating population at mutation-drift equilibrium). Therefore, rejection of the standard neutral model usually cannot be interpreted as unambiguous evidence for selection. One way out of this conundrum is to recognize that population demographic history affects patterns of variation at all loci in a genome in a similar manner, whereas natural selection acts upon specific loci ([@pbio-0020286-Cavalli-Sforza1]; [@pbio-0020286-Przeworski2]; [@pbio-0020286-Andolfatto1]; [@pbio-0020286-Nielsen1]). Thus, by sampling a large number of unlinked loci throughout the genome, it is in principle possible to distinguish between selection and demography. For instance, [@pbio-0020286-Akey1] recently used this approach to infer the presence of selection in a genome-wide collection of single nucleotide polymorphisms (SNPs). However, studies based on SNPs that were initially identified in a small sample and subsequently genotyped in a larger sample are not ideally suited for detecting selection, because ascertainment bias (i.e., a systematic bias introduced into a dataset because of the way in which the data were collected) complicates downstream analyses ([@pbio-0020286-Akey2]). However, DNA sequence data provides the opportunity to exhaustively catalog variation, which attenuates the problem of ascertainment bias and therefore is arguably the most powerful and direct approach for detecting selection. Here, we describe an extensive analysis of the molecular evolution of 132 genes that were comprehensively resequenced in 24 African-Americans and 23 European-Americans. In total, over 2.5 Mb of baseline reference DNA was sequenced, spanning 20 autosomal chromosomes and the X chromosome. The sampling of a large number of loci dispersed throughout the genome has allowed us to clarify the relative contributions of demography and selection to patterns of genetic variation at individual genes. Specifically, we developed a rigorous computational approach for taking into account multiple hypothesis tests and demographic history, and we found that while many apparent selective events can instead be explained by demography, there is also strong evidence for positive or balancing selection at eight genes in the European-derived population. In addition, we describe a striking example of a previously unreported recent selective sweep in European-Americans that spans four contiguous genes on Chromosome 7. More generally, our data provide insight into the demographic histories of African-American and European-American populations and have important implications for genetic association studies of complex diseases, as several of the genes showing evidence of selection have been implicated in susceptibility to complex human diseases. **Results** {#s2} =========== Statistical Tests Reveal Many Deviations from Neutrality {#s2a} -------------------------------------------------------- We resequenced 132 genes primarily involved in inflammation, blood clotting, and blood pressure regulation and discovered a total of 12,890 SNPs ([Table S1](#st001){ref-type="supplementary-material"}). We first characterized patterns of genetic variation by calculating several common summary statistics of the within-population allele frequency distribution, including Tajima\'s D, Fu and Li\'s D\*, Fu and Li\'s F\*, and Fay and Wu\'s H. As is conventionally done, we initially determined whether these statistics were significantly different from what is expected under a standard neutral model by performing coalescent simulations under the simplifying assumption of no recombination. In total, 28 genes in the European-American sample and ten genes in the African-American sample were nominally significant (i.e., the observed test statistic differed from neutral expectations at *p* \< 0.05) in one or more tests of the allele frequency distribution ([Figure 1](#pbio-0020286-g001){ref-type="fig"}). Thus, the European-American sample contained nearly three times as many significant genes as the African-American sample, and only three genes were significant in both samples *(ABO, IL1RN,* and *TNFRSF1B)*. ::: {#pbio-0020286-g001 .fig} Figure 1 ::: {.caption} ###### Scatter Plot of Neutrality Test Statistics in European- and African-Americans Genes that are nominally significant (*p* \< 0.05) in European-Americans (EA), African-Americans (AA), or both populations are denoted by red, blue, and green circles, respectively. Genes that are not significant are shown as black dots. Two-sided tests were used for Tajima\'s D, Fu and Li\'s D\*, and Fu and Li\'s F\*, and a one-sided test was used for Fay and Wu\'s H. :::  ::: The direction of Tajima\'s D, Fu and Li\'s D\*, and Fu and Li\'s F\* is potentially informative about the evolutionary and demographic forces that a population has experienced. For example, negative values reflect an excess of rare polymorphisms in a population, which is consistent with either positive selection or an increase in population size. Positive values indicate an excess of intermediate-frequency alleles in a population and can result from either balancing selection or population bottlenecks. In the European-American sample, we observed eleven significantly positive and five significantly negative values for one or more of these three test statistics ([Figure 1](#pbio-0020286-g001){ref-type="fig"}). In the African-American sample, we observed two significantly positive and five significantly negative values for one or more of the test statistics ([Figure 1](#pbio-0020286-g001){ref-type="fig"}). The observations of both significantly positive and significantly negative values of Tajima\'s D, Fu and Li\'s D\*, and Fu and Li\'s F\*, combined with the largely nonoverlapping set of significant genes, could reflect selective pressures unique to one population (i.e., local adaptation), different demographic histories, spurious results, or most likely some complex combination of all of these factors. Although these results are intriguing, their interpretation is confounded by two issues: (1) We have not corrected for multiple hypothesis tests, and (2) rejection of the standard neutral model can result from either selective or demographic forces. In the subsequent sections, we develop approaches to address these issues with the dual goals of identifying genes that possess strong evidence of natural selection and of inferring population demographic history. Correcting for Multiple Hypothesis Tests {#s2b} ---------------------------------------- In order to robustly correct for multiple hypothesis tests, the conventional practice of assuming no recombination when determining significance is not appropriate, because it results in conservative *p* values ([@pbio-0020286-Wall1]) and hence decreases the statistical power to detect deviations from neutrality. Although recombination can easily be incorporated into coalescent simulations, in practice it is difficult to accurately estimate recombination rates, which vary substantially across the genome ([@pbio-0020286-Yu1]; [@pbio-0020286-McVean1]). To model the stochastic behavior and uncertainty in local rates of recombination, we reassessed the significance of Tajima\'s D, Fu and Li\'s D\*, Fu and Li\'s F\*, and Fay and Wu\'s H by coalescent simulations that incorporate recombination rates sampled from a Gamma(2, 0.5 × 10^--8^) distribution (see [Materials and Methods](#s4){ref-type="sec"}). Finally, we corrected each statistic for multiple tests using the positive false discovery rate (FDR; [@pbio-0020286-Storey1]) method, which determines the predicted proportion of "false positives" for the number of significant observations. In the European-American sample, we observed 22 genes that were significant at a FDR of 5% (i.e., we expect approximately one false positive in this set of genes) for one or more tests of the allele frequency distribution ([Tables 1](#pbio-0020286-t001){ref-type="table"} and [S2](#s5){ref-type="sec"}). Thus, the number of significant genes in the European-American sample, after incorporating recombination and correcting for multiple tests, is very similar to the initial results where recombination was ignored and multiple tests were not corrected for. However, in the African-American sample there were no genes significant at a FDR of 5% for any of the tests of the allele frequency distribution (unpublished data). This result is consistent with the relatively small number of significant genes that were initially found before correcting for multiple tests ([Figure 1](#pbio-0020286-g001){ref-type="fig"}). Genes with the smallest FDR in African-Americans were *ABO, F2RL1,* and *IL17B,* which each had a FDR of 13.5% for Fu and Li\'s D\*. ::: {#pbio-0020286-t001 .table-wrap} Table 1 ::: {.caption} ###### Significant Genes in European-Americans after Correcting for Multiple Tests :::  D, D\*, F\*, and H denote Tajima\'s D, Fu and Li\'s D\*, Fu and Li\'s F\*, and Fay and Wu\'s H, respectively. Nominal *p* values determined from 10^4^ coalescent simulations with recombination are shown in the column next to each statistic. The *p* values that are significant after correcting for multiple tests (FDR = 5%) are shown in bold ::: Distinguishing between Selective and Demographic Forces {#s2c} ------------------------------------------------------- Although neutrality tests of the allele frequency distribution reveal many significant deviations, it is impossible to unambiguously interpret these data as evidence for natural selection, because the null model used to assess significance makes unrealistic assumptions about population demographic history. In principle, it is possible to distinguish between demography and selection, because demography affects all loci in the genome, whereas selection acts upon specific loci. Thus, by sampling a large number of loci dispersed throughout the genome, we can begin to construct a more realistic null hypothesis by which to evaluate the evidence for or against selection ([@pbio-0020286-Kreitman1]). To this end, we used the empirical data to explore four different demographic models ([Figure 2](#pbio-0020286-g002){ref-type="fig"}A), which we could then use to account for demographic influences on tests of natural selection. For each model, we used coalescent theory to simulate data over a broad range of parameters and identified the particular combination of parameters that most closely matched summary statistics (average Tajima\'s D, Fu and Li\'s D\*, and Fu and Li\'s F\*) of the observed data. Of the four demographic models, the European-American data are most consistent with a bottleneck occurring approximately 40,000 y ago, which is nearly identical to a previously reported estimate ([@pbio-0020286-Sabeti1]). However, the confidence intervals for the observed summary statistics are broad, and various aspects of the data are also consistent with other models ([Figure 2](#pbio-0020286-g002){ref-type="fig"}B). The African-American data are most consistent with either an exponential expansion or a relatively old and severe bottleneck ([Figure 2](#pbio-0020286-g002){ref-type="fig"}). Similarly, using DNA sequence variation from ten unlinked, noncoding loci, [@pbio-0020286-Pluzhnikov1] found that an African Hausa sample was consistent with a recent population expansion (although they did not consider bottleneck models). ::: {#pbio-0020286-g002 .fig} Figure 2 ::: {.caption} ###### Summary of the Four Demographic Models Considered in Each Population \(A) Schematic diagram of each demographic model and its associated parameters (see [Materials and Methods](#s4){ref-type="sec"} for details). Parameter values that match the observed data most closely for European-Americans (EA) and African-Americans (AA) are shown below the diagrams. \(B) Average and 95% confidence intervals of Tajima\'s D (blue bars), Fu and Li\'s D\* (red bars), and Fu and Li\'s F\* (pale yellow bars) for the observed data and each demographic model (using the parameters that most closely match the empirical data). Results from the standard neutral model (Constant) are also shown. :::  ::: We reestimated the significance of Tajima\'s D, Fu and Li\'s D\*, Fu and Li\'s F\*, and Fay and Wu\'s H in each population for each of the four demographic models using the best-fit parameter values. All simulations included recombination and correction for multiple tests using the FDR method (with a FDR of 5%) as described above. Population history can clearly have a profound effect on tests of natural selection ([Figure 3](#pbio-0020286-g003){ref-type="fig"}A and [3](#pbio-0020286-g003){ref-type="fig"}B; see also [@pbio-0020286-Simonsen1]; [@pbio-0020286-Przeworski1]), and given the uncertainty in our knowledge of human demographic history, it is challenging to ascribe unusual patterns of genetic variation to either demography or selection. To address this problem, we identified genes whose statistical evidence for selection was robust to demographic history. We conservatively defined demographically robust selection genes as those that demonstrated significant evidence for selection in all five demographic models. We identified eight demographically robust selection genes in European-Americans, and zero in African-Americans ([Figure 3](#pbio-0020286-g003){ref-type="fig"}C; [Table 2](#pbio-0020286-t002){ref-type="table"}). Thus, out of the 22 genes originally found to be significant (at a FDR of 5%) under a standard neutral model, our estimates suggest that demographic history can potentially account for approximately two-thirds of these observations. ::: {#pbio-0020286-g003 .fig} Figure 3 ::: {.caption} ###### The Influence of Demographic History on Tests of Selection (A and B) The significance of observed values of Tajima\'s D (red), Fu and Li\'s D\* (pale yellow), Fu and Li\'s F\* (pale blue), and Fay and Wu\'s H (dark blue) were reassessed for each best-fit demographic model in European-Americans (A) and African-Americans (B). Results from the standard neutral model (Constant) are shown for comparison. The number of significant genes for each demographic model is noted above each category in (A) and (B). For example, there were a total of 19 significant test statistics across all four tests of neutrality assuming a bottleneck model for Europeans, which define ten unique genes. Therefore, each gene is supported by approximately two (19/10) tests of neutrality. \(C) The distribution of the number of significant genes across the five demographic models in European-Americans and African-Americans. For example, in European-Americans, 40 genes were significant in at least one of the demographic models, and 27 genes were significant in at least two of the demographic models. :::  ::: ::: {#pbio-0020286-t002 .table-wrap} Table 2 ::: {.caption} ###### Demographically Robust Selection Genes in European-Americans :::  Biological Process terms were assigned using the Panther classification scheme ([@pbio-0020286-Thomas1]) ::: Evidence for a Recent Selective Sweep on Chromosome 7q in European-Americans {#s2d} ---------------------------------------------------------------------------- One particularly interesting region of the genome is located at 7q and contains four contiguous demographically robust selection genes (*EPHB6, TRPV6, TRPV5,* and *KEL;* [Figure 4](#pbio-0020286-g004){ref-type="fig"}A). Collectively, the entire 115-kb region bears many of the hallmarks of a locus subject to a recent selective sweep: an excess of high-frequency-derived alleles ([Figure 4](#pbio-0020286-g004){ref-type="fig"}B); an overall excess of rare polymorphisms, which results in an extreme skew of the site frequency spectrum reflected by sharply negative values of Tajima\'s D ([Figure 4](#pbio-0020286-g004){ref-type="fig"}C); and a significant reduction in the amount of nucleotide diversity ([Figure 4](#pbio-0020286-g004){ref-type="fig"}D). The signature of positive selection is seen only in European-Americans, suggesting that *EPHB6, TRPV6, TRPV5,* and/or *KEL* possess specific alleles that have conferred local adaptation to a unique environmental pressure in European-derived populations. Consistent with this hypothesis, we observed strong levels of population subdivision ([Figure 4](#pbio-0020286-g004){ref-type="fig"}E) across the entire 115-kb region. The closest genes centromeric to *EPHB6* and telomeric to *KEL* are approximately 42 kb and 64 kb away, respectively, suggesting that one or more of these four genes is the target of selection. However, we have not surveyed patterns of DNA sequence variation outside of the region delimited by *EPHB6* and *KEL,* and therefore it is possible that the signature of selection extends even further. Based on the level of genetic variation on the putatively selected haplotype (see [Materials and Methods](#s4){ref-type="sec"}), we can provide a rough estimate of the time back to the selective sweep as approximately 10,000 y ago. Although this number should be interpreted cautiously, it suggests that selection operated recently. ::: {#pbio-0020286-g004 .fig} Figure 4 ::: {.caption} ###### A Strong Signature of Positive Selection Spanning 115 kb on Chromosome 7q (A--D) Exons for *EPHB6, TRPV6, TRPV5,* and *KEL* are shown as gray vertical lines. A dashed black line indicates the boundary between *EPHB6* and *TRPV6* exons, which are approximately 1 kb apart. Transcriptional orientation is indicated by the arrows below exon positions. SNPs found in European-Americans and African-Americans are shown below. Noncoding, synonymous, and nonsynonymous SNPs are denoted as black, blue, and red vertical bars, respectively. The positions of three nonsynonymous SNPs in *TRPV6* are shown with asterisks. For each of the resulting nonsynonymous amino acid changes, the most frequent amino acid in European-Americans is given first. The frequency of derived alleles, P~D~ (B), sliding window plots of Tajima\'s D (C), and nucleotide diversity, π (D), are shown across the entire region. Gaps in the sliding window plots indicate positions where sequence data were not obtained. In (B--D), European- and African-American data are shown in red and black, respectively. \(E) The distribution of F~ST~ across the 115-kb region. The average F~ST~ for all SNPs across the 132 genes is shown as a dashed red line. The dashed green line indicates the threshold for significantly (*p* \< 0.01) large values of F~ST~, determined by coalescent simulations. :::  ::: **Discussion** {#s3} ============== In summary, we have found that both population demographic history and natural selection shaped patterns of DNA sequence variation in the 132 genes studied here. By studying multiple unlinked loci dispersed throughout the genome, we were able to develop a rigorous computational approach to distinguish between the confounding effects of natural selection and demographic history on patterns of genetic variation. Using this strategy, we found that approximately two-thirds of the genes that were initially significant could be accounted for by population demographic history. Thus, our analyses clearly demonstrate the importance of considering both neutral and nonneutral forces when interpreting DNA sequence variation. An interesting feature of our data is that the majority of deviations from neutrality, and all of the demographically robust selection genes, are not shared between the two population samples, suggesting that local adaptation has played an important role in recent human evolutionary history. Consistent with this observation, several possible examples of local adaptation in humans have previously been reported ([@pbio-0020286-Stephens1]; [@pbio-0020286-Rana1]; [@pbio-0020286-Hollox1]; [@pbio-0020286-Tishkoff2]; [@pbio-0020286-Currat1]; [@pbio-0020286-Fullerton1]; [@pbio-0020286-Gilad1]; [@pbio-0020286-Hamblin1]; [@pbio-0020286-Rockman1]). We hypothesize that the stronger signature of selection in the European-derived population may reflect the exposure of non-African populations to novel and evolutionarily recent selective pressures (e.g., unique dietary, climatic, and cultural environments) as modern humans migrated out of Africa and spread throughout the world. In contrast, the African-derived population may have experienced fewer evolutionarily recent selective forces. Theoretical studies have demonstrated that the power to detect a selective sweep is generally greatest if it occurred less than approximately 0.1 N~e~ generations ago (i.e., approximately 20,000--25,000 y ago \[[@pbio-0020286-Kim1]; [@pbio-0020286-Przeworski1]\]), which is consistent with our hypothesis that signatures of selection in European-Americans reflect recent selective events. However, it is important to note that we have surveyed less than 1% of all human genes, and many of the genes that we did analyze are involved in mediating inflammatory and immune responses; thus our results may not be representative of the genome at large. Interestingly, [@pbio-0020286-Glinka1] found that European-derived populations of Drosophila melanogaster demonstrated abundant evidence for recent selective sweeps, whereas African populations did not, which is strikingly similar to our results in humans. An alternative explanation for why we observed fewer significant results in African-Americans than in European-Americans is that African-Americans are an admixed population ([@pbio-0020286-Parra1]), and the admixture process may mask the signature of selection. However, simulation studies in which we constructed an artificially admixed European-American sample with African-American chromosomes resulted in an increase in significant genes relative to the observed data (unpublished data). Therefore, to the extent that our simulations recapitulate the dynamics of the admixture process in African-Americans, admixture is unlikely to explain the discrepancies between the two samples. It is important to point out that some genes that do not meet our rigorous definition of a high-confidence selection gene may have nonetheless been targets of selection, such as *ABO* in African-Americans ([Table S2](#st002){ref-type="supplementary-material"}). In this initial survey we have elected to be conservative and identify genes that possess the strongest signatures of selection. Ultimately, it will be necessary to confirm our results in geographically diverse populations (a more comprehensive sampling of African populations is particularly needed), as well as in replicate samples of the populations we studied, and to functionally characterize the suspected targets of selection. Recently, [@pbio-0020286-Clark1] presented an evolutionary analysis of 7,645 orthologous human-chimp-mouse gene trios by looking for accelerated rates of synonymous and nonsynonymous nucleotide substitution in either the human or the chimp lineages. In total, 50 genes overlap between our dataset and theirs ([Table S3](#st003){ref-type="supplementary-material"}), including three demographically robust selection genes (*TRPV6, EPHB6,* and *DCN;* see [Table 2](#pbio-0020286-t002){ref-type="table"}). All three of the demographically robust selection genes also demonstrate statistically significant evidence (*p* \< 0.05) of accelerated evolution in either the human *(TRPV6* and *EPHB6)* or chimp *(DCN)* lineage. In addition, [@pbio-0020286-Clark1] found evidence for accelerated evolution in seven genes along the human lineage that did not demonstrate evidence for selection in our dataset ([Table S3](#st003){ref-type="supplementary-material"}). This observation may simply reflect either false negatives in our analysis or false positives in [@pbio-0020286-Clark1]. However, it is important to note that the statistical methods and data used to detect selection in [@pbio-0020286-Clark1] (divergence between species) are quite different from our methods (polymorphism within species), so completely overlapping results are not expected. More specifically, the analyses of [@pbio-0020286-Clark1] will preferentially detect selective events between species, whereas our analyses will preferentially identify selection operating within species. In other words, these two methods are complimentary and may potentially detect selection operating over different time scales. In this respect, it is particularly interesting that the genes we identified as possessing the strongest evidence for recent selection in one human population also show evidence of selection in the human or chimp lineage following their divergence ([@pbio-0020286-Clark1]). The strongest signature of selection that we observed occurs on Chromosome 7q in European-Americans. The signature of selection extends for at least 115 kb and spans the genes *EPHB6, TRPV6, TRPV5,* and *KEL*. To our knowledge, this is the largest footprint of selection that has been described in the human genome, and likely reflects the combination of strong and recent selective pressures and reduced recombination in this region (the average ratio of genetic to physical distance, cM/Mb, is approximately 0.68 according to the deCode map). Based on our current data it is impossible to identify which gene (or perhaps genes) has been the target (or targets) of selection. However, *TRPV6* is a particularly interesting candidate, as it possesses three nonsynonymous amino acid substitutions (C157R, M378V, and M681T) that are each nearly fixed for the derived allele in European-Americans, show significant frequency differences between European-Americans and African-Americans, and are located in the most significant regions of both Tajima\'s D and reduced nucleotide diversity ([Figure 4](#pbio-0020286-g004){ref-type="fig"}). The program PolyPhen ([@pbio-0020286-Ramensky1]) predicts that the C157R replacement may alter protein structure. Recently, TRPV6 was shown to be up-regulated in prostate cancer ([@pbio-0020286-Wissenbach1]), and a susceptibility locus for aggressive prostate cancer was mapped to the *TRPV6* region (7q31--33; [@pbio-0020286-Paiss1]). These observations, combined with the large difference in disease prevalence between Europeans and African-Americans ([@pbio-0020286-Crawford1]), make *TRPV6* a strong candidate gene for prostate cancer susceptibility and/or aggressiveness. *TRPV6,* as well as *TRPV5,* constitute the rate-limiting step in kidney, intestine, and placenta calcium absorption ([@pbio-0020286-Nijenhuis1]; [@pbio-0020286-van1]). Interestingly, Northern European populations have very high frequencies of the lactase persistence allele (*LCT\*P;* [@pbio-0020286-Hollox1]), which allows digestion of fresh milk throughout adulthood. It is widely accepted that strong selection has driven *LCT\*P* to high frequency in Northern Europeans, beginning sometime after the domestication of animals approximately 9,000 y ago ([@pbio-0020286-Feldman1]; [@pbio-0020286-Hollox1]; [@pbio-0020286-Bersaglieri1]). What has been debated, however, is the specific selective advantage conferred by lactase persistence ([@pbio-0020286-Holden1]). Our finding that *TRPV6* and/or *TRPV5* have been under strong selective pressure in Northern Europeans suggests that increased calcium absorption may have been the driving force behind selection for lactase persistence, which was originally hypothesized by [@pbio-0020286-Flatz1]. Although additional studies are clearly needed, our results provide additional insight into the molecular mechanisms of adaptation to a new dietary niche (i.e., high-lactose diets). More generally, our results have several implications for mapping genes underlying complex human diseases. Specifically, four of the high-confidence selection genes have been implicated in various complex diseases ([Table 3](#pbio-0020286-t003){ref-type="table"}). If genes underlying complex diseases have experienced differential selective pressures, then this could in part explain the failure of many studies to replicate disease associations across populations ([@pbio-0020286-Florez1]; [@pbio-0020286-Moore1]). Finally, our data are consistent with the notion that variation in genes that was once beneficial may have become detrimental in the environmental and cultural milieu of contemporary human populations, akin to the "thrifty gene" hypothesis for type II diabetes ([@pbio-0020286-Neel1]). ::: {#pbio-0020286-t003 .table-wrap} Table 3 ::: {.caption} ###### Disease Associations with Demographically Robust Selection Genes :::  ::: Materials and Methods {#s4} ===================== {#s4a} ### **DNA samples and sequencing** {#s4a1} Human DNAs were obtained from the Coriell Institute (Camden, New Jersey, United States). We analyzed DNA from 24 African-Americans from the Human Variation Panel, African-American Panel of 50 (HD50AA) and DNA from 23 European-Americans derived from various CEPH pedigrees. We also sequenced each gene in a common chimpanzee (Pan troglodytes) to determine the derived allele for Fay and Wu\'s H test. These data were generated under the auspices of the SeattleSNPs Program for Genomic Applications, which resequences candidate genes involved in inflammatory processes in humans. In general, we resequenced the complete genomic region for each gene, including introns and approximately 2 kb 5′ of the gene and 1 kb 3′ of the gene using Big-Dye terminator chemistry on an ABI 3700 or ABI 3730XL (Applied Biosystems, Foster City, California, United States). For several exceptionally large genes, such as *F13A1,* less than complete coverage was obtained (see [Table S1](#st001){ref-type="supplementary-material"}). All variants occurring once in the sample were confirmed with an additional sequencing run. Further experimental details and all of the raw data can be found at our website (<http://pga.gs.washington.edu/>). ### **Data analysis** {#s4a2} We calculated the following summary statistics of nucleotide variation for each gene: θ\^= *S/a~n~,* where *S* is the number of segregating sites, and *n* is the sample size ([@pbio-0020286-Watterson1]); , where *h~i~* is an unbiased estimate of nucleotide diversity for the *i*th segregating site (see equation 12 in [@pbio-0020286-Tajima1]) and *η~S~*, which is the number of singletons ([@pbio-0020286-Fu1]). From these statistics we calculated several tests of the standard neutral model including Tajima\'s D ([@pbio-0020286-Tajima1]), Fu and Li\'s D\* ([@pbio-0020286-Fu1]), Fu and Li\'s F\* ([@pbio-0020286-Fu1]), and Fay and Wu\'s H statistic ([@pbio-0020286-Fay1]). In calculating Fu and Li\'s F\*, we used the formulas provided in [@pbio-0020286-Simonsen1], which correct a typographical error in the original description of the method ([@pbio-0020286-Fu1]). For a discussion of the similarities and differences of Tajima\'s D, Fu and Li\'s D\*, Fu and Li\'s F\*, and Fay and Wu\'s H, see [@pbio-0020286-Fu1], [@pbio-0020286-Simonsen1], and [@pbio-0020286-Przeworski1]. We initially assessed the significance of these statistics by comparing the observed values to 10^4^ coalescent simulations ([@pbio-0020286-Hudson1]), conditional on the observed sample size and number of segregating sites, assuming a standard neutral model with no recombination. Coalescent simulations were performed using the program ms (obtained from R. Hudson\'s Web site \[<http://home.uchicago.edu/~rhudson1/source.html>\]). In order to correct for multiple tests, we repeated the coalescent simulations as described above, but included recombination. Following [@pbio-0020286-Pluzhnikov1], for each of the 10^4^ coalescent realizations, we sampled the recombination rate from a Gamma(2, 0.5 × 10^--8^) distribution whose expectation equals the average genome-wide recombination rate of 10^--8^/generation ([@pbio-0020286-Hamblin1]). The positive FDR method was used to correct for multiple hypothesis tests using the software QVALUE ([@pbio-0020286-Storey1]; <http://faculty.washington.edu/~jstorey/qvalue/>). We quantified the allele frequency differences between the European- and African-American samples by the statistic F~ST~ as described in [@pbio-0020286-Akey1]. All of the analyses described above excluded insertion/deletion polymorphisms, but their inclusion does not affect any of our conclusions (unpublished data). We assigned PANTHER Biological Process terms ([@pbio-0020286-Thomas1]) to each gene. We estimated the time since the selective sweep for the Chromosome 7q region in European-Americans by analyzing the amount of nucleotide diversity that has accumulated on the selected haplotype as described in [@pbio-0020286-Rozas1]. We assumed that *TRPV6* is the target of selection and the selected haplotype is defined by the C157R, M378V, and M681T polymorphisms. If mutations are Poisson-distributed, the expected number of segregating sites in a genealogy is *E*\[*S*\] = *μE*\[*T*\], where *S, μ,* and *T* denote segregating sites, neutral mutation rate of the locus, and total branch length of the genealogy, respectively. Assuming a star-shaped genealogy, *E*\[*T*\] = *n* × *t*, where *n* is the number of selected haplotypes. Thus, the time back to the selective sweep, *t,* can be estimated by *S*/(*nμ*). For *TRPV6* in European-Americans, *n* = 45 (i.e., 45 out of 46 haplotypes carry C157, M378, and M681), *S* = 11, and *μ* = 2.5 × 10^--5^. ### **Demographic modeling** {#s4a3} We assessed the impact of demographic history on the robustness of the statistical tests of neutrality by using coalescent theory to simulate data under four different population histories, including a bottleneck, exponential expansion, population structure according to an island model that allows symmetric migration between demes, and population structure assuming population splitting with no subsequent migration. For each model we simulated data under a wide variety of parameters by conditioning on the observed sample size and θ\^~W~ for each population. The bottleneck model is specified by the parameters *F* (the inbreeding coefficient) and *t* (the time in years measured from the present) at which the bottleneck occurred. Values of *F* and *t* considered were *F* = \[0.05, 0.075, ... , 0.40\] and *t* = \[10,000, 20,000, ... , 100,000\]. The exponential expansion model is determined by the parameters *α* (the growth rate/generation) and *t* (the time, in years measured from the present, at which the population began increasing in size). Values considered for *α* and *t* were: *α* = \[0.0005, 0.001, ... , 0.01\] and *t* = \[10,000, 20,000, ... , 100,000\]. The population structure under an island model is specified by the population migration rate between demes, *M* = 4*N~o~m,* where *N~o~* and *m* are the effective subpopulation size and fraction of migrants in each subpopulation per generation, respectively. Values of *M* considered were *M* = \[1, 2, ... , 10\]. The structure model assuming population splitting with no subsequent migration is determined by the parameter *t* (the time in years since the populations diverged). Values of *t* considered were *t* = \[1,000, 2,000, ..., 10,000\]. In all simulations we assumed an effective population size of 10,000 and a generation time of 25 y in order to facilitate comparisons to a previous study ([@pbio-0020286-Sabeti1]). The parameter space for each model included a full grid search, so we tested 160, 100, 10, and 10 parameter combinations for the bottleneck, expansion, structure (island), and structure (splitting) models, respectively. We performed 10^4^ simulations for each parameter combination. For each demographic model, we calculated the average value of Tajima\'s D, Fu and Li\'s D\*, and Fu and Li\'s F\* and compared the results to the observed values of these statistics. For the bottleneck and exponential expansion models, we identified the parameter values that most closely matched the observed data by identifying the parameter combination that minimized the function , where *T~Oi~* and *T~Si~* denote the observed and simulated averages of Tajima\'s D, Fu and Li\'s D\*, and Fu and Li\'s F\*. For the demographic models of population structure we selected parameter values that matched the observed F~ST~. Finally, we reassessed the significance of the observed values of Tajima\'s D, Fu and Li\'s D\*, Fu and Li\'s F\*, and Fay and Wu\'s H by 10^4^ coalescent simulations for each demographic model using the best-fit parameter values. Supporting Information {#s5} ====================== Table S1 ::: {.caption} ###### Summary Statistics of the 132 Genes (266 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S2 ::: {.caption} ###### Neutrality Test Statistics (534 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S3 ::: {.caption} ###### Overlap of Genes Analyzed by [@pbio-0020286-Clark1] (87 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a3} ----------------- LocusLink ID numbers (<http://www.ncbi.nlm.nih.gov/LocusLink/>) for the genes discussed in this paper are *ABO* (28), *ACE2* (59272), *APOH* (350), *BDKRB2* (624), *BF* (629), *C2* (717), *CCR2* (1231), *CD36* (948), *CEBPB* (1051), *CRF* (10882), *CRP* (1401), *CSF2* (1437), *CSF3* (1440), *CSF3R* (1441), *CYP4A11* (1579), *CYP4F2* (8529), *DCN* (1634), *EPHB6* (2051), *F10* (2159), *F11* (2160), *F12* (2161), *F13A1* (2162), *F2* (2147), *F2R* (2149), *F2RL1* (2150), *F2RL2* (2151), *F2RL3* (9002), *F3* (2152), *F5* (2153), *F7* (2155), *F9* (2158), *FGA* (2243), *FGB* (2244), *FGG* (2266), *FGL2* (10875), *FSBP* (10646), *GP1BA* (2811), *ICAM1* (3383), *IFNG* (3458), *IGF2* (3481), *IGF2AS* (51214), *IL10* (3586), *IL10RA* (3587), *IL10RB* (3588), *IL11* (3589), *IL12A* (3592), *IL12B* (3593), *IL13* (3596), *IL15RA* (3601), *IL17B* (27190), *IL19* (29949), *IL1A* (3552), *IL1B* (3553), *IL1R1* (3554), *IL1R2* (7850), *IL1RN* (3557), *IL2* (3558), *IL20* (50604), *IL21R* (50615), *IL22* (50616), *IL24* (11009), *IL2RB* (3560), *IL3* (3562), *IL4* (3565), *IL4R* (3566), *IL5* (3567), *IL6* (3569), *IL8* (3576), *IL9* (3578), *IL9R* (3581), *IRAK4* (51135), *ITGA2* (3673), *ITGA8* (8516), *JAK3* (3718), *KEL* (3792), *KLK1* (3816), *KLKB1* (3818), *KNG* (3827), *LTA* (4049), *LTB* (4050), *MAP3K8* (1326), *MC1R* (4157), *MMP3* (4314), *MMP9* (4318), *NOS3* (4846), *PFC* (5199), *PLAT* (5327), *PLAU* (5328), *PLAUR* (5329), *PLG* (5340), *PON1* (5444), *PON2* (5445), *PPARA* (5465), *PPARG* (5468), *PROC* (5624), *PROCR* (10544), *PROS1* (5627), *PROZ* (8858), *PTGS2* (5743), *SCYA2* (6347), *SELE* (6401), *SELL* (6402), *SELP* (6403), *SELPLG* (6404), *SERPINA5* (5104), *SERPINC1* (462), *SERPINE1* (5054), *SFTPA1* (6435), *SFTPA2* (6436), *SFTPB* (6439), *SFTPC* (6440), *SFTPD* (6441), *SMP1* (23585), *STAT4* (6775), *STAT6* (6778), *TF* (7018), *TFPI* (7035), *TGFB3* (7043), *THBD* (7056), *TIRAP* (114609), *TNF* (7124), *TNFAIP1* (7126), *TNFAIP2* (7127), *TNFAIP3* (7126), *TNFRSF1A* (7132), *TNFRSF1B* (7133), *TRAF6* (7189), *TRPV5* (56302), *TRPV6* (55503), *VCAM1* (7412), *VEGF* (7422), and *VTN* (7448). Coriell (<http://coriell.undmj.edu/>) repository numbers for human genomic DNAs sequenced for this study are as follows. DNAs from African-Americans were NA17101--NA17116 and NA17133--NA17140. DNAs from European-Americans were NA06990, NA07019, NA07348, NA07349, NA10830, NA10831, NA10842--NA10845, NA10848, NA10850--NA10854, NA10857, NA10858, NA10860, NA10861, NA12547, NA12548, and NA12560. We would like to thank members of the SeattleSNPs team (M. Ahearn, T. Armel, E. Calhoun, M. Chung, C. Hastings, P. Keyes, P. Lee, S. Kuldanek, M. Montoya, C. Poel, E. Toth, and N. Rajkumar) for cataloging the variation data. We would also like to thank D. Akey and D. Crawford for critical reading of this manuscript and J. Fay for helpful discussions. This work was supported by a National Science Foundation Postdoctoral Research Fellowship in Interdisciplinary Informatics (JMA) and grants from the National Heart Lung and Blood Institute Program for Genomic Applications (HL66682 to DAN and MJR; HL66642 to LK), the National Institute of Mental Health (MH59520 to LK), and the National Institutes of Health Pharmacogenetics Research Network (U01 HL69757 to DAN). LK is a James S. McDonnell Centennial Fellow. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** MJR, CSC, and DAN conceived and designed the experiments. JMA, MAE, MDS, and LK analyzed the data. JMA, DAN, and LK wrote the paper. Academic Editor: Lon Cardon, University of Oxford Citation: Akey JM, Eberle MA, Rieder MJ, Carlson CS, Shriver MD, et al. (2004) Population history and natural selection shape patterns of genetic variation in 132 genes. PLoS Biol 2(10): e286. FDR : false discovery rate SNP : single nucleotide polymorphism

PubMed Central

2024-06-05T03:55:47.368694

2004-9-7

PMC515368

Introduction {#s1} ============ The von Hippel-Lindau *(VHL)* gene is a tumor suppressor that is mutated in the majority of both hereditary and sporadic, clear-cell renal carcinomas ([@pbio-0020289-Kaelin1]). In hereditary VHL disease affected individuals are also predisposed to pheochromocytomas and retinal/central nervous system hemangioblastomas and develop multiple benign lesions in the kidney and other organs. Despite more than a decade of intensive investigation following identification of the defective gene in 1993 ([@pbio-0020289-Latif1]), the nature of the VHL tumor suppressor mechanism and how it relates to the physiological function of VHL remains unclear ([@pbio-0020289-Kaelin1]). To date, the best-understood function of VHL is as a ubiquitin ligase that affects oxygen-dependent proteolytic targeting of the α subunits of hypoxia-inducible factor (HIF) ([@pbio-0020289-Maxwell1]; [@pbio-0020289-Ohh2]). Oxygen-dependent hydroxylation of two HIF-α prolyl residues by HIF prolyl hydroxylases ([@pbio-0020289-Epstein1]; [@pbio-0020289-Ivan1]; [@pbio-0020289-Jaakkola1]) promotes interaction with VHL and targets HIF-α for degradation by the ubiquitin-proteasome pathway. In VHL-defective cells HIF-α subunits are stabilized and HIF is constitutively activated, resulting in the upregulation of HIF target genes ([@pbio-0020289-Maxwell1]). Whether this, or other putative VHL pathways, accounts for the tumor suppressor action is the subject of active investigation ([@pbio-0020289-Kondo1], [@pbio-0020289-Kondo2]; [@pbio-0020289-Maranchie1]). For instance, a number of different VHL-dependent cellular phenotypes have been defined by contrasting VHL-defective cells with transfectants re-expressing wild-type VHL ([@pbio-0020289-Kaelin1]). These have highlighted effects of VHL on invasiveness, branching morphogenesis, and matrix assembly ([@pbio-0020289-Ohh1]; [@pbio-0020289-Koochekpour1]; [@pbio-0020289-Davidowitz1]; [@pbio-0020289-Kamada1]; [@pbio-0020289-Esteban-Barragan1]). However, mechanistic links to VHL function have not yet been defined and it is unclear whether or not these effects are secondary to dysregulation of HIF. This has led to attempts to define the existence, or otherwise, of non-HIF, VHL-regulated pathways by comparing patterns of gene expression induced by VHL inactivation with those induced by hypoxia ([@pbio-0020289-Wykoff1]; [@pbio-0020289-Zatyka1]; [@pbio-0020289-Jiang3]). The observed patterns are not fully concordant, suggesting that there may be non-HIF, VHL-regulated pathways. However, these studies leave important uncertainties since HIF dysregulation might have secondary effects on pathways that are not themselves responsive to hypoxia and VHL might target hypoxia pathways other than HIF. To address this we have used a genetic approach in *Caenorhabditis elegans.* Whereas mammalian cells possess three HIF-α isoforms that are targeted by VHL, C. elegans has a single HIF-α homolog (HIF-1) and a single VHL homolog (VHL-1), simplifying the genetic approach ([@pbio-0020289-Epstein1]; [@pbio-0020289-Jiang1]). Since homozygous *vhl-1* and *hif-1* loss-of-function worms are viable, we created *hif-1; vhl-1* worms and compared the effects of *vhl-1* inactivation on gene expression in wild-type and HIF-1--defective backgrounds. Our results clearly demonstrate the existence of both HIF-dependent and HIF-independent pathways of VHL-dependent gene expression. HIF-1--dependent effects of *vhl-1* inactivation on gene expression were also produced by inactivation of the HIF prolyl hydroxylase homolog EGL-9. In contrast, the HIF-1--independent effects of *vhl-1* inactivation were not observed in *egl-9* loss-of-function worms but were seen in a panel of mutant worms *(dpy-18, let-268, gon-1, mig-17,* and *unc-6)* bearing defects in genes involved in extracellular matrix function, supporting the existence of a conserved non-HIF pathway connecting VHL with an as yet unknown extracellular matrix function. Results {#s2} ======= Effect of VHL-1 Inactivation on Gene Expression in C. elegans {#s2a} ------------------------------------------------------------- As a first step in defining VHL-1--dependent pathways in *C. elegans,* a whole-genome microarray was probed to compare transcript patterns in *vhl-1* versus wild-type worms (*n* = 1). From this array a set of genes (selected for amplitude of differential expression, signal intensity, quality of array signal, and putative function) was assayed quantitatively by ribonuclease (RNase) protection ([Table 1](#pbio-0020289-t001){ref-type="table"}). Of the 14 genes analyzed, six (F22B5.4, unknown function; *nhr-57,* predicted nuclear hormone receptor; *fmo-12,* predicted flavin monooxygenase; *egl-9,* HIF-1 prolyl hydroxylase \[[@pbio-0020289-Epstein1]\]; *phy-2,* procollagen prolyl 4-hydroxylase α subunit \[[@pbio-0020289-Friedman1]\]; and *cah-4,* predicted carbonic anhydrase) were strikingly downregulated by VHL-1 ([Figure 1](#pbio-0020289-g001){ref-type="fig"}A; [Table 2](#pbio-0020289-t002){ref-type="table"}, column B). Further analysis in synchronized worm populations indicated that the VHL-1--dependent effects were observed in all stages ([Figure 1](#pbio-0020289-g001){ref-type="fig"}B and unpublished data). ::: {#pbio-0020289-g001 .fig} Figure 1 ::: {.caption} ###### HIF-1--Dependent Effects of VHL-1 Inactivation Representative RNase protection assays of genes that were differentially expressed in the *vhl-1* versus wild-type microarray in (A) mixed-stage and (B) synchronized populations of worm. All genes are regulated by the VHL-1/HIF-1/EGL-9 pathway. \(C) Regulation of *nhr-57* mRNA in *egl-9; vhl-1* worms and by *egl-9* RNAi and DIP in *vhl-1* worms. For RNAi experiments controls were L4440 vector alone (−) and C17G10.1, an irrelevant putative dioxygenase. F21C3.5 is a constitutively expressed gene used to control for RNA integrity. RNase protection assays were performed using worms cultured under normoxic conditions, unless otherwise indicated. :::  ::: ::: {#pbio-0020289-t001 .table-wrap} Table 1 ::: {.caption} ###### Top 30 Upregulated Genes in the *vhl-1* versus Wild-Type Microarray Comparison and Confirmation of Selected Genes by RNase Protection Assays :::  Confirmation of selected genes: Y, reproducible upregulation of gene in *vhl-1*/wild-type worms as confirmed by RNase protection assays; N, no reproducible upregulation of gene as tested by RNase protection assays; ---, not determined. Gene name refers to the three-letter gene name where available and open reading frame name (WormBase); details of name changes and primers in arrays are from [http://worm-chip.stanford.edu/pcr.all\_primers.plus\_gels.4--10-02.txt](http://worm-chip.stanford.edu/pcr.all_primers.plus_gels.4–10-02.txt). Description is the predicted protein, annotated by Proteome/Incyte; ---, protein of unknown function. Microarrays and RNase protection assays were performed using worms cultured under normoxic conditions ::: ::: {#pbio-0020289-t002 .table-wrap} Table 2 ::: {.caption} ###### Differential Expression of VHL-1--Regulated Genes in Mutants Affecting the HIF-1/VHL-1/EGL-9 Pathway :::  Column A, data from the microarray comparison of *vhl-1* versus wild-type worms. Columns B to I, data from RNase protection assays. The figures represent the (fold) differences in expression averaged for the indicated number *(n)* of independent comparisons. Statistical analysis of differential expression was performed where *n* ≥ 3; \*, *p* \< 0.05. Statistical analysis was performed for differences between *vhl-1* versus wild-type (column B) and *egl-9* versus wild-type (column C); †, *p* \< 0.05. N, normoxia; H, hypoxia (0.1% oxygen) ::: Analysis of the EGL-9/HIF-1 Pathway {#s2b} ----------------------------------- To determine the extent to which disruption of the conserved EGL-9/HIF-1 pathway mediates these effects we studied wild-type, *hif-1, vhl-1,* and *egl-9* single mutant worms and *hif-1; vhl-1* and *egl-9; hif-1* double mutant worms. Apart from the mild phenotype of the *vhl-1* worms (slightly uncoordinated, slow growth, and reduced brood size) and the egg-laying defective phenotype of *egl-9,* none of the worm strains showed obvious phenotypic abnormalities. Interestingly, *hif-1* corrected the phenotype of *egl-9.* The findings indicate that all six genes are strongly regulated by the EGL-9/HIF-1 pathway ([Figure 1](#pbio-0020289-g001){ref-type="fig"}A; [Table 2](#pbio-0020289-t002){ref-type="table"}). All six genes were inducible by hypoxia in wild-type worms ([Table 2](#pbio-0020289-t002){ref-type="table"}, column D) and strikingly upregulated in *egl-9* worms ([Table 2](#pbio-0020289-t002){ref-type="table"}, column C). *hif-1* inactivation abrogated the upregulation by the *egl-9* mutation ([Table 2](#pbio-0020289-t002){ref-type="table"}, columns C and F) and strikingly reduced induction by hypoxia ([Table 2](#pbio-0020289-t002){ref-type="table"}, columns D and G). Computational analysis revealed that five (F22B5.4, *nhr-57, fmo-12, egl-9,* and *cah-4*) of the six genes contained a potential HIF-1 binding core motif (RCGTG) within an arbitrarily defined region (−1,000 to +250 nucleotides) that was conserved in Caenorhabditis briggsae ([Table 3](#pbio-0020289-t003){ref-type="table"}), suggesting that these genes are direct HIF-1 transcriptional targets. ::: {#pbio-0020289-t003 .table-wrap} Table 3 ::: {.caption} ###### Evolutionarily Conserved HBS Consensus Sequences :::  The sequence column shows alignments between C. elegans and C. briggsae that conserve a match (bold type) to the mammalian HBS consensus motif, RCGTG, where R = A or G. Matches were found on both the sense and antisense strands; matches that are perfect palindromes (CACGTG) can be considered equally good matches to both the sense and antisense strand. The antisense matches are oriented to demonstrate alignment with the consensus motif. The positions of the aligned sequences are shown relative to the translation initiation site. The gene *cah-4* has two alternate first exons (denoted "a" and "b" in WormBase); both were evaluated. In addition to the genes shown, C01B4.7, F56A4.10, C01B4.9, and C01B4.6 were screened for conserved RCGTG motifs, but none was found ::: Though the six genes all conformed to the above patterns to demonstrate regulation by the EGL-9/HIF-1 pathway ([Figure 1](#pbio-0020289-g001){ref-type="fig"}A; [Table 2](#pbio-0020289-t002){ref-type="table"}), there were differences. First, for some genes (F22B5.4, *nhr-57,* and *fmo-12*) expression in normoxia was entirely dependent on HIF-1, whereas other genes retained substantial normoxic expression in *hif-1* worms ([Figure 1](#pbio-0020289-g001){ref-type="fig"}A). Second, three genes *(nhr-57, egl-9,* and *cah-4)* showed modest upregulation, and one gene *(phy-2)* showed modest downregulation, by hypoxia that was independent of HIF-1, VHL-1, and EGL-9 ([Table 2](#pbio-0020289-t002){ref-type="table"}, columns G--I). Finally, for certain genes, upregulation was clearly greater in *egl-9* than *vhl-1* worms, results being particularly striking for *nhr-57* ([Table 2](#pbio-0020289-t002){ref-type="table"}, columns C and B). To pursue this, we created *egl-9; vhl-1* double mutants and also exposed *vhl-1* worms to *egl-9* RNAi. Both procedures increased *nhr-57* expression, indicating that EGL-9 has non--VHL-1--mediated effects on this pathway ([Figure 1](#pbio-0020289-g001){ref-type="fig"}C). Interestingly, the effects of genetic inactivation of *egl-9* in the *vhl-1* background were not mimicked by the dioxygenase inhibitor 2,2′-dipyridyl (DIP), suggesting that the VHL-1--independent repressive effects on *nhr-57* may be nonenzymatic. Evidence for a VHL-1--Dependent, HIF-1--Independent Pathway {#s2c} ----------------------------------------------------------- To address directly whether HIF-1--independent, VHL-1--mediated pathways exist, we performed further microarray comparisons of RNA from *hif-1; vhl-1* and *hif-1* worms (*n* = 3). Fewer genes showed differential expression than in the *vhl-1* versus wild-type array; however, persisting differential expression did suggest the existence of VHL-1 pathways that are independent of HIF-1 ([Table 4](#pbio-0020289-t401){ref-type="table"}). To test this, a number of genes were selected for further validation by RNase protection assay on the basis of amplitude of differential expression, *p* value, signal intensity, and quality of array signal. Of the 25 genes analyzed by RNase protection assay ([Table 4](#pbio-0020289-t401){ref-type="table"}), six (C01B4.7, F56A4.10, C01B4.9, and C01B4.8, all predicted transmembrane proteins belonging to the major facilitator superfamily \[InterPro: IPR007114 and IPR005828\]; F56A4.2, a predicted C-type lectin \[InterPro: IPR001304\]; and C01B4.6, a predicted aldose epimerase \[InterPro: IPR008183\]) showed clear downregulation by VHL-1 in a HIF-1--independent manner ([Figure 2](#pbio-0020289-g002){ref-type="fig"}A; [Table 5](#pbio-0020289-t005){ref-type="table"}, column C). These effects were observed across essentially all developmental stages of the worm ([Figure 2](#pbio-0020289-g002){ref-type="fig"}B). Computational analysis revealed that only one (C01B4.8) of the five HIF-1--independent, VHL-1--dependent genes tested (C01B4.7, F56A4.10, C01B4.9, C01B4.8, and C01B4.6; no single ortholog of F56A4.2 could be identified in C. briggsae) contained a potential HIF-1 binding site (HBS) within an arbitrarily defined region that was conserved in C. briggsae (see [Table 3](#pbio-0020289-t003){ref-type="table"}). This contrasts with the HIF-1--dependent, VHL-1--dependent genes validated by RNase protection assay (see [Figure 1](#pbio-0020289-g001){ref-type="fig"}A), for which potential HBSs could be defined for five of the six genes tested (see [Table 3](#pbio-0020289-t003){ref-type="table"}). ::: {#pbio-0020289-g002 .fig} Figure 2 ::: {.caption} ###### HIF-1--Independent Effects of VHL-1 Inactivation RNase protection assays of genes that were differentially expressed in the *hif-1; vhl-1* versus *hif-1* microarrays in (A) mixed-stage and (B) synchronized populations of worm. The results confirm the existence of VHL-1--dependent, HIF-1--independent effects on gene expression. :::  ::: ::: {#pbio-0020289-t401 .table-wrap} Table 4 ::: {.caption} ###### Upregulated Genes in the *hif-1; vhl-1* versus *hif-1* Microarray Comparisons and Confirmation of Selected Genes by RNase Protection Assays :::  Confirmation of selected genes: Y, reproducible upregulation of gene in *hif-1; vhl-1*/*hif-1* worms as confirmed by RNase protection assays; N, no reproducible upregulation of gene as tested by RNase protection assays; asterisk, not assayed, riboprobe could not be constructed; NS, no signal by RNase protection assay; ---, not determined. Microarrays and RNase protection assays were performed using worms cultured under normoxic conditions. C35B8.1, C46A5.3, R03D7.5, T11F9.8, and ZK1010.7 were also tested by RNase protection assay based on microarray data; these genes did not show reproducible upregulation by RNase protection assays ::: ::: {#pbio-0020289-t005 .table-wrap} Table 5 ::: {.caption} ###### Differential Expression of HIF-1--Independent, VHL-1--Regulated Genes in Mutants Affecting the HIF-1/VHL-1/EGL-9 Pathway and Procollagen Hydroxylases :::  Columns A and B, data from microarray comparisons; columns C to K, data from RNase protection assays. The figures represent the (fold) differences in expression averaged for the indicated number *(n)* of independent comparisons. Statistical analysis of differential expression was performed where *n* ≥ 3; \*, *p* \< 0.05. N, normoxia; H, hypoxia (0.1% oxygen) ::: Interestingly, all six genes validated by RNase protection assay to be negatively regulated by VHL-1 in a HIF-1--independent manner localize within 45 kb on Chromosome V (although they were not situated in physical proximity on the array). We applied single-linkage clustering (nearest-neighbor method) ([@pbio-0020289-Sneath1]; [@pbio-0020289-Dillon1]; [@pbio-0020289-Roy1]) to identify spatial clusters of genes considered to be negatively regulated by VHL-1 in a HIF-1--independent manner from the microarray data ([Table 4](#pbio-0020289-t006){ref-type="table"}) and random sampling to evaluate the significance of such clusters. Using a clustering threshold of 96,985 bp (see [Materials and Methods](#s4){ref-type="sec"}), one cluster of ten genes and four clusters of two genes were identified ([Figure 3](#pbio-0020289-g003){ref-type="fig"}A). On 100,000 simulated datasets of 57 randomly selected genes (equal number to that of VHL-1--dependent, HIF-1--independent genes; [Table 4](#pbio-0020289-t401){ref-type="table"}), the frequency of observed cluster sizes was as follows: one gene, 5,043,442; two genes, 298,425; three genes, 18,198; four genes, 1,190; five genes, 66; six genes, 4. No clusters of more than six genes were observed. Therefore, the cluster of ten VHL-1--regulated (HIF-1--independent) genes, which extends over 110 kb to include F56A4.9, Y45G12C.9, Y45G12C.12, and Y45G12C.2 in addition to the six genes validated by RNase protection assays, can be considered statistically significant to *p* ≪ 10^−5^ ([Figure 3](#pbio-0020289-g003){ref-type="fig"}B). Recent C. elegans genomic assemblies (for example, WS120) have shown that the entire 110-kb region containing the coregulated gene cluster is arranged in tandem with a second nearly identical segmental duplication of the locus (\>99.9% identical in alignment). At this level of identity, our microarray and RNase protection analyses cannot discriminate between the two copies of each gene, so for all of our analyses we have only used the names of the distal copy and genes from the proximal copy were excluded from computational analyses. ::: {#pbio-0020289-g003 .fig} Figure 3 ::: {.caption} ###### Chromosomal Clustering of VHL-1--Dependent (HIF-1--Independent) Genes \(A) Chromosomal localization of VHL-1--dependent, HIF-1--independent genes. The positions of the genes from [Table 4](#pbio-0020289-t401){ref-type="table"} are indicated by vertical ticks along the C. elegans chromosomes (shown to scale). Where two such genes are too close to be clearly resolved, the tick is marked by an asterisk. The single significant spatial clustering of VHL-1--dependent, HIF-1--independent genes is indicated by a red rectangle. The histogram under each chromosome shows the gene density (deeper bar, greater density) calculated as a sliding window of 100,000 bp moving with 10,000-bp increments along each chromosome. Dark blue indicates total annotated gene density, and light blue indicates the density of genes from the microarray that passed preliminary quality control. \(B) Organization of the VHL-1--regulated (HIF-1--independent) gene cluster from Chromosome V. The relative positions and sizes of gene transcription units are shown to scale, with genes transcribed left to right above the horizontal line and right to left below the line. Names in black indicate genes that passed all selection criteria to be considered upregulated in *hif-1; vhl-1* versus *hif-1* worms (see [Table 4](#pbio-0020289-t401){ref-type="table"}). Genes with a mean \>2.0-fold upregulation are indicated by green boxes, 1.5- to 2-fold are yellow, and \<1.5-fold are red. Genes for which no data were obtained are shown as light grey. :::  ::: ::: {#pbio-0020289-t006 .table-wrap} Table 6 ::: {.caption} ###### C. elegans Strains and Alleles :::  ^a^ Note that *eDf18* carries a weak *gon-1* mutation that renders the CB4504 strain temperature sensitive for the Gon phenotype ::: Extracellular Matrix Link to Novel VHL-1 Pathway {#s2d} ------------------------------------------------ Since ubiquitin ligases commonly recognize more than one substrate, we considered whether these HIF-1--independent genes might be regulated by prolyl hydroxylation of another VHL-1 substrate by EGL-9. However, this was not supported by any differential expression in *egl-9; hif-1* versus *hif-1* worms ([Figure 4](#pbio-0020289-g004){ref-type="fig"}A; [Table 5](#pbio-0020289-t005){ref-type="table"}, column E). Nevertheless, two genes, C01B4.7 and C01B4.8, were upregulated in *hif-1* worms by hypoxia and the 2-oxoglutarate dioxygenase inhibitors, DIP and dimethyloxalylglycine (DMOG) ([Figure 4](#pbio-0020289-g004){ref-type="fig"}; [Table 5](#pbio-0020289-t005){ref-type="table"}, column F), suggesting that another enzyme in this class might be involved. The procollagen prolyl hydroxylases DPY-18, PHY-2, and PHY-3 ([@pbio-0020289-Friedman1]; [@pbio-0020289-Riihimaa1]) and the procollagen lysyl hydroxylase LET-268 ([@pbio-0020289-Norman1]) were tested as potential candidates. A clear pattern was observed. All six VHL-1--regulated, HIF-1--independent genes were reproducibly downregulated by DPY-18 and LET-268 but not by PHY-2 or PHY-3 ([Figure 5](#pbio-0020289-g005){ref-type="fig"}A; [Table 5](#pbio-0020289-t005){ref-type="table"}, columns H--K). The strain carrying the heterozygous *let-268* mutation is heterozygous for *unc-4, dpy-10,* and *unc-52;* however, the VHL-1--dependent, HIF-1--independent genes were not differentially expressed in *unc-4, dpy-10,* or *unc-52* worms, indicating that the effects were due to LET-268 (unpublished data). Further experiments on *dpy-18; hif-1* double mutant worms clearly indicated that the effects of DPY-18 on this group of genes were (like the effects of VHL-1) HIF-1 independent ([Figure 5](#pbio-0020289-g005){ref-type="fig"}C and unpublished data). ::: {#pbio-0020289-g004 .fig} Figure 4 ::: {.caption} ###### Responses of VHL-1--Dependent, HIF-1--Independent Genes to *egl-9* Inactivation, Hypoxia, and 2-Oxoglutarate Dioxygenase Inhibitors RNase protection assays showing regulation of VHL-1--dependent, HIF-1--independent genes by (A) EGL-9 and hypoxia and (B) pharmacological inhibitors of 2-oxoglutarate dioxygenases: DIP and DMOG. None of the genes is regulated by EGL-9, but two genes (C01B4.7 and C01B4.8) show modest induction by hypoxia, DIP, and DMOG. :::  ::: ::: {#pbio-0020289-g005 .fig} Figure 5 ::: {.caption} ###### Sensitivity of VHL-1--Regulated Genes to Defects in Extracellular Matrix-Associated Proteins RNase protection assays showing altered expression of VHL-1--regulated genes that are HIF-1 independent (upper six panels) and HIF-1 dependent (F22B5.4) in worms bearing mutations affecting (A) procollagen prolyl and lysyl hydroxylases and (B) other extracellular matrix-associated proteins. A common pattern of upregulation is observed in *hif-1; vhl-1, vhl-1, dpy-18, let-268, gon-1, mig-17,* and *unc-6* worms but not other mutants. This contrasts with the HIF-1--dependent gene F22B5.4, which is upregulated in *vhl-1* worms but none of the other mutants. \(C) RNase protection assay for C01B4.9 illustrating DPY-18--mediated changes in expression that are independent of HIF-1. :::  ::: Downregulation by DPY-18 and LET-268 is consistent with the positive effects of hypoxia, DIP, and DMOG, since all these stimuli inhibit DPY-18 and LET-268. However, the involvement of a lysyl, as well as a prolyl, hydroxylase suggests that the effects were unlikely to arise from failure of hydroxylation of a second prolyl hydroxylation substrate recognized by VHL-1 and were more likely to be related to a common function of DPY-18 and LET-268, such as a function in extracellular matrix formation. To pursue this, we tested the effects of defects in proteins involved in other aspects of extracellular matrix formation (either in the cuticle or basement membrane) that are distinct from protein hydroxylation. These experiments indicated that the six genes were, to varying extents, upregulated in the basement membrane-associated *gon-1* (heterozygote), *mig-17,* and *unc-6* mutant worms but not in the cuticle-associated *dpy-11, bli-4,* or *sqt-3* mutant worms ([Figure 5](#pbio-0020289-g005){ref-type="fig"}B and unpublished data). In contrast, none of the HIF-1--dependent genes was upregulated in these mutants ([Figure 5](#pbio-0020289-g005){ref-type="fig"}B and unpublished data). GON-1 and MIG-17 encode secreted metalloproteases and UNC-6 encodes a netrin; all are thought to be involved in basement membrane remodeling/cell migration during gonadal morphogenesis ([@pbio-0020289-Hedgecock1]; [@pbio-0020289-Blelloch1]; [@pbio-0020289-Nishiwaki1]). Conversely, DPY-11 (a thioredoxin) and BLI-4 (a serine endoprotease) are both involved in collagen formation in the worm cuticle ([@pbio-0020289-Thein1]) and SQT-3 encodes a cuticular collagen. These results therefore extend the characterization of the VHL-1--dependent, HIF-1--independent pathway and support a connection with extracellular matrix/basement membrane function. Discussion {#s3} ========== By comparing the effects of *vhl-1* inactivation in different genetic backgrounds, these data clearly distinguish HIF-1--dependent and --independent effects of VHL-1 on gene expression. Somewhat surprisingly, all of the VHL-regulated genes analyzed fell into one of two patterns: independent of HIF-1 and EGL-9 and dependent on DPY-18, LET-268, GON-1, MIG-17, and UNC-6, or the reverse, suggesting that they reflect perturbation of two discrete aspects of VHL-1 function. The HIF-1--dependent expression pattern of all six genes chosen for detailed analysis from the *vhl-1* versus wild-type array underlines the importance of the HIF-1 pathway in VHL-1 function. Computational analysis revealed that five of these genes (F22B5.4, *nhr-57, fmo-12, egl-9,* and *cah-4*) have at least one HIF-1 binding core motif (RCGTG) that is conserved in C. briggsae within an arbitrarily defined (−1,000 to +250 nucleotides) promoter region, suggesting that they are direct HIF-1 transcriptional targets. Several genes (*egl-9,* HIF prolyl hydroxylase; *phy-2,* procollagen prolyl 4-hydroxylase α subunit; and *cah-4,* carbonic anhydrase) have mammalian homologs that are HIF targets ([@pbio-0020289-Ivanov1]; [@pbio-0020289-Takahashi1]; [@pbio-0020289-Epstein1]), emphasizing the extent of conservation of the pathway. Others, such as flavin monooxygenase *fmo-12* and the nuclear hormone receptor *nhr-57,* are apparently novel HIF-1 target genes. Interestingly, some of these HIF-1--dependent genes were partly downregulated by EGL-9 in a VHL-1-- and iron-independent manner, suggesting that, in addition to the HIF-1/VHL-1 pathway, EGL-9 regulates HIF-1 transcriptional activity via a novel pathway. Remarkably, among the candidate genes tested from the *hif-1; vhl-1* versus *hif-1* screens, all six that showed reproducible (HIF-1--independent) regulation by VHL-1 were located within 45 kb on Chromosome V. Analysis of the microarray data revealed that there was indeed a single, highly significant (*p* \< 10^−5^) chromosomal cluster of genes negatively regulated by VHL-1 in a HIF-1--independent manner and that in total this cluster extended over 110 kb to include F56A4.9, Y45G12C.9, Y45G12C.12, and Y45G12C.2 in addition to the six genes validated by RNase protection assay. The chromosomal localization of genes in C. elegans is not random, with functionally related genes located close to one another ([@pbio-0020289-Roy1]) or even organized into operons ([@pbio-0020289-Blumenthal1]). Even though, based on the absence of spliced leader SL2 sequences ([@pbio-0020289-Blumenthal2]) and the presence of inverse transcriptional orientations, the genes in this cluster do not appear to be within the same operon, there may be a functional relevance to their physical proximity. Four of the six genes validated by RNase protection assay (C01B4.7, F56A4.10, C01B4.9, and C01B4.8) encode membrane transporters of the major facilitator superfamily, a family of transporters involved in passive transport of small solutes. C01B4.9 clusters phylogenetically with monocarboxylate transporters, and C01B4.7, F56A4.10, and C01B4.8 cluster with sodium phosphate transporters (unpublished data). Both gene families have been subject to rounds of gene duplication in the vertebrate and nematode lineages. As such, it is not possible to define one-to-one orthologous relationships for these genes between C. elegans and *Homo sapiens.* Nevertheless, the genomic clustering, predicted functional similarities, and common pattern of perturbed expression across an extensive set of mutant worms suggest that the upregulation of the genes in *vhl-1* worms reflects the disturbance of a specific function of VHL-1. The common effects of inactivating mutations in *vhl-1* and in genes that manifest functional overlap in extracellular matrix formation---*dpy-18* and *let-268* (encoding procollagen hydroxylases) ([@pbio-0020289-Friedman1]; [@pbio-0020289-Norman1]), *gon-1* and *mig-17* (encoding secreted metalloproteases) ([@pbio-0020289-Blelloch1]; [@pbio-0020289-Nishiwaki1]), and *unc-6* (encoding the extracellular guidance protein, netrin) ([@pbio-0020289-Hedgecock1])---suggest a related function for this HIF-1--independent VHL pathway. Interestingly, VHL-defective renal carcinoma cells demonstrate a variety of matrix-related abnormalities, including abnormal fibronectin assembly, defective formation of fibrillar adhesions, and changes in branching morphogenesis and migration ([@pbio-0020289-Ohh1]; [@pbio-0020289-Koochekpour1]; [@pbio-0020289-Davidowitz1]; [@pbio-0020289-Kamada1]; [@pbio-0020289-Esteban-Barragan1]). These abnormalities can be corrected by transfection of renal carcinoma cells with wild-type *vhl,* indicating that they are attributable, either directly or indirectly, to VHL loss of function. Furthermore, immunoprecipitation studies using renal carcinoma cell extracts have indicated that VHL binds to fibronectin ([@pbio-0020289-Ohh1]). Most tumor-associated VHL mutants, when transfected into VHL-defective renal carcinoma cells, are defective in both complementing HIF dysregulation and fibronectin binding ([@pbio-0020289-Kaelin1]). However, mutations associated with type 2C (predisposition to pheochromocytoma only) VHL disease complement defective HIF regulation but bind fibronectin with lower affinity than wild-type VHL ([@pbio-0020289-Hoffman1]). Though the precise link to abnormal matrix assembly remains unclear, this has suggested a HIF-independent function of VHL. The present study supports the existence of a HIF-independent pathway connected with extracellular matrix function and suggests that this may be a highly conserved function of VHL that is potentially amenable to genetic analysis in model organisms. Materials and Methods {#s4} ===================== {#s4a} ### Strains and culturing conditions. {#s4a1} Worms were studied as mixed-stage populations or as synchronized populations following brief exposure to sodium hypochlorite. Exposure to hypoxia (2% or 0.1% oxygen), DIP (200 μM), and DMOG (1 mM) was for 18 h ([@pbio-0020289-Epstein1]). RNA interference (RNAi) was performed by feeding worms Escherichia coli strain HT115(DE3) expressing double-stranded (ds) RNA on Nematode Growth Medium containing 1 mM isopropyl-β-[D]{.smallcaps}-thiogalactopyranoside (ITPG) and 50 μg/ml ampicillin for 72 h. Plasmids for ds RNA production were derivatives of the L4440 vector and were obtained from J. Ahringer (Cambridge, United Kingdom); ds RNA sequences are available on WormBase (<http://www.wormbase.org>). Wild-type worms were Bristol strain (N2); mutant strains were obtained from the *Caenorhabditis* Genetics Center ([Table 6](#pbio-0020289-t006){ref-type="table"}). Strains were maintained at room temperature except for the temperature-sensitive *gon-1* worms (maintained at 18 °C). The double mutants *hif-1; vhl-1, egl-9; hif-1,* and *dpy-18; hif-1* were constructed using either *fog-2* or *unc-51* to mark *hif-1*(+); the double mutant *egl-9; vhl-1* was constructed using *unc-42* to mark *egl-9*(+); PCR was used to confirm homozygosity. ### Microarray screening. {#s4a2} Microarray comparisons of wild-type versus *vhl-1* worms and *hif-1* versus *hif-1; vhl-1* worms were performed on independent samples of RNA (*n* = 1 and 3, respectively), using near full-genome C. elegans DNA microarrays ([@pbio-0020289-Jiang1]). Total RNA was extracted from mixed-stage populations of worm cultured under normoxic conditions using Tri-reagent (Sigma, Poole, Dorset, United Kingdom) and mRNA purified using oligo-dT beads (Qiagen, Crawley, West Sussex, United Kingdom). cDNA synthesis and microarray hybridization and scanning were performed as described previously ([@pbio-0020289-Jiang1]). Cy5-dUTP was used to label cDNA from wild-type and *hif-1* worms and Cy3-dUTP was used to label cDNA from *vhl-1* and *hif-1; vhl-1* worms. The arrays were computer normalized by the default procedure in the Stanford Microarray Database (SMD); primary array data are available on the SMD (<http://genome-www.stanford.edu/microarray>) and are also shown in [Tables S1 through S4](#st001){ref-type="supplementary-material"}. Fold change was calculated as the ratio of the means of Cy3-dUTP intensity to normalized Cy5-dUTP intensity (normalized to correct for signal differences between Cy3-dUTP and Cy5-dUTP intensities across the whole array) with median background intensities subtracted from both signal intensities to correct for the background (see SMD). Genes with background-corrected signal intensities below zero or with array spots that were flagged in the SMD as being unreliable were discarded as a preliminary quality control. For the *hif-1* versus *hif-1; vhl-1* microarray comparisons (*n* = 3) the log~2~ fold change was calculated as the mean of the three log~2~ transformed fold changes. To test for significant upregulation, the mean log~2~ fold change was compared with zero using a Student\'s *t* test. The genes were ranked by amplitude of fold upregulation and a subset of genes was selected for potential validation by RNase protection assays (see [Tables 1](#pbio-0020289-t001){ref-type="table"} and [4](#pbio-0020289-t401){ref-type="table"}) based on the following criteria: (a) *t* test, *p* \< 0.10 (for the *hif-1* versus *hif-1; vhl-1* microarray comparisons, *n* = 3); (b) mean Cy3-dUTP and Cy5-dUTP background-corrected signal intensities exceeding 300 and 100 U, respectively (lower intensities than these were difficult to detect by RNase protection assay); and (c) high spot quality as judged by manual inspection. For the *hif-1* versus *hif-1; vhl-1* microarray comparisons (*n* = 3), genes (which had been filtered as described above) were considered to be differentially expressed if the mean fold change was greater than 2.0 (see [Table 4](#pbio-0020289-t401){ref-type="table"}). ### RNase protection assays. {#s4a3} Assays were performed on total RNA from mixed-stage populations of worm cultured under normoxic conditions, unless otherwise indicated. Details of riboprobe templates are provided in [Table 7](#pbio-0020289-t007){ref-type="table"}; details of genes tested are shown in [Tables 1](#pbio-0020289-t001){ref-type="table"} and [4](#pbio-0020289-t401){ref-type="table"}. Quantification was performed using a phosphorimager (Molecular Dynamics, Sunnyvale, California, United States) and related to an internal control assay for the constitutively expressed F21C3.5 (protein with similarity to mouse prefoldin subunit 6). Where *n* ≥ 3, the log~2~ fold change was calculated from the mean of the log~2~ transformed fold changes and statistical significance was calculated by comparing the mean log~2~ fold change with zero using a Student\'s *t* test. ::: {#pbio-0020289-t007 .table-wrap} Table 7 ::: {.caption} ###### Sequence and Length of Riboprobes :::  ^a^ Note that the protected region of the *egl-9* transcript does not overlap the *sa307* deletion in the JT307 *egl-9* strain ::: ### Computational analyses {#s4a4} \(1) Identification of potential HBSs (see [Table 3](#pbio-0020289-t003){ref-type="table"}). Orthologs of C. elegans genes were identified in the C. briggsae genome assembly (cb25) as reciprocal best matches by BLASTN, initiated with the C. elegans gene coding sequence (a single ortholog of F56A4.2 could not be defined). Translation initiation sites (well-annotated surrogates for transcriptional start sites; none of these genes are annotated as having spliced 5′ UTRs) were inferred in both C. briggsae and C. elegans through alignment with the C. elegans coding sequence (WormBase, WS117). Sequences encompassing the 1,000 nucleotides upstream to 250 nucleotides downstream of the translation start sites for orthologous genes were aligned using DNA Block Aligner ([@pbio-0020289-Jareborg1]) with the following options: gap = 0.001 and blockopen = 0.005. Sequence alignments were searched with the HBS motif RCGTG ([@pbio-0020289-Camenisch1]), identifying cases where HBS-like motifs were conserved between both C. elegans and *C. briggsae.* \(2) Single-linkage analysis to determine spatial clusters of VHL-1--dependent, HIF-1--independent genes (see [Figure 3](#pbio-0020289-g003){ref-type="fig"}). A maximum distance for the linking of two clusters was determined by ranking the distance between 10,000 randomly selected pairs of genes from the same chromosome (but sampled over all six nuclear chromosomes) and selecting as a threshold the first percentile of the distribution, 96,985 bp. Intergene distances were calculated from the closest point between the annotated coding sequence of each gene; genes on separate chromosomes were considered to have an infinite intergene distance. Simulations were performed using genes selected at random from genes that were represented on the microarray and that passed preliminary quality control criteria. All genomic coordinates were based on genomic assembly WS120 and the associated WormBase annotation of genes obtained from the University of California, Santa Cruz Genome Browser (<http://genome.ucsc.edu/>). The software used for simulations and clustering was implemented in Perl and is available on request. Supporting Information {#s5} ====================== Primary microarray data can be viewed at <http://genome-www.stanford.edu/microarray>. Table S1 ::: {.caption} ###### *vhl-1* versus Wild-Type Microarray Comparison Primary microarray data for the *vhl-1* (green, channel 1) versus wild-type (red, channel 2) comparison. (6.5 MB XLS). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S2 ::: {.caption} ###### *hif-1; vhl-1* versus *hif-1* Microarray Comparisons I Primary microarray data for the three independent *hif-1; vhl-1* (green, channel 1) versus *hif-1* (red, channel 2) microarray comparisons. Continued in Tables [S3](#st003){ref-type="supplementary-material"} and [S4](#st004){ref-type="supplementary-material"}. (6.6 MB XLS). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S3 ::: {.caption} ###### *hif-1; vhl-1* versus *hif-1* Microarray Comparisons II Continuation of [Table S2](#st002){ref-type="supplementary-material"}. (6.6 MB XLS). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S4 ::: {.caption} ###### *hif-1; vhl-1* versus *hif-1* Microarray Comparisons III Continuation of Tables [S2](#st002){ref-type="supplementary-material"} and [S3](#st003){ref-type="supplementary-material"}. (6.6 MB XLS). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a5} ----------------- Primary array data have been deposited in ArrayExpress (<http://www.ebi.ac.uk/arrayexpress/>) under accession number E-SMDB-23. The *H. sapiens VHL* gene discussed in this paper can be found in Online Mendelian Inheritance in Man (OMIM) under accession number 608537 (<http://www.ncbi.nlm.nih.gov:80/entrez/dispomim.cgi?id=608537>). The C. elegans genes discussed in this paper (*bli-4,* C01B4.6, C01B4.7, C01B4.8, C01B4.9, *cah-4, dpy-10, dpy-11, dpy-18, egl-9,* F21C3.5, F22B5.4, F56A4.2, F56A4.9, F56A4.10, *fmo-12, fog-2, gon-1, hif-1, let-268, mig-17, nhr-57, phy-2, phy-3, sqt-3, unc-4, unc-6, unc-42, unc-51, unc-52, vhl-1,* Y45G12C.2, Y45G12C.9, and Y45G12C.12) can be found in the WormBase database by including the name of the gene at the end of the URL (e.g., for *bli-4,* <http://wormbase.org/db/gene/gene?name=bli-4>). ::: {#pbio-0020289-t402 .table-wrap} Table 4 ::: {.caption} ###### Continued :::  ::: The authors would like to thank Richard Mott and James Lund for helpful discussions. This work was supported by the Wellcome Trust, the British Heart Foundation, and the Medical Research Council (United Kingdom); the Agency for Science, Technology and Research (Singapore); and the National Center for Research Resources (United States of America). **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** TB, KWL, ACRE, CWP, JH, and PJR conceived and designed the experiments. TB, KWL, ACRE, MJ, and DOR performed the experiments. TB, KWL, CWP, JMG, MST, JH, and PJR analyzed the data. MST designed and implemented the computational analyses. SKK, JH, and PJR contributed reagents/materials/analysis tools. TB and PJR wrote the paper. Academic Editor: Christopher Kemp, Fred Hutchinson Cancer Research Center Citation: Bishop T, Lau KW, Epstein ACR, Kim SK, Jiang M, et al. (2004) Genetic analysis of pathways regulated by the von Hippel-Lindau tumor suppressor in *Caenorhabditis elegans.* PLoS Biol 2(10): e289. DIP : 2,2′-dipyridyl DMOG : dimethyloxalylglycine HBS : HIF-1 binding site HIF : hypoxia-inducible factor RNase : ribonuclease SMD : Stanford Microarray Database VHL : von Hippel-Lindau

PubMed Central

2024-06-05T03:55:47.372571

2004-9-7

PMC515369

Introduction {#s1} ============ Phylogeography has advanced our understanding of the spatial distribution of genetic diversity within and between species ([@pbio-0020290-Avise1]). However, empirical evidence of temporal change in genetic diversity in a single locality over time has not yet been placed in a population genetic or phylogeographic framework over ecologically long periods of time. In this paper we attempt to determine variation in genetic diversity experienced by populations of two mammalian species in situ and to place that diversity in the context of a changing environment through time. We view this approach as "phylochronology," or the study of populations in space and time using phylogenetic and population genetic methods. Similar studies have not used such a long temporal record ([@pbio-0020290-Pergams1]), have not considered gene flow ([@pbio-0020290-Lambert1]; [@pbio-0020290-Pergams1]), or have used a spatially averaged sample as a proxy for a single locality ([@pbio-0020290-Leonard1]). Our study takes advantage of a continuous, well-sampled mammalian fossil sequence spanning the last 3,000 years (Lamar Cave, Yellowstone National Park, Wyoming, United States). Lamar Cave has an extraordinarily complete representation of the local species in the vicinity, with over 10,000 identified mammalian specimens representing over 80% of the mammal species in the local habitat ([@pbio-0020290-Hadly3]). Late-Holocene climatic change, including the Medieval Warm Period (1,150 to 650 years before present \[ybp\]) and Little Ice Age (650 to 50 ybp) ([@pbio-0020290-Soon1]), affected the local abundances of common small mammals in a manner consistent with their habitat preferences ([@pbio-0020290-Hadly1]). We focus on two mesic habitat specialists, Microtus montanus (montane vole) and Thomomys talpoides (northern pocket gopher), species that presently are widespread in mountain habitats of western North America. Due to their preferences for wetter habitats, both responded demographically by increasing in relative abundance during wetter climates and declining during warmer climates. M. montanus showed an increase in abundance relative to other common rodents during periods of wet, cool climate in Yellowstone ([@pbio-0020290-Hadly1]). A 40% decline in M. montanus abundance occurred from 2,525 ybp to about 470 ybp during the Medieval Warm Period. Because T. talpoides also demonstrates a preference for mesic montane conditions, shifts in their relative abundance mimic the response seen in *Microtus,* decreasing by 50% between 2,525 and 470 ybp. In addition, T. talpoides showed a significant reduction in body size during this time ([@pbio-0020290-Hadly2]). These data highlight the influence of climatic change on the population dynamics and phenotypic response of these species, especially during warming events. Although the population responses of T. talpoides and M. montanus are similar, the ways in which they respond to climatic change at the genetic level are predicted to diverge because of differences in dispersal ability and population substructure. Advancement of ancient DNA (aDNA) techniques allows us to investigate directly the impacts of these environmental perturbations on neutral genetic diversity concurrent with these species population responses. We obtained ancient and modern mitochondrial DNA sequences from M. montanus and used previously published data for *T. talpoides.* Although these two species are broadly similar in body size (*M. montanus,* 50--100 g; *T. talpoides,* 75--150 g) and are principally herbivorous, they differ in their natural history. T. talpoides is characterized by low population densities (1--62 gophers/hectare \[ha\]), a fossorial mode of life, maximum dispersal distances of a few hundred meters, and fiercely territorial behavior ([@pbio-0020290-Verts1]). Populations of T. talpoides from Lamar Cave exhibit very little genetic variation through time but considerable genetic differences between present localities ([@pbio-0020290-Hadly5]). This spatiotemporal pattern suggests that late-Holocene gene flow did not influence modern genetic variation of T. talpoides within localities over relatively short time scales (hundreds to thousands of years) despite the absence of obvious migration barriers. M. montanus achieves higher average population densities (60--186 voles/ha) ([@pbio-0020290-Sullivan1]) than *T. talpoides.* In addition, genetic studies of closely related species and other arvicoline species *(Microtus pennsylvanicus, Microtus longicaudus,* and *Microtus agrestis)* have found little evidence for population subdivision over the scale of hundreds of kilometers ([@pbio-0020290-Plante1]; [@pbio-0020290-Conroy2]; [@pbio-0020290-Bjornstadt1]; [@pbio-0020290-Jaarola1]), as expected from the ability and proclivity of voles to disperse hundreds to thousands of meters, resulting in migration between populations on generational time scales ([@pbio-0020290-Jenkins1]; [@pbio-0020290-Lidicker1]). Such demography, however, implies that historical gene flow may be difficult to detect in M. montanus if only genetic data from the modern animals are used. This is contrary to the pattern expected in T. talpoides because this species has high genetic differentiation between extant populations, and past movement between such populations would be relatively easy to ascertain by the historic presence of unique, divergent haplotypes ([@pbio-0020290-Hadly5]). The primary advantage of a phylochronologic approach as opposed to a single time slice for understanding mammalian response is the ability to reveal changes in genetic variation through time. This is in contrast to modern genetic studies that seek to reconstruct demographic history based on inferences from past climate or geologic records (e.g., [@pbio-0020290-Storz1]; [@pbio-0020290-Lessa1]). Our study separates demographic and genetic response explicitly, allowing us to understand the microevolutionary forces responsible for the differences in species response over a time scale relevant to evolution within species. This approach is particularly powerful when coupled with environmental data so that perturbations may be linked to organismal response. In order to reveal these microevolutionary forces from our serial ancient data, it was necessary to explore the influence of sampling from the fossil record and to determine how variation in stochastic evolutionary forces (gene flow, drift, and mutation rate) might influence the record of gene diversity over time. Thus, we combined four methods of estimating population size, determining statistical significance, and assessing gene diversity through time. (1) We derived independent ecological estimates of population size through time from abundances of M. montanus and T. talpoides fossil specimens and modern population densities. (2) We calculated gene diversity over this 3,000-year period to determine the impact of environmental perturbations on genetic effective population size and gene diversity in M. montanus and *T. talpoides.* Unlike previous aDNA work ([@pbio-0020290-Consuegra1]; [@pbio-0020290-Hofreiter1]; [@pbio-0020290-Lambert1]; [@pbio-0020290-Orlando1]; [@pbio-0020290-Paxinos1]), we used mitochondrial DNA sequences for samples taken through time from a single locality. (3) While there have been advances in the use of nuclear markers for ancient genetic analyses, we also confine our analyses to more easily derived mitochondrial DNA (mtDNA) data, thus limiting our analyses to a single locus, usually seen as a neutral marker within mammalian species ([@pbio-0020290-Moritz1]). Our approach also constrains us to the fossil sample sizes from this locality, which are extremely large for ancient DNA studies, but limited relative to population genetic studies. Thus, we assessed the statistical power of a single locus for our empirical data using a neutral population model. (4) We used a neutral population model and serial coalescent simulations to determine whether our observed genetic data reflect our ecological estimates of population size and to evaluate statistical significance in changes of gene diversity through time. Despite similar population-level responses to climatic change of the late Holocene, we expected differences in gene diversity change for the two species. For *T. talpoides,* we predicted that changes in genetic variation through time would be dominated by drift, as suggested by the modern life history characteristics of small effective population size, low dispersal, and high amounts of population substructure. Therefore, as the ecological effective population size of T. talpoides declined with warmer climates, we expected genetic variation to decline. For *M. montanus,* we predicted that changes in genetic variation through time may be influenced more by migration, as suggested by large effective population sizes, high rates of dispersal, and low amounts of population substructure. As a result, past declines in ecological estimates of population size of M. montanus would not necessarily have resulted in a decrease in genetic variation. Results {#s2} ======= Fossil Abundance {#s2a} ---------------- Our assessment of population response to climatic change ([Figure 1](#pbio-0020290-g001){ref-type="fig"}) depends on reconstruction of population size. Fossil relative abundances give a hint of the census size through time while genetic data (gene diversity, [Figure 1](#pbio-0020290-g001){ref-type="fig"}) should yield independent assessments of the effective population size. The relationship between these measures varies, although most studies suggest that the estimate of effective size derived from ecological data is higher than that derived from genetic data*(N~e\_ecol~* \>\> *N~e\_gen~)* ([@pbio-0020290-Frankham1]; [@pbio-0020290-Kalinowski1]). However, we can convert census size estimates at any point in time into effective size estimates and vice versa. This allows us to compare explicitly ecological and genetic measures of population size. [Figure 2](#pbio-0020290-g002){ref-type="fig"} shows *N~e\_ecol~* estimates based on low-, high-, and moderate-density estimates for T. talpoides ([Figure 2](#pbio-0020290-g002){ref-type="fig"}A) and M. montanus ([Figure 2](#pbio-0020290-g002){ref-type="fig"}B). For M. montanus they range from 218,652 to 436,981 for low-density estimates and from 677,825 to 1,354,650 for high-density estimates. For T. talpoides low-density estimates range from 2,219 to 5,015 and high-density estimates range from 4,586 to 10,488 individuals in the 7-km radius around Lamar Cave. ::: {#pbio-0020290-g001 .fig} Figure 1 ::: {.caption} ###### Proportional Population Size and Gene Diversity of M. montanus and T. talpoides \(A) Proportional population size (relative abundance) of M. montanus and T. talpoides (*n* = 8,589 fossils) by years before present. \(B) Gene diversity (H) of M. montanus and T. talpoides by years before present; 95% confidence intervals are shown. Squares indicate *M. montanus;* triangles indicate *T. talpoides.* :::  ::: ::: {#pbio-0020290-g002 .fig} Figure 2 ::: {.caption} ###### Estimates of *N~e\_gen~* and *N~e\_ecol~* \(A) T. talpoides and (B) M. montanus through time. Circles and dashed lines show *N~e\_ecol~* estimates based on low-, high-, and moderate-density estimates. *N~e\_gen~* estimates (\[A\] triangles and \[B\] rectangles) are based on *θ~S~* estimates from Arlequin. Standard errors for *N~e\_gen~* are represented. :::  ::: Genetic Data: M. montanus {#s2b} ------------------------- The genetic evidence we have assembled from M. montanus suggests that the sequences we obtained for this study are target mtDNA. Of the 312 bp we sequenced for *M. montanus,* 96.5% of all the mutations were third-position codon changes, with first- and second-position mutations accounting for 3.5% and 0%, respectively. These ratios of variation are concordant with expectations for within-species variation and small overall sequence divergences ([@pbio-0020290-Yang1]). Nucleotide base composition is similar to that of other *Microtus* species ([@pbio-0020290-Conroy2]; [@pbio-0020290-Jaarola1]), with an excess of adenine (31.2%) and a deficit of guanine (15.7%) (χ^2^; α = 0.71). Most of the mutations are synonymous (97.7%); the transition-to-transversion ratio of the entire data set was 4.1 to 1, which is consistent with expectations for mammalian cytochrome *b* and evolution in other *Microtus* species ([@pbio-0020290-Conroy2]; [@pbio-0020290-Jaarola1]). Fossil and modern transition-to-transversion ratios are similar (3.1 and 4.2, respectively). All M. montanus sequences are reciprocally monophyletic (including M. pennsylvanicus as outgroup taxon) and translated successfully. Together with the frequency distribution of our pairwise differences, the prevalence of silent and third-position codon changes, and the standard of obtaining both forward and reverse fragments of overlapping sequence regions, these data permit us to conclude that the genetic diversity we have sampled represents authentic mitochondrial population variation and is unlikely to be from nuclear copies or pseudogenes. A total of 282 experiments included 47 fossil extractions and 1,644 PCRs. Eighty-eight percent of our aDNA specimens yielded readable sequence data, with no relationship found between success rate and age of the specimen (*R* ^2^ = 0.004, not significant). All but one (out of 121) of the extraction controls were negative. When sequenced, this extraction blank BLASTed similar to *Montanus townsendii,* a taxon we had never worked on in the facility; this sequence has not since been amplified in the lab, and that extraction was not used further. Out of 87 successfully amplified samples and one sequence obtained from GenBank (AF119280), we identified 17 haplotypes within four haplogroups *(A--D)* of M. montanus ([Figure 3](#pbio-0020290-g003){ref-type="fig"}A). The distribution of haplotypes within haplogroups suggests that our groups are defined appropriately. Each haplogroup was defined by at least 3% sequence divergence (≥10 bp) from other haplogroups in the 312-bp cytochrome *b* fragment. The majority of individuals (98.8%) fall within haplogroups *A* and *D,* with 84% of the samples within one substitution of the locally ancestral haplotype *A* ([Figure 3](#pbio-0020290-g003){ref-type="fig"}A). ::: {#pbio-0020290-g003 .fig} Figure 3 ::: {.caption} ###### Haplotype Networks for M. montanus and T. talpoides Haplotype networks ([@pbio-0020290-Clement1]) for (A) M. montanus and (B) T. talpoides from Lamar Cave fossils and from modern specimens collected within a 400-km radius of Lamar Cave. Haplogroups for both species are indicated as *A--D.* Each haplogroup within a species is defined by at least 3% sequence divergence within the cytochrome *b* fragment. M. montanus haplogroup *B* is taken from GenBank. Haplogroup *C* is a sample from outside our 400-km radius (NK5897, Mono County, California; Museum of Southwestern Biology \#53376). Light shading shows modern samples; dark shading shows fossil samples; bars indicate substitutions; cytochrome *b* sequence positions are indicated by number above base designation. Numbers within parentheses indicate sample sizes for each haplotype. :::  ::: The maximum uncorrected sequence divergence for our complete spatial and temporal data set was 4.5%, demonstrated between haplogroups *A* and *B.* Given that the highest average rodent divergence rate for cytochrome *b* is 6% to 10% per million years ([@pbio-0020290-Irwin1]), these haplogroups have been evolving separately for at least 450,000 years. A similar age (422,000 years) is found when using a rate of 2.3% per million years for third-position transversions ([@pbio-0020290-Conroy1]). The maximum uncorrected sequence divergence of M. montanus from throughout the Lamar Cave temporal sequence was 4.2%. The maximum sequence divergence of 19 modern individuals from populations of this species within Yellowstone National Park and surroundings was 3.8%. Genetic Data: T. talpoides {#s2c} -------------------------- Protocols for T. talpoides are found in [@pbio-0020290-Hadly5]. A haplotype network of this species shows three haplogroups and a total of eight haplotypes from 76 specimens ([Figure 3](#pbio-0020290-g003){ref-type="fig"}B). For *T. talpoides,* 98.0% of all the mutations were third-position codon changes, with first- and second-position mutations accounting for 0% and 2.0%, respectively. Nucleotide base composition shows an excess of thymine (34.7%) and a deficit of guanine (11.4%). All mutations were synonymous; the transition-to-transversion ratio of the entire data set was 4.0 to 1. Gene Diversity through Time {#s2d} --------------------------- Estimates of gene diversity, nucleotide diversity, and number of segregating sites differed between M. montanus and T. talpoides ([Table 1](#pbio-0020290-t001){ref-type="table"}). The estimates for M. montanus were higher than those for *T. talpoides,* as predicted by life history traits including higher ecological effective population size and higher dispersal between populations. ::: {#pbio-0020290-t001 .table-wrap} Table 1 ::: {.caption} ###### Summary Statistics, Sample Sizes, Sequence Length, and Number of Haplotypes for the Ancient DNA Samples for T. talpoides and M. montanus from Lamar Cave, Wyoming :::  Gene diversity, number of segregating sites, and nucleotide diversity for each time interval were calculated with Arlequin v. 2.000 ([@pbio-0020290-Jaarola2]; [@pbio-0020290-Schneider1]) ::: Our raw data on these species show similar relative abundance patterns but disparate trends in gene diversity (see [Figure 1](#pbio-0020290-g001){ref-type="fig"}). These patterns cannot be linked directly to relative abundance because gene diversity estimates depend on true population size as well as sampling. We investigate how both of these parameters impact the observed trend in the following sections. Comparing Ecological and Genetic Estimates of Effective Size {#s2e} ------------------------------------------------------------ Ecological estimates of population size for *T. talpoides (N~e-tt\_ecol~)* exhibit the same trend as the genetic estimates *(N~e-tt\_gen~),* namely a population size decline of \>50% after 1,500 ybp (see [Figure 2](#pbio-0020290-g002){ref-type="fig"}A). For *T. talpoides, N~e-tt\_gen~* is consistently higher than *N~e-tt\_ecol~.* When compared to those for M. montanus (see [Figure 2](#pbio-0020290-g002){ref-type="fig"}B), the estimates of total ecological and genetic effective population size are much lower. Additionally, unlike with *M. montanus,* the ecological and genetic effective sizes follow similar trends through the entire time period sampled, indicating that T. talpoides is acting as a closed population. For *M. montanus,* the estimates of effective size derived from the ecological data *(N~e-mm\_ecol~)* are lower than those derived from genetic data *(N~e-mm\_gen~)* for all time points (see [Figure 2](#pbio-0020290-g002){ref-type="fig"}B). While the estimates are not expected to be identical, comparison of their trends is instructive. Both genetic and ecological estimates follow similar trends between 2,525 and 845 ybp, after which the two estimates follow opposite trajectories. While the ecological size decreases by 50%, the genetic estimates show an initial decline of 30%, followed by an increase in population size equivalent to the pre--1,438-ybp level. This demonstrates that although the population is not recovering ecologically, it does recover genetically from population decline between 1,438 and 845 ybp. Effects of Sampling {#s2f} ------------------- For *M. montanus,* our observed data were within the 95% confidence intervals for both sets of simulations (*n~sample~* and *n* = 100; mutation rate = 4% per million years; moderate density values used to calculate abundance) for all time points except for 2,525 ybp, where observed gene diversity was significantly lower than could have been calculated given our sample size ([Figure 4](#pbio-0020290-g004){ref-type="fig"}). Since the observed gene diversity is lower than expected, we repeated simulations for five additional combinations of mutation rate and abundance (low mutation rate, high abundance; low mutation rate, moderate abundance; low mutation rate, low abundance; moderate mutation rate, low abundance; high mutation rate, low abundance). Results revealed that the observed gene diversity at 2,525 ybp was within the lower fifth percentile of the predicted distribution for two of the five combinations (when both mutation rate and abundance were low and for low mutation rate, moderate abundance). The overlap between the 95% confidence intervals for both sets of simulations suggests that sampling bias does not significantly impact the observed patterns of gene diversity except at 1,438 ybp (*n* = 4), suggesting that we do not have sufficient power to detect processes at this time period. ::: {#pbio-0020290-g004 .fig} Figure 4 ::: {.caption} ###### Expected and Observed Gene Diversity of M. montanus Boxes represent the 95th, 50th, and fifth percentiles for expected gene diversity of M. montanus given estimates of *N~e-mm\_ecol~* at 2,525, 1,438, 845, 470, and 166 ybp and the associated sample sizes (*n* = 7, 7, 18, 4, and 6) based on the Ewens sampling distribution (assumed mutation rate = 4% per million years per bp for a 312-bp fragment). Bars represent the 95th and fifth percentiles for a sample size of 100 at the same points in time. Diamonds represent observed gene diversity from empirical genetic data. The empirical data for each time unit fall within the expected ranges of gene diversity, except those for 2,525 ybp, which are much too low for the seven samples to detect, suggesting that observed gene diversities are not limited by sample size. :::  ::: Although the gene diversity for T. talpoides is not different given expectations from a closed population, we attempted to determine the statistical limitations of these data. Investigation of the effects of sampling for T. talpoides revealed that given the smaller number of base pairs (64 bp), we do not have enough statistical power to reject the null hypothesis. For every sampling time point, the Ewens distribution predicted that only one haplotype would be present in the genetic samples (unpublished data). As a result, the predicted gene diversity was zero for all time points. Because the observed gene diversity for T. talpoides was higher than predicted we also simulated five combinations of mutation rate and abundance, which could result in a higher predicted diversity (high mutation rate, high abundance; high mutation rate, moderate abundance; high mutation rate, low abundance; moderate mutation rate, high abundance; low mutation rate, high abundance) for 166 ybp (*n* = 34). Results for all five combinations predicted presence of a single haplotype. Since we do not have adequate statistical power given the genetic data for *T. talpoides,* we did not conduct significance tests for this species. However, the observed values of gene diversity are not unexpected from dynamics within a closed population. Significance of Changes in Gene Diversity in M. montanus {#s2g} -------------------------------------------------------- For eight of the nine combinations of mutation rate and effective size, we could reject the null hypothesis (closed population, no selection, changes in abundance inferred through fossil abundance) based on the observed change in M. montanus gene diversity throughout the entire time series (2,525 to 166 ybp) ([Table 2](#pbio-0020290-t002){ref-type="table"}). The expected distribution of change in gene diversity given the null hypothesis and based on moderate M. montanus densities and moderate mutation rate is shown in [Figure 5](#pbio-0020290-g005){ref-type="fig"} ([Table 2](#pbio-0020290-t002){ref-type="table"} shows all combinations), along with the observed change. However, given the observed change in gene diversity specifically between 2,525 and 845 ybp, we were able to reject the null hypothesis for all nine combinations of mutation rate and effective size. These results suggest that M. montanus was not acting as a closed population during this period of time. Because the serial coalescent model presented here does not discriminate between selection and migration, either of these processes could have caused the observed change in gene diversity. ::: {#pbio-0020290-g005 .fig} Figure 5 ::: {.caption} ###### Distribution of Change in Gene Diversity for M. montanus between 2,525 and 166 ybp, Based on Serial Coalescent Simulations Sampling is modeled at two points in time. *N~e-mm\_ecol~* estimates from [Figure 2](#pbio-0020290-g002){ref-type="fig"} are used to specify demographic history. Eight of the nine combinations of mutation rate and density allow us to reject the null hypothesis for a closed population ([Table 2](#pbio-0020290-t002){ref-type="table"}). This figure illustrates simulation results for moderate density and moderate mutation rate (4% per million years per bp). The probability of the observed change (shown by dashed arrow) is significant (*p* = 0.015). :::  ::: ::: {#pbio-0020290-t002 .table-wrap} Table 2 ::: {.caption} ###### The Average (over 1,000 Simulations) Expected Change in Gene Diversity for Nine Combinations of Density and Mutation Rate for M. montanus between 2,525 and 166 ybp :::  Significance values for observed change are in parentheses. All combinations other than moderate mutation rate and low density allow rejection of the null hypothesis of a closed population ::: Our results may depend on our assumptions of equilibrium population size prior to 2,525 ybp. We investigated sensitivity to this assumption by modeling a population bottleneck in M. montanus prior to 2,525 ybp. We modeled population reduction to 104,577 (0.75 × *N~e-mm\_ecol2525ybp~*), 69,718 (0.5 × *N~e-mm\_ecol2525ybp~*), and 34,859 (0.25 × *N~e-mm\_ecol2525ybp~*) prior to 2,525 ybp. For a 75% bottleneck prior to 2,525 ybp, we were no longer able to reject the null hypothesis of a closed population. These results indicate that an extreme bottleneck where population size was reduced to 75% or more might result in the observed change in gene diversity. Our simulations reveal that unless an extreme bottleneck happened prior to 2,525 ybp, we can be confident that the observed change in gene diversity is not likely to be from events that occurred immediately prior to our historic data, and thus is due to migration or selection. Coalescent simulations used to investigate the significance of the observed gene diversity value for the modern samples demonstrated that the null hypothesis of past population size change could be rejected at only two of the nine mutation rate and density combinations. These results reveal that given data from only the modern samples, it was not possible to reject the null hypothesis of past population size change in *M. montanus.* Historic genetic data allow us to discriminate between population processes over millennia much better than do modern data alone. Other Evidence for Migration {#s2h} ---------------------------- Independent lines of genetic and demographic evidence also point to the influence of gene flow in M. montanus populations. Estimates of *β*-diversity (used here to measure haplotypic turnover) from haplotypic data reveal that turnover was highest between 2,525 and 845 ybp (*β~2525--845ybp~* = 3; *β~845--166ybp~* = 1.5; and *β~2525--166ybp~* = 1.5). Closer examination of the haplotypic distributions demonstrates that five novel haplotypes appeared by 845 ybp, three of which are ≥3.2% different from the most common haplogroup *(A)* ([Figure 3](#pbio-0020290-g003){ref-type="fig"}A) at 2,525 ybp, further implicating gene flow between 2,525 and 845 ybp. Discussion {#s3} ========== Our results demonstrate different genetic responses by two species of small mammals to changes in population size driven by climatic change. Fossil abundance data reveal population decline for both T. talpoides and M. montanus between 1,438 and 470 ybp, a period spanning the Medieval Warm Period ([@pbio-0020290-Hadly1]). For *T. talpoides,* the genetic response is directly related to changes in population size: Decrease in population size results in lowered gene diversity. M. montanus demonstrates the opposite relationship: A decrease in population size (between 1,438 and 166 ybp) results in an increase in gene diversity. We attempted to statistically validate our results by the use of serial coalescent simulations to demonstrate that the change in gene diversity of M. montanus between 2,525 and 845 ybp is significantly different from that expected based on the decrease in ecological estimates of population size. Taken together, these results indicate a departure from conditions of equilibrium (closed population without selection) for *M. montanus.* Our results have the following possible explanations: (1) the sampling area for fossils changed, (2) the local population size expanded, (3) selection occurred, and/or (4) gene flow occurred. Selection versus Gene Flow {#s3a} -------------------------- Results from all three of our analyses suggest that gene flow could be responsible for the patterns in gene diversity observed in our empirical data. Additionally, recent results of experimental studies of density dispersal dynamics in the root vole, *Microtus oeconomus,* indicate that migration occurs most frequently in and between low-density patches ([@pbio-0020290-Andreassen1]). These results indicate that density and dispersal in voles may be inversely related, a finding that is consistent with our results. An alternative explanation is that selection is governing the observed gene diversity patterns. While cytochrome *b* may not be under intense selection ([@pbio-0020290-Irwin1]), it is linked to other portions of the mitochondrial genome that may be selectively advantageous in particular environments. Using cytochrome *b* as a marker for the accumulations of adaptations elsewhere on the genome may yield information about the effects of selection on local populations through time. Further exploration is necessary to investigate and identify the presence of locally adapted mtDNA and the rates of evolutionary change necessary to produce the variation in gene diversity we have observed (e.g., [@pbio-0020290-Pergams1]). Conclusions {#s3b} ----------- Here we demonstrate, using a phylochronologic approach, that it is possible to distinguish the dynamic processes that govern gene diversity over relatively short time scales (hundreds to thousands of years). We have documented environmental change, population response, genetic diversity change, and the correlations between the three. Without serial data, we would capture just a single record of these historic processes: modern genetic diversity. Although it is possible to hypothesize about historic events using modern data, phylochronology affords a unique look into the past and the potential ability to separate cause from effect. In particular, we show that M. montanus has a history recording responses both within populations (fluctuations in population size, possible selection) and between populations (gene flow). Discrepancy between the ecological and genetic estimates of population size and significant changes in haplotypic diversity prior to the Medieval Warm Period implicate increased gene flow into the Lamar Cave M. montanus population. Additionally, the observed haplotypic turnover in the Yellowstone population during this period suggests that as abundance of M. montanus declined through the last 845 years, relatively more individuals carried newly introduced haplotypes. Our results indicate that the presently observed widespread genetic variation across the geographic range in this species arose not because gene flow was equivalent through all populations through time, but because during particular time periods, certain local populations (and/or genotypes) declined while others expanded. Our data show that even with a prolonged ecological population size decline, the genetic diversity of M. montanus was maintained. In contrast, gene flow has not played a significant role in the recent genetic history of *T. talpoides.* This species, instead, responded more as a closed population over this time. The disparate nature of population response to climatic change of these two species is likely due to differences in demographic dispersal patterns between their populations. Such differences in species demography have resulted in differential genetic response to climatic change, even when ecological response is similar. Thus, genetic response to environmental change can be viewed as "individualistic," similar to unique adjustments of species ranges ([@pbio-0020290-Root1]). Life history traits such as dispersal ability contribute to the overall gene diversity of species in both space and time. If life history has such a large impact for common species, such differences will be particularly important in understanding how entire communities are affected by global change. Ultimately, knowledge from such analyses will lead to distinct, and perhaps predictable, patterns of species persistence through climatic changes, insights that will prove invaluable to future conservation of biodiversity. Materials and Methods {#s4} ===================== {#s4a} ### Fossil locality {#s4a1} Lamar Cave contains well-stratified, thoroughly radiocarbon-dated deposits, which display high fidelity to the local mammalian community ([@pbio-0020290-Hadly1], [@pbio-0020290-Hadly3]). The most common animals from Lamar Cave are also the most common in the sagebrush grassland ecosystem in which Lamar Cave is located. Relative abundances are based on the entire data set of 10,597 specimens (except for those in [Figure 1](#pbio-0020290-g001){ref-type="fig"}A, which uses the five most common small mammals \[*n* = 8,589\]) and are concordant with expectations of taxonomic diversity in montane mammal communities of western North America ([@pbio-0020290-Hadly4]). The cumulative number of bones in Lamar Cave is correlated with time, demonstrating a constant "rain" of bones from the past to the present environment (*R* ^2^ = 0.86; unpublished data). ### Age assignment {#s4a2} The historical *Microtus* samples encompass 15 of 16 radiocarbon-dated stratigraphic levels from Lamar Cave ([@pbio-0020290-Hadly1]), with a maximum radiocarbon age of 2,860 ± 70 ybp (CAMS-20356). Data from the stratigraphic levels were pooled into five discrete intervals for Lamar Cave representing the last 3,000 years ([@pbio-0020290-Hadly1]). Interval boundaries were based on the stratigraphic pattern of deposition observed during excavation as well as a detailed radiocarbon chronology. Thus each interval represents a biologically significant packet of specimens. Age for each interval was assigned as the midpoint of the span of the calibrated radiocarbon ages ([@pbio-0020290-Hadly1]). ### Ecological estimates of population size {#s4a3} Absolute population sizes in both ancient and modern communities are difficult to estimate. However, relative abundance changes of the small mammals are consistent with the climatically caused changes in habitats and the habitat preferences documented by modern trapping data proximate to the fossil site ([@pbio-0020290-Hadly3], [Figure 1](#pbio-0020290-g001){ref-type="fig"}A). By calculating the area preferred by M. montanus and T. talpoides using the geographical information system, we were able to standardize the relationship between taphonomy and population size in these species. This was possible because the collection radius of the fossils from Lamar Cave has been documented to be less than 7 km by using strontium isotopes ([@pbio-0020290-Porder1]). We then estimated ecological effective population size from relative abundance through time, current population density, and the current area of preferred habitat in the collection radius. The current area of M. montanus habitat within the 7-km radius totals 1,992 ha; for T. talpoides it totals 971 ha. High, moderate, and low densities for T. talpoides ([@pbio-0020290-Verts1]) and M. montanus ([@pbio-0020290-Sullivan1]) were used in association with the corresponding areas of occupied habitat to estimate current census size. Because the rate of accumulation of bones through time is constant, the current census size was indexed against the percentage of *Microtus* bones from the uppermost level of Lamar Cave and used to calculate historic census sizes through time. In order to compare ecological estimates of effective population size to genetic estimates, we estimated mitochondrial effective population size assuming that *N~e~/N~census~* = 0.5 ([@pbio-0020290-Storz2]; estimates from small mammals), and given that mitochondrial effective size is *N~e~*/4. Our estimates of *N~e~* were based entirely on relative fossil abundance and current ecological data and assume a mitochondrial effective size; thus we labelled them *N~e-tt\_ecol~ (T. talpoides)* and *N~e-mm\_ecol~ (M. montanus).* ### Genetic data from *M. montanus* {#s4a4} From the fossil material, we used the upper first molar (from only one side of the jaw per level) to avoid sampling the same individual multiple times. Some *Microtus* species are cryptic with respect to these teeth. Genetic diagnosis indicated that these 78 fossil samples are derived from multiple arvicoline species, with a total of 47 M. montanus specimens (see [Table 1](#pbio-0020290-t001){ref-type="table"}). Fossil samples (*Microtus* molariform teeth) ranged from 3.5 to 15.3 mg (average, 8.8 mg). We used two previously described extraction methods on the fossil teeth ([@pbio-0020290-Hadly6]). We obtained a 312-bp fragment of mitochondrial cytochrome *b* from ancient samples (*n* = 47) using the following primers: forward primers (5′--3′), CLETH 37 TAY AAY ATA ATY GAA ACH TGA A (5′ end of cyt *b* 319 equals Mus musculus 14458), CLETH 37L AYG GMT CTT AYA ACA TAA TCG AAA CAT G (cyt *b* 311, M. musculus 14450), MMONT 1 CAG TAA TTA CAA AYC TWC TAT CA (cyt *b* 452, M. musculus 14591), and MMONT 3 AGT GAA TCT GAG GGG GCT TCT CAG TAG A (cyt *b* 485, M. musculus 14621); reverse primers (5′--3′), ARVIC 08 CAG ATY CAY TCY ACT AGT GTT G (cyt *b* 473, M. musculus 14612), ARVIC 08L CTC AGA TTC ACT CTA CTA GTG TTG TG (cyt *b* 471, M. musculus 14610), MMONT 4 TTR TTT GAT CCT GTT TCG TGT AGG AAT A (cyt *b* 595, M. musculus 14631), and MMONT 2L TTG ACT GTG TAG TAA GGG TGA AAT GGG A (cyt *b* 653, M. musculus 14792). Attempts were made to amplify the region in two overlapping fragments using CLETH37/ARVIC 08, CLETH37L/ARVIC 08L, and MMONT 1/MMONT 2L. However, the low rate of success for MMONT 1/MMONT 2L (40%) in the fossil samples necessitated the breaking of the second fragment into two overlapping fragments using MMONT 1/MMONT 4 and MMONT 3/MMONT 2L. Ancient DNA samples were run on an ABI PRISM 310 Genetic Analyzer in the post-PCR lab, and modern DNA samples were run on an ABI PRISM 377 Sequencer in a separate sequencing facility (Protein and Nucleotide Facility, Beckman Center, Palo Alto, California, United States). Fragments were sequenced in both directions, primer regions were overlapped, and sequences with any ambiguous sites were rerun until completely resolved in order to provide additional corroboration and eliminate ambiguity. We adhered to strict extraction and amplification protocols ([@pbio-0020290-Hadly6]). The protocol further included (1) independent sequence corroboration of two samples (J. Mountain lab, Anthropological Sciences, Stanford University), (2) processing of modern samples using personnel and reagents in another lab (D. Petrov lab, Biological Sciences, Stanford University) all physically separate from the aDNA facility, (3) monitoring contamination with several extraction and PCR controls, (4) primer design specific to arvicoline species, and (5) no prior or concurrent history of working with murid species in any of the DNA facilities involved. Modern (spatial) genetic sampling was obtained from a variety of sources including modern skins, modern liver tissue, museum skins, and teeth derived from modern raptor pellets. DNA was successfully extracted from 16 modern skins *(n* = 13 *M. montanus; n* = 3 other *Microtus)* and 12 teeth *(n* = 5 *M. montanus; n* = 7 other *Microtus)* from the vicinity (within 10 km) of the fossil site. In addition, DNA was successfully extracted from 51 modern specimens (liver tissue and museum skins) collected within a 400-km radius of Lamar Cave. Of these 51 samples, 22 were derived from museum skins *(n* = 21 *M. montanus; n =* 1 *M. longicaudus)* and 29 from liver tissue *(n* = 9 *M. montanus; n* = 18 other *Microtus)* of specimens trapped in the field. We followed the animal tissue protocol using the Qiagen Dneasy Tissue Kit (Qiagen, Valencia, California, United States) on 6.5 mg of liver samples and 7.5 mg of tooth samples (*n* = 3). For museum samples, DNA was extracted from the ventral skin incision (0.5 to 10.3 mg; average, 2.7 mg). We amplified the entire cytochrome *b* gene (1,143 bp) for some modern skins (*n* = 13) and liver tissue (*n* = 29) with the following primers: forward primers (5′--3′), MVZ 05 CGA AGC TTG ATA TGA AAA ACC ATC GTT (cyt *b* −51, M. musculus 14088) and ARVIC 07 AAA GCC ACC CTC ACA CGA TT (cyt *b* 514, M. musculus 14653); reverse primers (5′--3′) MICRO 06 GGA TTA TTT GAT CCT GTT TCG T (cyt *b* 602, M. musculus 14741) and VOLE 14 TTT CAT TAC TGG TTT ACA AGA C (cyt *b* 1170, M. musculus 15309). DNA was extracted from a total of 81 modern *Microtus* samples, and 79 of those yielded successful amplification. Of these 79 samples, 48 were positively identified as M. montanus. Of the M. montanus samples, 41 specimens were successfully haplotyped. Sequences have been deposited in GenBank ([see Supporting Information](#s5){ref-type="sec"}). ### Genetic data from *T. talpoides* {#s4a5} We have built upon the previously published T. talpoides data set ([@pbio-0020290-Hadly5]) with additional temporal sampling (*n* = 3) (see [Table 1](#pbio-0020290-t001){ref-type="table"}). T. talpoides from Lamar Cave demonstrated remarkable continuity in gene diversity through time, with only three haplotypes present, all of which differ from each other by one synonymous third-position mutation. The majority of the fossil T. talpoides specimens are from haplotype *A* (83%), which is not found elsewhere in a 400-km radius around Lamar Cave. This constancy in the genetic lineage of T. talpoides within a single locality persists in spite of climatic changes and concurrent significant population and body size changes ([@pbio-0020290-Hadly5]). ### Data analysis {#s4a6} Arlequin v. 2.000 ([@pbio-0020290-Jaarola2]; [@pbio-0020290-Schneider1]) was used to calculate gene diversity, number of segregating sites, and nucleotide diversity for each time interval (see [Figure 1](#pbio-0020290-g001){ref-type="fig"}B and [Table 1](#pbio-0020290-t001){ref-type="table"}). Additionally, genetic data were used to estimate the mitochondrial effective population size *(N~e~)* for all points in the past. Assuming a neutral model of molecular evolution, *θ~S~* (where *S* is the number of segregating sites; *θ~S~* = 2 *N~e~μ*, where *μ* is the mutation rate for the complete sequence per generation) was used to estimate *N~e~. θ~S~* was preferred over other estimators of *θ* since *θ~H~* is biased for single locus estimates, *θ~k~* does not incorporate sequence information, and *θ~π~* has higher variance. These estimates are determined entirely from genetic data; thus we label them *N~e-tt\_gen~ (T. talpoides)* and *N~e-mm\_gen~ (M. montanus).* We assumed a range of *μ* (2% \[low\], 4% \[moderate\], or 10% \[high\] per million years per bp) ([Table 2](#pbio-0020290-t002){ref-type="table"}). Additionally, to estimate haplotype turnover we calculated *β*-diversity (used traditionally in ecology to measure species turnover) using Cody\'s index ([@pbio-0020290-Cody1]). For *β*-diversity each haplotype was treated as a "species." ### Effects of sampling {#s4a7} Given the general limitations of obtaining aDNA sequence data, sample sizes will always present challenges to ancient population genetic studies. To investigate whether our samples are adequate for addressing temporal gene diversity in both species, we used a neutral population model to evaluate expected diversity. [@pbio-0020290-Ewens1] derived expressions for the sampling distribution characterizing a closed population of size *N,* a gene with mutation rate *μ,* and an infinite alleles model given a sample size *n.* Here we use the Ewens distribution to ascertain the ability of our data to detect variation in values of gene diversity. In addition, this distribution allows us to (1) predict the distribution of expected gene diversity at each point in time (independently) given estimates of *N~e\_ecol~,* a moderate mutation rate of 4% per million years, typed sequence length (312 bp for M. montanus and 63 bp for T. talpoides), and sample size (*M. montanus: n~166ybp~* = 7, *n~470ybp~* = 7, *n~845ybp~* = 18, *n~1438ybp~* = 4, and *n~2525ybp~* = 6; *T. talpoides: n~166ybp~* = 34, *n~470ybp~* = 5, *n~845ybp~* = 29, *n~1438ybp~* = 4, and *n~2525ybp~* = 11) and (2) predict change in the expected gene diversity distribution for a large sample size of 100. A modified version of MONTE CARLO ([@pbio-0020290-Slatkin1], [@pbio-0020290-Slatkin2]) was used to generate possible allele configurations for a given set of parameters *(θ, K,* and *n),* and gene diversity was calculated for each configuration. Simulations were repeated 1,000 times for each sampling time point, resulting in the 95% confidence intervals for the distribution of predicted gene diversities given a closed, selectively neutral population through time. ### Serial coalescent simulations {#s4a8} We assessed the significance of the observed changes in gene diversity between time points using simulations. The serial coalescent was used as a framework to model M. montanus evolution during the past 2,525 years. Simulations were repeated 1,000 times to generate genetic data and changes in gene diversity for the null hypothesis. The significance of the observed changes was then inferred by comparing it to the generated null values (we thus used a Monte Carlo significance test). Estimates of *N~e\_ecol~* were used to set up a null hypothesis specifying population size change through time in a closed population. We assumed that past changes in population size between intervals were due to exponential growth or decline. The effective population size at two time points and the time between the points was used to calculate a growth rate. The estimated growth rates were *r~1438--2525ybp~* = −0.000178, *r~845--1438ybp~* = 0.000732, *r~470--845ybp~* = 0.0004385, and *r~166--470ybp~* = 0.0003212. The null hypothesis thus corresponds to effective population sizes at particular points in the past and exponential growth or decline between those intervals, and represents an ecologically realistic description of the past 2,525 years. We assumed a constant population size prior to 2,525 ybp, as we have no data before this point in time. The coalescent program SIMCOAL ([@pbio-0020290-Excoffier1]) was modified to incorporate temporal sampling (also known as heterochronous sampling): *n~1~* samples modeled back in time, with *n~2~* samples added to the genealogy at a time point *t~1~* generations in the past. The serial coalescent ([@pbio-0020290-Rodrigo1]; [@pbio-0020290-Drummond2]) has been used to estimate parameters such as *μ* for HIV and ancient mtDNA (most recently using an MCMC approach; [@pbio-0020290-Drummond1]; [@pbio-0020290-Lambert1]). In this paper, we present what we believe to be its first application as a simulation tool used to predict change in gene diversity for the null model of population size change described above. Running the model 1,000 times provides the expected distribution for change in gene diversity. Using *N~e-mm\_ecol~* estimates based on high-, moderate-, and low-density estimates for M. montanus (186, 126, and 60 voles/ha; [@pbio-0020290-Sullivan1]) and a high, moderate, and low mutation rate (*μ* = 10%, 4%, and 2% per million years per bp; sequence length = 312 bp; finite sites mutation model, no rate heterogeneity), we investigated the significance of observed changes in gene diversity at all nine parameter combinations over the entire time span (2,525 to 166 ybp) and between 2,525 and 845 ybp (spanning the Medieval Warm Period). Additionally, we also investigated the relevance of temporal data to our ability to reject the null hypothesis. Significance tests were repeated assuming the observed data were from only the most recent genetic samples. Again, simulations were repeated for all nine combinations of mutation rate and density estimates. Supporting Information {#s5} ====================== Accession Numbers {#s5a1} ----------------- Sequences for the successfully haplotyped M. montanus specimens described in [Materials and Methods](#s4){ref-type="sec"} have been deposited in GenBank under accession numbers AY660606 to AY660629. We thank J. Zinck for her assistance in the lab, C. Roseman for his initial work with effective size estimates, and J. Mountain and D. Petrov for use of their molecular facilities. We particularly appreciate the comments and discussion from T. Oliver. Several comments improved earlier versions of this manuscript and we thank D. Ackerly, A. Barnosky, P. Ehrlich, M. Feldman, J. Mountain, S. Palumbi, J. Patton, A. Paytan, D. Petrov, T. Root, S. Schneider, P. Vitousek, members of the Hadly lab, Jay Storz, and two anonymous reviewers\' suggestions. Financial support was provided by National Science Foundation award DEB \#0108541 to EAH. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** EAH conceived and designed the experiments. YLC, MV, KOK, PAS, and CJC performed the experiments. EAH, UR, YLC, and MV analyzed the data. EAH and UR contributed reagents/materials/analysis tools. EAH, UR, YLC, MV, and KOK wrote the paper. Academic Editor: Craig Moritz, University of California, Berkeley Citation: Hadly EA, Ramakrishnan U, Chan YL, van Tuinen M, O\'Keefe K, et al. (2004) Genetic response to climatic change: Insights from ancient DNA and phylochronology. PLoS Biol 2(10): e290. aDNA : ancient DNA bp : base pair ha : hectare mtDNA : mitochondrial DNA *N~e\_ecol~* : estimate of effective size derived from ecological data *N~e\_gen~* : estimate of effective size derived from genetic data *N~e-mm\_ecol~* : *N~e\_ecol~* for *M. montanus* *N~e-mm\_gen~* : *N~e\_gen~* for *M. montanus* *N~e-tt\_ecol~* : *N~e\_ecol~* for *T. talpoides* *N~e-tt\_gen~* : *N~e\_gen~* for *T. talpoides* ybp : years before present

PubMed Central

2024-06-05T03:55:47.376033

2004-9-7

PMC515370

Introduction {#s1} ============ Deinococcus radiodurans is a non-spore-forming bacterium notable for its capacity to tolerate exposure to ionizing radiation ([@pbio-0020304-Battista1]). The D~37~ dose for D. radiodurans R1 is approximately 6,500 Gy, at least 200-fold higher than the D~37~ dose of Escherichia coli cultures irradiated under the same conditions. The energy deposited by 6,500-Gy γ radiation should introduce thousands of DNA lesions, including hundreds of double-stranded breaks ([@pbio-0020304-Smith1]). The mechanisms responsible for this species\' resilience are poorly described, and recent analyses of DNA-damage-induced changes in the proteome ([@pbio-0020304-Lipton1]) and transcriptome ([@pbio-0020304-Liu1]) of D. radiodurans cultures have done little to improve our understanding of *D. radiodurans\'* radioresistance ([@pbio-0020304-Edwards1]; [@pbio-0020304-Narumi1]). For most species, the intracellular generation of strand breaks has lethal consequences; exposed free ends serve as substrates for intracellular exonucleases that degrade the genome. However, in D. radiodurans the presence of strand breaks does not result in a catastrophic loss of genetic information ([@pbio-0020304-Dean1]; [@pbio-0020304-Lett1]; [@pbio-0020304-Vukovic-Nagy1]). Instead, this species appears to have the ability to control DNA degradation postirradiation by synthesizing proteins that prevent extensive digestion of the genome, and it has been suggested that the DNA degradation observed in this species is an integral part of the process of DNA repair, generating single-stranded DNA that promotes homologous recombination and restitution of the damaged genome ([@pbio-0020304-Battista2]). When D. radiodurans is exposed to a high dose of ionizing radiation, a number of genes are induced that lack readily identifiable homologues among known prokaryotic proteins ([@pbio-0020304-Liu1]; [@pbio-0020304-Tanaka1]). Among these is the gene designated DR0423. This locus is one of the most highly induced genes in *Deinococcus* following γ-irradiation, with expression increasing 20- to 30-fold relative to an untreated control. Although originally annotated as a "hypothetical" protein ([@pbio-0020304-White1]), a more detailed analysis ([@pbio-0020304-Iyer1]) has identified an evolutionary relationship between DR0423p and the important eukaryotic recombination protein Rad52. Rad52 is part of a larger family of proteins exhibiting structural similarity but little sequence homology, including the prokaryotic Redβ, RecT, and Erf proteins ([@pbio-0020304-Passy1]; [@pbio-0020304-Iyer1]). In this report, we provide evidence for a DNA end-protection system in D. radiodurans and characterize the DR0423 protein as a component of that system. Our studies suggest that DNA end protection might be particularly important to this species in the context of long-term survival during desiccation and recovery in a nutrient-poor environment. Results {#s2} ======= Transcripts Corresponding to the Coding Sequence Designated DR0423 Increase in Response to Sublethal Doses of Ionizing Radiation {#s2a} -------------------------------------------------------------------------------------------------------------------------------- During the course of microarray studies intended to establish which R1 loci respond to ionizing radiation, it was noted that transcripts of DR0423 were among the mostly highly induced ([@pbio-0020304-Tanaka1]). As an independent confirmation of these microarray results, the expression of this gene was monitored using quantitative real-time PCR. Total RNA was isolated from exponential-phase cultures of R1 immediately after and at 30 and 60 min following exposure to 3,000-Gy ionizing radiation. Changes in transcript abundance for the *recA* (DR2340), *gap* (DR1343), and DR0423 genes were determined as previously described ([@pbio-0020304-Earl1]). The results of these analyses are listed in [Table 1](#pbio-0020304-t001){ref-type="table"}. Consistent with previous results, levels of *recA* transcript increased postirradiation ([@pbio-0020304-Narumi2]; [@pbio-0020304-Bonacossa1]; [@pbio-0020304-Satoh1]), whereas *gap* induction remained unchanged ([@pbio-0020304-Earl1]). The *gap* gene encodes glyceraldehyde 3-phosphate dehydrogenase and does not respond to DNA damage. Within one-half hour postirradiation, levels of DR0423 transcript increased 20- to 30-fold, suggesting that DR0423p may be a previously unrecognized component of the cell\'s defense against ionizing-radiation-induced damage. ::: {#pbio-0020304-t001 .table-wrap} Table 1 ::: {.caption} ###### Relative Expression of the *ddrA*, *recA,* and *gap* Genes of D. radiodurans R1 following Exposure to 3,000-Gy Ionizing Radiation :::  Relative expression was determined before and after irradiation by calculating transcript abundance using quantitative RT-PCR. The numbers in this table are the ratio of transcript present postirradiation to that present preirradiation. Values are the means of ratios calculated from three independent experiments (*n* = 6). The ranges of values obtained are included in parentheses adjacent to each mean. A value greater than one indicates an increase in expression in response to ionizing radiation ::: Deletion of DR0423 Sensitizes D. radiodurans R1 to Ionizing Radiation and Mitomycin C {#s2b} ------------------------------------------------------------------------------------- The DR0423 gene was inactivated by deletion in D. radiodurans R1, as described elsewhere ([@pbio-0020304-Funayama1]), and the resulting strain designated TNK104. Confirmation of the gene deletion is provided in [Figure 1](#pbio-0020304-g001){ref-type="fig"}. Deletion of DR0423 does not alter the growth rate of the culture (approximately 1.5-h doubling time), or decrease the efficiency of natural transformation (approximately 5 × 10^−5^ rifampicin-resistant transformants per colony-forming unit ) relative to R1, indicating that DR0423p is not essential for the processes of DNA replication or homologous recombination. To establish whether DR0423p was necessary for DNA damage tolerance, TNK104 was evaluated for its ability to survive ionizing radiation and mitomycin C. Aliquots of exponential-phase cultures were exposed to these DNA-damaging agents. TNK104 exhibits increased sensitivity to both agents relative to the wild-type R1 strain ([Figure 2](#pbio-0020304-g002){ref-type="fig"}), but cultures only displayed significant ionizing radiation sensitivity at doses in excess of 5,000 Gy. Since expression of DR0423 increases in response to ionizing radiation, and its gene product contributes to the DNA damage resistance of this species, we have chosen to designate this gene as *DNA damage response A (ddrA)*. ::: {#pbio-0020304-g001 .fig} Figure 1 ::: {.caption} ###### Verification of Gene Deletions \(A) Verification of *ddrA* and *recA* gene deletions by PCR analysis. Purified PCR fragments were amplified from the genomic DNA of strains R1, TNK104, TNK106, and TNK110 using primers that flank the coding sequences for *ddrA* and *recA.* Products were separated on a 0.8% agarose gel to establish whether the fragment size corresponded to the gene-replacement cassette. The left panel depicts the replacement of *ddrA* in TNK104 and TNK110. The right panel depicts the replacement of *recA* in TNK106 and TNK110. Expected sizes of the wild-type and mutant sequences are given in the figure above each image of the agarose gel. \(B) Verification of the *ddrA* gene deletion by restriction analysis of purified PCR products. Purified PCR fragments were amplified from the genomic DNA of strains R1, TNK104, and TNK110, using primers that flank the coding sequences for *ddrA.* Products were restricted with EcoR1 (left panel) and EcoRV (right panel) to verify their identity. Products were separated on a 0.8% agarose gel to establish whether the restriction fragment corresponded with the expected sizes as illustrated in the figure above each image of the agarose gel. \(C) Verification of the *recA* gene deletion by restriction analysis of purified PCR products. Purified PCR fragments were amplified from the genomic DNA of strains R1, TNK106, and TNK110, using primers that flank the coding sequences for *recA.* Products were restricted with PvuII (left panel) and BglII (right panel) to verify their identity. Products were separated on a 0.8% agarose gel to establish whether the restriction fragment corresponded with expected sizes as illustrated in the figure above each gel. :::  ::: ::: {#pbio-0020304-g002 .fig} Figure 2 ::: {.caption} ###### DNA Damage Sensitivity of D. radiodurans Cells Lacking DdrA Function \(A) Representative survival curves for D. radiodurans strain TNK104 *ΔddrA* (squares) and D. radiodurans R1 (circles) following exposure to γ radiation. Survival of strains; values are the mean ± standard deviation of three independent experiments; *n* = 9. \(B) Representative survival curves for D. radiodurans strain TNK104 *ΔddrA* (squares) and D. radiodurans R1 (circles) following exposure to mitomycin C. Values are the mean ± standard deviation of three independent experiments; *n* = 9. :::  ::: A *ddrA recA* Double Mutant Is More Sensitive to Ionizing Radiation Than Either Single Mutant {#s2c} --------------------------------------------------------------------------------------------- The *recA* gene was deleted from R1 and TNK104 (see [Figure 1](#pbio-0020304-g001){ref-type="fig"}B and [1](#pbio-0020304-g001){ref-type="fig"}C), resulting in strains TNK106 *(ΔrecA)* and TNK110 *(ΔrecA, ΔddrA),* respectively. Deinococcal strains lacking *recA* function are considered the most ionizing-radiation-sensitive strains described for this species ([@pbio-0020304-Moseley1]; [@pbio-0020304-Gutman1]). However, as indicated in [Figure 3](#pbio-0020304-g003){ref-type="fig"}, TNK110 is 3- to 5-fold more sensitive to ionizing radiation than the *ΔrecA* strain, indicating that DNA damage response A protein (DdrA), at least in part, contributes to *D. radiodurans\'* survival by a mechanism that is independent of RecA function. ::: {#pbio-0020304-g003 .fig} Figure 3 ::: {.caption} ###### DdrA Functions in a RecA-Independent DNA Repair Process Representative survival curves for D. radiodurans strains TNK106 *ΔrecA* (closed circles) and TNK110 *ΔddrA ΔrecA* (open triangles) following exposure to lower levels of γ radiation. All values are the mean ± standard deviation of three independent experiments; *n* = 9. :::  ::: Evidence That the DdrA Protein Contributes to Genome Restitution {#s2d} ---------------------------------------------------------------- To determine if loss of DdrA affected genome restitution and stability postirradiation, we followed the recovery of cultures of R1 and TNK104 following a 5,000-Gy dose of γ radiation. Initially, exponential-phase cultures were harvested, suspended in 10 mM MgSO~4~, and irradiated. No carbon source was added. Restoration of the genome was monitored by pulsed-field gel electrophoresis, and aliquots retrieved from the recovering cultures were used to determine viability. Cultures were left in this medium and sampled at 24-h intervals over a 120-h time course. The gel depicted in [Figure 4](#pbio-0020304-g004){ref-type="fig"} illustrates the reassembly of the genomes of irradiated R1 cells. There are 11 NotI sites in the D. radiodurans genome, and when restricted, most of the resulting fragments can be separated by pulsed-field gel electrophoresis as seen in the lane (C) corresponding to the unirradiated control. Immediately after irradiation, the introduction of DNA double-stranded breaks results in the disappearance of the higher molecular weight NotI fragments, but the pattern of fragments is restored in 24--48 h, indicating that R1 is repairing double-stranded breaks under these conditions, in spite of the absence of nutrients. This pattern persists throughout the rest of the time course, indicating that once reformed the genome is stable. Despite genome restitution, R1 cultures held in MgSO~4~ are not as proficient at recovering from ionizing-radiation-induced damage as cultures that are allowed to recover in rich medium ([Figure 4](#pbio-0020304-g004){ref-type="fig"}; data not shown). Even if plated immediately after exposure, R1 cultures suspended in MgSO~4~ exhibit a modest 2-fold reduction in viability when exposed to 5,000-Gy γ radiation relative to R1 cultures irradiated in rich medium (see [Figure 2](#pbio-0020304-g002){ref-type="fig"}). The longer the culture is held in MgSO~4~ ([Figure 5](#pbio-0020304-g005){ref-type="fig"}), the greater the reduction in viability. After 120 h, approximately 10% of the irradiated R1 population remains viable. In comparison, 80% of an unirradiated exponential-phase population of R1 is viable when kept in 10 mM MgSO~4~ for 5 d (data not shown). ::: {#pbio-0020304-g004 .fig} Figure 4 ::: {.caption} ###### Genome Recovery in the Absence of Nutrients Depends on DdrA \(A) Pulsed-field gel electrophoresis analyses of D. radiodurans strain RI recovery over a 120 h time course in 10 mM MgSO~4~ following 5,000-Gy γ radiation. \(B) Pulsed-field gel electrophoresis analyses of D. radiodurans strain TNK104 (*ΔddrA*) recovery following 5,000-Gy γ radiation. :::  ::: ::: {#pbio-0020304-g005 .fig} Figure 5 ::: {.caption} ###### DdrA Protein Effects on In Vivo Survival and Genome Preservation following Exposure to Ionizing Radiation in the Absence of Nutrients \(A) Survival of D. radiodurans R1 and TNK104 cultures held in 10 mM MgSO~4~ for 120 h following exposure to 5,000-Gy γ radiation. Samples were obtained at 24-h intervals. All values are the mean ± standard deviation of three independent experiments; *n* = 9 \(B) Changes in DNA content in cultures of R1 and TNK104 recovering from exposure to 5,000-Gy γ radiation in MgSO~4~. The DNA concentration at each time point is expressed as a percentage of that present in each strain prior to irradiation. :::  ::: Irradiated TNK104 cultures are significantly more vulnerable to ionizing radiation during a prolonged incubation in MgSO~4~ ([Figure 5](#pbio-0020304-g005){ref-type="fig"}A). TNK104 cultures exhibit only 0.1% survival after 120 h, a 100-fold reduction relative to identically treated R1 cultures. Also, in sharp contrast to the R1 cultures (see [Figure 4](#pbio-0020304-g004){ref-type="fig"}A), there is no evidence of genome reassembly in the TNK104 cells over this time course (see [Figure 4](#pbio-0020304-g004){ref-type="fig"}B), suggesting that failure to reassemble the genome contributes to the lower viability observed in TNK104 cultures. We directly examined the influence of DdrA on the fate of genomic DNA (see [Figure 5](#pbio-0020304-g005){ref-type="fig"}B) by monitoring changes in DNA content as the cultures of R1 and TNK104 recovered from exposure to 5,000 Gy in MgSO~4~. An aliquot of each unirradiated culture was isolated and total DNA concentration for 10^6^ cfu calculated. Following irradiation, the DNA content of a volume corresponding to the original 10^6^ cfu of each culture was determined. The DNA concentration at each time point in [Figure 5](#pbio-0020304-g005){ref-type="fig"}B is expressed as a percentage of that present in each strain prior to irradiation. Immediately after irradiation, the genomic DNA in the R1 culture was reduced by approximately 18%, a value consistent with previous findings ([@pbio-0020304-Dean1]; [@pbio-0020304-Lett1]; [@pbio-0020304-Vukovic-Nagy1]) that indicate that 20%--25% of the genomic DNA of D. radiodurans will be degraded and expelled from the cell following exposure to 5,000-Gy γ radiation. In contrast, genomic DNA degradation in the strain lacking DdrA approached 55%. Thus, the presence of DdrA has a greater than 3-fold effect on the preservation of genomic DNA during early times after irradiation. In the succeeding 120 h, the R1 genomic DNA was reduced by a total of 31%, while the loss of genomic DNA increased to 64% in TNK104. These results suggest that DdrA has a direct effect on the preservation of genomic DNA following extreme insults. We also examined genome restitution in a rich medium (TGY broth). Consistent with the survival curve depicted in [Figure 2](#pbio-0020304-g002){ref-type="fig"}, we found that when TNK104 cells are exposed to 5,000 Gy their genomes reassemble with kinetics identical to those of the wild-type R1 culture ([@pbio-0020304-Grimsley1]; [@pbio-0020304-Mattimore1]); the genome reforms in less than 6 h (data not shown). Thus, DdrA appears to contribute to genome reconstruction in D. radiodurans following irradiation, but this role was only obvious in cultures suspended in MgSO~4~. There could be at least two explanations for this observation. First, the action of DdrA may overlap with the activity of at least one other protein, and while each redundant activity is functional in rich medium, only DdrA is functional in cultures held in MgSO~4~. Alternatively, the primary role of DdrA could be the passive protection of exposed 3′ DNA ends at the sites of DNA strand breaks. Under conditions with limiting nutrient availability, DdrA could contribute to genome restitution simply by preventing the massive genomic degradation evident in [Figure 5](#pbio-0020304-g005){ref-type="fig"}B. In a rich medium, active DNA repair may render DdrA-mediated DNA protection less important. The Purified DdrA Protein Binds the 3′ Ends of Single-Stranded DNA and Protects Them from Digestion by an Exonuclease {#s2e} --------------------------------------------------------------------------------------------------------------------- The *ddrA* gene was cloned and expressed in *E. coli,* and the protein was purified to homogeneity ([Figure 6](#pbio-0020304-g006){ref-type="fig"}). The identity of the purified protein was confirmed by N-terminal sequencing and mass spectrometry. The deduced N-terminal sequence was MKLSDV, matching the predicted sequence of the first six amino acids perfectly (with the initiating methionine retained). The measured mass of the protein was 23,012.8 ± 3.46 Da, in good agreement with the 23,003.38 Da predicted. In two gel-filtration experiments using a Sephacryl S300 column calibrated with molecular weight standards, DdrA eluted as a sharp peak with an apparent mass in the two different trials of 218 and 190 kDa (data not shown). These results suggest that DdrA is an oligomer in solution with 8--10 subunits. Whereas these results are preliminary, Rad52 protein and other members of this family function as large oligomeric rings ([@pbio-0020304-Passy1]; [@pbio-0020304-Iyer1]; [@pbio-0020304-Singleton1]) ::: {#pbio-0020304-g006 .fig} Figure 6 ::: {.caption} ###### Purification of the DdrA Protein The first lane contains molecular weight markers. The second and third lanes contain crude extracts from E. coli strain pEAW298 (DdrA overproducer) in which the *ddrA* gene is uninduced or induced, respectively. The final lane contains purified DdrA protein. :::  ::: DdrA exhibited no ATPase, helicase, recombinase, or nuclease activity (data not shown). However, it bound to single-stranded DNA as determined by an electrophoretic mobility-shift assay (EMSA) ([Figure 7](#pbio-0020304-g007){ref-type="fig"}). Binding to duplex DNA depended on the presence of a 3′ single-stranded extension at one end ([Figure 7](#pbio-0020304-g007){ref-type="fig"}), indicating that the protein has some affinity for a free 3′ end in single-stranded DNA. This binding was not disrupted by a challenge with a 1,000- to 2,000-fold excess of a duplex oligonucleotide with a 5′ single-stranded extension ([Figure 7](#pbio-0020304-g007){ref-type="fig"}). ::: {#pbio-0020304-g007 .fig} Figure 7 ::: {.caption} ###### DdrA Protein Binds to Single-Stranded DNA with Free 3′ Ends Four sets of EMSAs are presented, with the gels and electrophoresis conditions carefully matched. DNA substrate concentrations are 0.7 nM in each case, reported as total molecules. In each set, the first three lanes show the effects of the indicated concentration of DdrA protein. The fourth and fifth lanes are identical to the second and third lanes, respectively, except that they are treated with proteinase K to demonstrate that the DNA has not been altered. In set D, the sixth and seventh lanes are identical to the third lane (with 4 μM DdrA protein), except that they have been challenged with a 1,000-fold or 2,000-fold excess of unlabeled oligo with a 5′ extension, respectively. The unlabeled challenge oligo is the same as that used in reaction set C. \(A) Single-stranded oligonucleotides (51 nt in length), labeled on the 5′ end. \(B) 5′ end--labeled duplex DNA fragments (51 bp). \(C) 5′ end--labeled oligonucleotide, with a self-complementary sequence leading to the formation of an 18-bp hairpin and a 15-nt 5′ single-stranded extension. \(D) 3′ end--labeled oligonucleotide, with a self-complementary sequence leading to the formation of an 18-bp hairpin and a 16-nt 3′ single-stranded extension. The sequences of the single-stranded extensions in the oligos used in sets C and D are matched, except that an extra adenosine residue has been added to the oligo used in set D during the labeling process. Note that in set B, only the lower substrate band (unannealed oligonucleotides) is bound by DdrA, and the migration of the resulting complexes is identical to that shown in set A. :::  ::: DdrA also protected the single-stranded DNA from degradation by exonuclease I from *E. coli,* which digests single-stranded DNA from the 3′ end ([Figure 8](#pbio-0020304-g008){ref-type="fig"}). The DNA binding trials shown in [Figure 8](#pbio-0020304-g008){ref-type="fig"}A were scaled up, and the bound species was cut out of a preparative gel. The extracted protein comigrated with DdrA protein on a sodium dodecyl sulfate (SDS)-polyacrylamide gel ([Figure 8](#pbio-0020304-g008){ref-type="fig"}B), providing further confidence that the binding is due to DdrA and not a minor contaminant in the DdrA protein preparation. These results suggest that D. radiodurans possesses a novel DNA end-protection system and that DdrA is a component of that system. ::: {#pbio-0020304-g008 .fig} Figure 8 ::: {.caption} ###### DdrA Protein Protects 3′ Ends from Degradation by Exonuclease I \(A) This set of reactions uses the labeled duplex DNA illustrated. The oligos annealed to form this DNA are 51 and 37 nt in length and pair so as to leave a 14-nt 3′ extension. The shorter DNA is 5′ end--labeled. The first lane contains unreacted DNA, showing both the annealed duplex and the unannealed single-stranded DNA. The second lane shows the DNA after treatment with 3 units of exonuclease I for 7 min in a 15-μl reaction mixture. Note that the duplex DNA in the upper band has been shortened by removal of the single-stranded extension. In lanes 3 and 4, the DdrA protein (4 μM) has been incubated with the DNA, without and with the 3 units of exonuclease I, respectively. The DNA is bound by DdrA and shifted to the top of the gel. The reactions shown in lanes 5 and 6 are identical to those in lanes 3 and 4, but with SDS and proteinase K added to disrupt the DdrA--DNA complexes and reveal that the DNA has been minimally affected by exonuclease I. The final lane shows another reaction of the DNA with 3 units of exonuclease I, in the presence of 4 μM bovine serum albumin. Exonuclease I degrades single-stranded DNA in the 3′ to 5′ direction. \(B) The protein bound to the duplex DNA is DdrA. The reaction of lane 3 in (A) was scaled up and the protein--DNA complex excised from the gel as described in [Materials and Methods](#s4){ref-type="sec"}. The protein in this complex was subjected to electrophoresis on an SDS-polyacrylamide gel, shown here (lane 3). The control lanes contained prestained protein standards (lane 1) and purified DdrA protein (lane 2). The gel-extracted protein comigrated with DdrA. :::  ::: The eukaryotic Rad52 protein has a single-stranded annealing activity that may be important to its in vivo function ([@pbio-0020304-Mortensen1]; [@pbio-0020304-Sugiyama1]). We carried out several tests to determine if the DdrA protein had a similar annealing activity. In multiple trials using oligonucleotides of 30 and 51 nucleotides (nt) in length, no DNA strand annealing activity was detected over a range of DdrA concentrations and conditions (data not shown). Discussion {#s3} ========== The extraordinary resistance of D. radiodurans to DNA damage arose not as an adaptation to high levels of radiation, but rather as a response to desiccation ([@pbio-0020304-Mattimore1]). In an arid environment, dormant D. radiodurans cells would gradually accumulate DNA lesions of all kinds, including strand breaks. Since DNA repair is highly reliant on metabolic energy, and appropriate nutrients cannot be assured upon rehydration, it is not unreasonable to expect that this species possesses a means to efficiently repair accumulated damage that minimizes energy use. In this context, mechanisms must have evolved to maintain the genome and protect it from unnecessary degradation by nucleases and other agents. In this study we have identified functions associated with a "hypothetical" protein encoded by D. radiodurans R1 that contributes to this species\' capacity to tolerate exposure to ionizing radiation and mitomycin C. We propose that the DR0423 protein, which we have designated DdrA, is part of a DNA end-protection system. Induced in response to the appearance of strand breaks generated by ionizing radiation (or subsequent to desiccation), DdrA would cap the strand breaks and help stabilize the genome until such time as conditions were more amenable to systematic DNA repair. The results we have obtained both in vivo and in vitro are consistent with this hypothesis. When the *ddrA* gene is deleted from R1, an otherwise wild-type cell becomes more sensitive to DNA-damaging agents (see [Figure 2](#pbio-0020304-g002){ref-type="fig"}). We show that DdrA has at least two activities: DdrA contributes to genome restitution following irradiation (see [Figure 4](#pbio-0020304-g004){ref-type="fig"}), and purified DdrA binds the 3′ ends of single-stranded DNA and protects those ends from digestion by exonucleases (see [Figures 7](#pbio-0020304-g007){ref-type="fig"} and [8](#pbio-0020304-g008){ref-type="fig"}). Notably, the effects of a *ddrA* deletion are amplified if nutrients are not provided after exposure to ionizing radiation, and cells held this way for 5 d display a 100-fold reduction in viability relative to the wild-type cells (see [Figure 5](#pbio-0020304-g005){ref-type="fig"}). In these nutrient-poor conditions, cells lacking DdrA protein do not restore their chromosomes. Instead, the chromosomes are degraded extensively. Even though the R1 strain was able to restore its genome following irradiation and incubation in 10 mM MgSO~4~, there was no evidence of genome reassembly in similarly treated cultures of TNK104, the *ΔddrA* derivative of R1 (see [Figure 4](#pbio-0020304-g004){ref-type="fig"}). This result indicates that DdrA plays a qualified role in genome restitution. Clearly the protein is necessary for this process in cells held in MgSO~4~, and we suggest that TNK104\'s inability to reconstitute its genome under these conditions is likely to be related to the DNA degradation that is observed in this strain following irradiation (see [Figure 5](#pbio-0020304-g005){ref-type="fig"}B). DdrA is not needed if cells are allowed to recover in a nutrient-rich medium (see [Figure 2](#pbio-0020304-g002){ref-type="fig"}). This suggests that the function that DdrA mediates in genome restitution is either redundant or unnecessary when other repair processes are robust. If there is a protein with a redundant activity, it is evident only in rich medium. We do not know the identity of the redundant component, or understand why it is not functional in MgSO~4~. Since DdrA binds the 3′ ends of single-stranded DNA, we presume that this protein either has the same activity or is rendered unnecessary by a compensating activity possible only in a nutrient-rich environment (such as DNA synthesis to counter exonucleolytic degradation). If, instead, DdrA is part of a passive DNA protection system, this system may be critical under conditions in which active (energy-requiring) DNA repair is not possible, such as when cells are desiccated or held in a nutrient-free medium. DdrA may not be as important in a nutrient-rich environment, where active DNA synthesis and other DNA repair processes may compensate for the loss of DNA end protection. The increased sensitivity observed in TNK110 *(ΔrecA ΔddrA)* relative to TNK106 *(ΔrecA)* indicates that DdrA participates in a process that complements RecA-mediated survival mechanisms (see [Figure 3](#pbio-0020304-g003){ref-type="fig"}), rescuing some irradiated cells even in the absence of RecA function. Since DdrA is distantly but specifically related to the Rad52 family of eukaryotic proteins, as well as a family of phage-associated proteins that mediate single-stranded annealing ([@pbio-0020304-Iyer1]), we speculate that DdrA could be a component of a single-stranded annealing system that functions simultaneously with RecA-dependent homologous recombination. This possibility is consistent with an earlier report by [@pbio-0020304-Daly1] who documented RecA-independent genome restitution postirradiation. They reported that approximately 30% of the R1 genome is assembled in a *recA* background during the first 1.5 h after exposure, and they suggested that this process was single-stranded annealing. The DdrA protein could act directly or indirectly in any single-stranded annealing process that might occur in *Deinococcus.* Although the related Rad52 protein possesses a single-stranded annealing activity ([@pbio-0020304-Mortensen1]; [@pbio-0020304-Sugiyama1]), we have thus far failed to detect such an activity with DdrA protein. One of three explanations seems likely: (i) we have not yet identified suitable conditions for the assay of DdrA-dependent DNA strand annealing; (ii) DdrA is part of a complex, and other proteins are needed to observe activity; or (iii) DdrA does not possess such an activity. DdrA\'s capacity to protect the 3′ ends of single-stranded DNA from digestion should help maintain the integrity of DNA fragments generated following DNA damage, whether those fragments are a result of the direct action of the damaging agent or arise as a consequence of a repair process that cleaves the phosphodiester backbone. By limiting degradation, proteins that protect DNA ends should enhance DNA damage tolerance and cell survival; the stabilized fragments serve as a long-lived substrate for homologous recombination or single-stranded annealing. In other words, we suspect that the ability to preserve genetic information is one key to understanding DdrA function and, in a larger context, the DNA damage tolerance of this species. DNA binding proteins, such as DdrA, may be particularly important for surviving desiccation. Like ionizing radiation, the process of desiccation is inherently DNA damaging, introducing large numbers of DNA double-stranded breaks. Following an extended period of desiccation, broken DNA ends would presumably need to be protected to minimize loss of genetic information. We know of no precedent for an activity of this sort in bacteria, although its existence has been predicted at least once ([@pbio-0020304-Clark1]). Bacteriophage are known to encode proteins (e.g., the gene 2 protein of T4 \[[@pbio-0020304-Wang1]\]) that prevent exonucleolytic digestion of their genomes during infection, and given its sequence similarity to other phage proteins, it is possible that D. radiodurans acquired DdrA from a phage during its evolution. Since inactivation of DdrA reduces, but does not eliminate, the DNA damage resistance of *Deinococcus,* we suggest that other proteins with complementary functions, possibly designed to bind DNA ends with different structures, are also encoded by this species, and the protection provided by these proteins contributes significantly to DNA damage tolerance. By itself, DdrA protein does not enhance the radiation resistance of E. coli strains in which it has been expressed (L. Alice Simmons and J. Battista, unpublished data). It seems likely that *D. radiodurans,* and other bacteria with similar capacities to survive high DNA damage loads, employs multiple systems to repair its DNA. The DNA end-protection system we have begun to explore may be supplemented by special genome architectures ([@pbio-0020304-Levin-Zaidman1]), traditional DNA repair systems (some with unusual properties \[[@pbio-0020304-Kim1]\]), and perhaps novel enzymatic systems not previously examined. Although we have detected no apparent enzymatic activities in DdrA to augment its DNA binding function, further work is needed to determine if DdrA contributes to single-stranded annealing or other potential DNA repair pathways. Bound to 3′ DNA ends, DdrA would be at a focus of DNA repair activity once genome restitution was initiated. The evolutionary relationship of DdrA to Rad52 may also telegraph a facilitating role in other DNA repair processes. Materials and Methods {#s4} ===================== {#s4a} ### Strains, growth conditions, and treatment {#s4a1} Strains and plasmids used in this study are described in [Table 2](#pbio-0020304-t002){ref-type="table"}. All genes are identified as described in the published genome sequence (<http://www.tigr.org/tigr-scripts/CMR2/GenomePage3.spl?database=gdr>). All strains derived from D. radiodurans were grown at 30 °C in TGY broth (0.5% tryptone, 0.3% yeast extract, and 0.1% glucose) or on TGY agar (1.5% agar). E. coli strains were grown in Luria-Bertani (LB) broth or on LB plates at 37 °C. Plasmids were routinely propagated in E. coli strain DH5αMCR. D. radiodurans cultures were evaluated for their ability to survive exposure to DNA-damaging agents in exponential growth (OD600 = 0.08 − 0.15, 5 × 10^6^ − 1 × 10^7^ cfu/ml). All cultures were treated at 25 °C. Gamma irradiation was conducted using a model 484R ^60^Co irradiator (J. L. Shepherd and Associates, San Fernando, California, United States) at a rate of 30 Gy/min. Resistance to mitomycin C was determined by adding 1 mg of mitomycin C (Sigma, St. Louis, Missouri, United States) to 1-ml broth cultures of the D. radiodurans strain. Aliquots of the treated culture were removed at one-half-hour intervals over the next 2 h, washed in 10 mM MgSO4, and plated on TGY agar to determine viability. ::: {#pbio-0020304-t002 .table-wrap} Table 2 ::: {.caption} ###### Strains and Plasmids :::  ::: ### Construction of TNK104, TNK106, and TNK110 {#s4a2} The genes *ddrA* and *recA* were disrupted by targeted mutagenesis using techniques described previously ([@pbio-0020304-Funayama1]). A deletion cassette was created for each locus and transformed into an exponential-phase D. radiodurans R1 culture. Recombinants were selected on TGY plates containing an appropriate antibiotic. Since D. radiodurans is multigenomic, individual colonies were screened to determine if they were homozygous for the disruption by isolating genomic DNA from putative recombinants and using a PCR-based analysis to determine whether the gene of interest had been deleted. Details for how each strain was generated are given below. The construction of TNK104 began with the creation of a drug cassette capable of conferring hygromycin resistance on *D. radiodurans.* The hygromycin B phosphotransferase gene *(hyg)* from pHP45omega-hyg ([@pbio-0020304-Blondelet-Rouault1]) was spliced to the 120 bp of sequence immediately upstream of the initiation codon of the *D. radiodurans katA* gene (DR1998) ([@pbio-0020304-Funayama1]), using primers whose sequences overlapped. Subsequently, the *katA--hyg* fusion product was joined to PCR fragments ([@pbio-0020304-Horton1]) derived from the sequence 1.0 kbp immediately upstream and 0.9 kbp immediately downstream of *ddrA*. This hybrid fragment was cloned into pGEM-T (Promega, Madison, Wisconsin, United States), creating pTNK205. pTNK205 was propagated in E. coli DH5α-MCR. The deletion of *ddrA* was accomplished by transforming ([@pbio-0020304-Earl2]) an exponential-phase R1 culture with linear pTNK205. Hygromycin-resistant recombinants were selected on TGY plates containing 37.5 μg/ml hygromycin. To confirm gene replacement, primers, which anneal outside the coding sequence of *ddrA,* were used to generate PCR fragments from genomic DNA from hygromycin-resistant colonies and R1. The purified PCR products were restricted with EcoRI and EcoRV. The *hyg* gene contains an EcoRI site, but *ddrA* does not. *ddrA* contains an EcoRV site, but *hyg* does not. In the recombinant, designated TNK104, a single 1.3-kbp fragment, corresponding to the *katA--hyg* cassette was amplified, whereas there was no trace of the 0.85-kbp fragment, indicative of *ddrA* amplification (see [Figure 1](#pbio-0020304-g001){ref-type="fig"}A). EcoRI cleaved the product amplified from TNK104 into 0.2-kbp and 1.1-kbp fragments, while the R1-derived product remained intact ([Figure 1](#pbio-0020304-g001){ref-type="fig"}B). EcoRV digested the amplicon from R1 into fragments of 0.4 kbp and 0.45 kbp, but it did not affect the TNK104-derived product ([Figure 1](#pbio-0020304-g001){ref-type="fig"}C). We conclude that TNK104 carries a deletion of the *ddrA* coding sequence marked by the *katA--hyg* cassette and that the strain is homozygous for the deletion. The *recA* deletion strain TNK106 was constructed in a manner similar to that of TNK104. Initially, the *katA* promoter of D. radiodurans was fused to the chloramphenicol acetyltransferase gene *(cat)* from pBC (Stratagene, La Jolla, California, United States). This drug cassette was then spliced to PCR products corresponding to genomic DNA sequences 1.6 kbp upstream and 1.2 kbp downstream of *recA* by overlap extension, before being cloned into pGEM-T. The resulting plasmid was designated pTNK210. An exponential-phase R1 culture was transformed with the replacement cassette from pTNK210, and chloramphenicol-resistant recombinants were selected on TGY plates containing 3 μg/ml chloramphenicol. Genomic DNA of each recombinant was amplified to determine if the *recA* coding region was deleted. Purified PCR products amplified using primers that anneal to sequences flanking *recA* were treated with PvuII and BglII. The *cat* gene carries a PvuII site, but *recA* does not. *recA* contains a BglII site, but *cat* does not. A 1.3-kbp fragment, corresponding to the *katA--cat* cassette, was obtained from a recombinant designated TNK106, but DNA from this recombinant did not generate the 1.5-kbp fragment corresponding to *recA* ([Figure 1](#pbio-0020304-g001){ref-type="fig"}A). Amplifications of genomic DNA from R1 only produced the 1.5-kbp fragments ([Figure 1](#pbio-0020304-g001){ref-type="fig"}A). The 1.3-kbp PCR product from TNK106 was cleaved by PvuII to 0.5-kbp and 0.8-kbp fragments, whereas the 1.5 kbp from R1 remained intact ([Figure 1](#pbio-0020304-g001){ref-type="fig"}C). BglII cut the R1-derived 1.5 kbp to fragments of 0.45 kbp and 1.05 kbp, but not the product from TNK106 ([Figure 1](#pbio-0020304-g001){ref-type="fig"}C). We conclude that *recA* has been replaced by *katA--cat* in TNK106 and that the strain is homozygous for this allele. TNK110 is a double mutant in which *recA* and *ddrA* are deleted. This strain was constructed by deleting *recA* from TNK104 using the protocol described for the creation of TNK106. The construct was verified by the scheme used to identify *ddrA* deletion in TNK104 and *recA* deletion in TNK106 (see [Figures 1](#pbio-0020304-g001){ref-type="fig"}A and [1](#pbio-0020304-g001){ref-type="fig"}C). ### Pulsed-field gel electrophoresis {#s4a3} After irradiation at 5.0 kGy, cells were collected by centrifugation (6,000*g,* 15 min, 4 °C) and resuspended in either TGY broth or 10 mM MgSO4 solution, before being placed in a shaking incubator at 30 °C for 24 h. Aliquots of these cultures were removed at various time points, and cells were washed in 0.9% NaCl and suspended in 0.125 M EDTA (pH 8.0) at a density of 5 × 10^8^ cells/ml. The suspensions were mixed with low-melting-point agarose (Sigma) to obtain a final concentration of 0.8% agarose. Agarose blocks containing the cell suspension were incubated overnight at 37 °C in 0.05 M EDTA (pH 7.5) containing 1 mg/ml of lysozyme. After lysozyme treatment, agarose plugs were placed in ESP buffer (EDTA 0.5 M \[pH 9--9.5\], 1% lauroyl sarcosine, 1 mg/ml proteinase K) at 50 °C for 6 h, followed by a 2-d incubation at 37 °C. Prior to digestion with restriction enzymes, agarose plugs were washed once with TE buffer (pH 7.5) containing 1 mM phenylmethylsulfonyl fluoride and then four times with TE buffer (pH 7.5). DNA contained within the agarose plugs was digested with 10 U of NotI restriction enzyme (New England Biolabs, Beverly, Massachusetts, United States) overnight at 37 °C. Restriction digests were analyzed on 1% agarose gels in 0.5X TBE, using a CHEF-MAPPER electrophoresis system (Bio-Rad, Hercules, California, United States) at 6 V/cm for 22 h at 12 °C, with a linear pulse ramp of 10--60 s and a switching angle of 120°. Gels were stained with water containing 0.5 μg/ml ethidium bromide for 20 min and destained for 10 min in water. ### Quantitative real-time PCR {#s4a4} The protocol followed was the same as that described previously ([@pbio-0020304-Earl1]). Total RNA was extracted from 1-l cultures of irradiated and nonirradiated exponential-phase D. radiodurans cultures using TRI Reagent (Molecular Research Center, Cincinnati, Ohio, United States) following manufacturer\'s instructions. Cell disruption was accomplished by adding 100 μ1 of 0.1-mm zirconia/silica beads (Biospec Products, Bartlesville, Oklahoma, United States) and TRI Reagent to the cell paste from 1 l of cells and vigorously agitating this mixture for 6 min with a vortex mixer. Two micrograms of each DNase I--treated, purified RNA sample was converted to cDNA using SUPERSCRIPT II RNase H^−^ Reverse Transcriptase (Invitrogen, Carlsbad, California, United States) combined with 25 pmol of random hexamers to initiate synthesis. Conditions for this reaction followed the manufacturer\'s instructions. Approximately 100 bp of unique sequence from the genes encoding DdrA (DR0423), RecA (DR2340), and glyceraldehyde 3-phosphate dehydrogenase (DR1343) were amplified using the following primer sets: DR0423up (5′-GGTGCAGGACCGACTCGACGCCGTTTGCC-3′), DR0423down (5′-CCTCGCGGGTCACGCCGAGCACGGTCAGG-3′), DR2340up (5′-GTCAGCACCGGCAGCCTCAGCCTTGACCTC-3′), DR2340down (5′-GATGGCGAGGGCCAGGGTGGTCTTGC-3′), and DR1343up (5′-CTTCACCAGCCGCGAAGGGGCCTCCAAGC-3′), DR1343down (5′-GCCCAGCACGATGGAGAAGTCCTCGCC-3′). The PCR reaction (50 μ1) for amplifying these genes contained the appropriate primers at a final concentration of 0.2 μM, 1 μ1 of the cDNA template, and SYBR Green PCR Core Reagents (Applied Biosystems, Foster City, California, United States). Amplifications were carried out by incubating reactions at 95 °C for 3 min prior to 40 cycles of 30 s at 95 °C, followed by 30 s at 65 °C and 30 s at 72 °C. Data were collected and analyzed at each 72-°C interval. Each 96-well plate consisted of standard curves for each primer set run in duplicate. Standard curves were constructed using cDNA obtained from the unirradiated wild-type organism. A dilution series (1 to 1 × 10^−4^) of each experimental sample was generated and run in duplicate. Negative controls without a cDNA template were run on every plate analyzed. All assays were performed using the iCycler iQ Real-Time Detection System (Bio-Rad). All data were PCR-baseline subtracted before threshold cycle values were designated and before standard curves were constructed. Mean concentrations of the transcripts in each sample were calculated from the standard curves generated using the *recA* primer set. Induction levels were determined by dividing the calculated concentration of transcript from the irradiated sample by the concentration of transcript from the unirradiated sample for each strain. The mean concentration of the *gap* transcript, a housekeeping gene whose expression is unaffected by ionizing radiation, was also determined before and after irradiation for each strain. ### DNA content measurement in TNK104 and R1 cells {#s4a5} Overnight cultures growing in TGY medium were harvested at room temperature. Control culture aliquots were fixed with 1% toluene (final vol/vol), shaken vigorously, and stored at 4 °C. The fixed bacteria were diluted (1/10, 1/100, and 1/1,000) in 3 ml (final volume) of dilution buffer: 10 mM NaCl, 6.6 mM Na~2~SO~4~, 5 mM N′-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES; pH 7.0). The remaining cultures were centrifuged for 20 min at 4 °C at 7,000 rpm. Bacterial pellets were washed twice and resuspended in 10 mM MgSO~4~ for γ irradiation. Cell suspensions were irradiated at 5,000 Gy and incubated at 30 °C for 120 h. Aliquots were removed immediately following irradiation, at 48 h, and at 120 h postirradiation. Cells were toluene-fixed as described above; 100 μ1 of DAPI (stock solution 3 μg/ml) was added to each dilution tube and mixed. The fluorescence intensity was determined after excitation at 350 nm by measuring emission at 450 nm. ### Cloning, overexpression, and purification of DdrA {#s4a6} The *ddrA* gene was amplified using the genomic DNA from D. radiodurans strain R1. PCR primers were designed according to the *ddrA* gene sequence annotated in the genomic bank (<http://www.ncbi.nlm.nih.gov>). The gene was cloned in E. coli overexpressing plasmid pEAW298. DdrA-overproducing cells were lysed with lysozyme, and the protein was precipitated from the supernatant by adding ammonium sulfate to 30% saturation. The protein was purified with DEAE and hydroxyapatite chromatography to greater than 99% purity. The identity of the purified protein was confirmed by N-terminal sequencing (Protein and Nucleic Acid Chemistry Laboratory, Washington University School of Medicine, St. Louis, Missouri, United States) and accurate mass determination (Biotech Center, University of Wisconsin, Madison, Wisconsin, United States). The protein was transferred into the storage buffer (20 mM Tris-acetate, 80% cation \[pH 7.5\]/50% glycerol \[w/v\], 0.5 M NaCl, 0.1 mM EDTA, and 1 mM DTT) and stored at −80 °C. ### Determination of the extinction coefficient for pure DdrA protein {#s4a7} The extinction coefficient for DdrA protein was determined using a modification of a published procedure ([@pbio-0020304-Marrione1]). UV absorbance spectra were measured with a Cary 300 dual-beam spectrophotometer (Varian, Palo Alto, California, United States). The temperature was maintained using a circulating water bath. Cell-path length and bandwidth were 1 cm and 0.5 nm, respectively. The extinction coefficient for native DdrA protein was determined in the storage buffer, by comparing the absorbance spectra of the native protein to the absorbance spectra of the protein denatured in 6 M guanidine hydrochloride (Gnd--HCl) in storage buffer. The extinction coefficients at 280 nm of glycyl-[L]{.smallcaps}-tyrosylglycine and *N*-acetyl-[L]{.smallcaps}-tryptophanamide in 6 M GND--HCl are 1,280 M^−1^cm^−1^ and 5,690 M^−1^cm^−1^, respectively ([@pbio-0020304-Edelhoch1]). In the DdrA protein there are five tyrosine, five tryptophan, and two cysteine residues in a protein with a total molecular mass of 23 kDa. Even if all cysteine residues were involved in disulfide bonds, the contribution of cystine to the absorbance of DdrA protein is predicted to be less than 1% and was neglected from our calculations. The extinction coefficient at 280 nm for denatured DdrA protein in ɛ~denat,\ 280\ nm~ = 5 × 5,690 + 5 × 1,280 = 3.485 × 10^4^ M^−1^cm^−1^. Absorbance spectra of native and denatured (6 M GND--HCl) DdrA protein were scanned at 25 °C, from 320 to 240 nm, for five different dilutions and with two different protein preparations. DdrA protein was diluted in storage buffer or storage buffer plus 6 M GND--HCl (final concentration) in a total volume of 80 μ1 and was preincubated at 25 °C for 5 min before scanning. Each dilution was carried out in triplicate, and the absorbance values at 280 nm were averaged. The concentrations of native and denatured protein were equal to each other in each scan at each dilution. The extinction coefficient of native DdrA protein at 280 nm was determined according to the expression ([@pbio-0020304-Gill1]): ɛ~nat,\ 280\ nm~ = ɛ~denat,\ 280\ nm~ × Abs~nat,\ 280\ nm~/Abs~denat,\ 280\ nm~. We used five determinations with two different protein preparations, yielding an average extinction coefficient of ɛ~nat,\ 280\ nm~ = 2.8728 ± 0.1999 × 10^4^ M^−1^cm^−1^ in storage buffer at 25 °C. The A~280~/A~260~ ratio for the native DdrA protein is 1.575 ± 0.00091. The error in both cases is 1 s.d. ### DNA binding assay {#s4a8} The duplex oligonucleotide with a 3′ single-stranded extension was hairpin-forming oligonucleotide A (5′-TTA ACG ACC GTC GAC CTG CAG GTC GAC GGT CGT TAA CGT CTC TCA GAT TGT-3′), which was labeled at the 3′ terminus with \[α-^32^P\]ddATP, using terminal transferase. After labeling, hairpin formation generated an 18-bp duplex hairpin with a 16-nt 3′ extension. The duplex oligonucleotide with a 5′ single-stranded extension was hairpin-forming oligonucleotide B (5′-CGT CTC TCA GAT TGT TTA ACG ACC GTC GAC CTG CAG GTC GAC GGT CGT TAA-3′). The oligo was labeled at the 5′ end using \[γ-^32^P\] ATP and polynucleotide kinase. After labeling, hairpin formation generated a DNA with 18 bp in the hairpin duplex and a 15-nt 5′ extension. A blunt-ended duplex DNA fragment was prepared by annealing oligonucleotide C (5′-GGT CTT TCA AAT TGT TTA AGG AAG AAA CTA ATG CTA GCC ACG GTC CGA GCC-3′) ^32^P-labeled at its 5′ end, with unlabeled oligonucleotide D (5′-GGC TCG GAC CGT GGC TAG CAT TAG TTT CTT CCT TAA ACA ATT TGA AAG ACC-3′). The single-stranded oligonucleotide was the end-labeled oligo C. EMSAs for DNA binding were carried out in 15-μl reaction mixtures containing the reaction buffer (40 mM Tris-acetate \[pH 7.5\],10% glycerol \[w/v\], 0.1 M NaCl, 0.1 mM EDTA, 1 mM DTT) and 0.7 nM (60 nM nt) ^32^P-labeled duplex DNA. The reaction was initiated by adding the DdrA protein to the required concentration. The reaction mixture was incubated at 30 °C for 30 min and loaded onto a 10% native polyacrylamide gel. The electrophoresis was performed in 1X TBE (89 mM Tris-borate \[pH8.3\], 2 mM EDTA) at room temperature. After the electrophoresis was completed, the gel was dried and exposed with a Phosphoimager (Molecular Dynamics, Sunnyvale, California, United States). ### Identification of DdrA protein in DNA--protein complex {#s4a9} The general strategy of this experiment was to incubate a DNA duplex with a 3′ extension with DdrA protein, resolve the protein complex in native PAGE, excise the complex from the gel, extract the protein from the slice, and analyze the protein in SDS-PAGE. If the protein is DdrA, it will comigrate with DdrA protein in SDS-PAGE. A ^32^P-labeled oligonucleotide (30 nt; 5′-GTG CGC TCC GAG CTC AGC TAC CGC GAG GCC-3′) was annealed with a longer unlabeled oligonucleotide (50 nt; 5′-GGC CTC GCG GTA GCT GAG CTC GGA GCG CAC GAT TCG CAC TGC TGA TGT TC-3′). Annealing was carried out in a 40-μ1 solution containing 0.5 μM of each oligonucleotide in 25 mM Tris HCl (pH 8), 50 mM NaCl, and 12.5 mM MgC1~2~. The solution was heated briefly at 100 °C, by transferring the closed tube to a beaker of boiling water, and allowed to cool slowly overnight. The tube was refrigerated for several hours and then stored at −20 °C until use. The resulting labeled duplex DNA with a 3′ extension (0.7 nM) was incubated with 4 μM DdrA protein under the DNA binding conditions described above. The mixture was loaded onto a 10% native polyacrylamide gel. Electrophoresis was performed as described above. The gel was exposed with X-ray film to map the position of the protein--duplex complex. The complex was cut out of the gel. The gel slice was frozen in liquid nitrogen and crushed into a slurry with a plastic stick. The slurry was mixed with an equal volume of SDS-PAGE loading buffer and boiled for 3 min. The mixture was loaded onto a 12% SDS-PAGE gel and the protein present compared to molecular weight standards and purified DdrA protein. ### Exonuclease assay {#s4a10} The duplex with a 3′ extension was prepared by annealing oligonucleotide A (5′-CTA GCA TTA GTT TCT TCC TTA AAC AAT TTG AAA GAC C-3′), which was labeled at the 5′ terminus with \[γ-^32^P\]ATP, and cold oligonucleotide B (5′-GGT CTT TCA AAT TGT TTA AGG AAG AAA CTA ATG CTA GCC ACG GTC CGA GCC-3′). The annealing generated a 14-nt 3′ extension at one end of the short duplex. Before adding the exonuclease, the ^32^P-labeled duplex (60 nM nt) was preincubated with the DdrA protein at the indicated concentration in 15 μl of the exonuclease reaction buffer (40 mM Tris-acetate \[pH 7.5\], 0.1 M NaCl, 10 mM MgC12, 0.1 mM EDTA, 1 mM DTT, 10% glycerol) at room temperature for 10 min. In the control experiment, the DdrA protein was replaced with bovine serum albumin. Exonuclease I was added to 200 U/ml and the reaction mixture was incubated at 37 °C for 30 min. After the incubation was complete, the reactions 5 and 6 were deproteinized with 0.2% SDS and 0.2 mg/ml proteinase K at 37 °C for 15 min. The DNA--protein complexes were resolved in the native polyacrylamide gel as above. Supporting Information {#s5} ====================== Accession Numbers {#s5a1} ----------------- The GenBank (<http://www.ncbi.nlm.nih.gov/Genbank/index.html>) accession numbers for the genes and gene products discussed in this paper are *ddrA*/DR0423 (NP\_294146), Erf (P04892), *gap/*DR1343 (NP\_295066), Rad52 (P06778), *recA/*DR2340 (NP\_296061), RecT (P33228), and Redβ (P03698). The work described in this article was supported by grants GM52725 from the United States National Institutes of Health (to MMC), and DEFG0201ER63151 from the United States Department of Energy (to JRB) and a MURI subcontract award (N000014-01-1-0852) from the Naval Research Laboratory (to JRB). The authors also thank L. Alice Simmons for assistance in generating the survival curves, and Elizabeth A. Wood for constructing the strains used to overproduce the DdrA protein. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** MMC and JRB conceived and designed the experiments. DRH, MT, SVS, EJ, and AME performed the experiments. DRH, MT, SVS, EJ, AME, MMC, and JRB analyzed the data. MMC and JRB wrote the paper. Academic Editor: Steve Elledge, Harvard Medical School ¤1 Current address: Department of Immunology and Medical Zoology, School of Medicine, Kagoshima University, Kagoshima City, Kagoshima, Japan ¤2 Current address: Proteomics, R&D, Promega Corporation, Madison, Wisconsin, United States of America ¤3 Current address: Department of Microbiology and Molecular Genetics, Harvard Medical School, Boston, Massachusetts, United States of America Citation: Harris DR, Tanaka M, Saveliev SV, Jolivet E, Earl AM, et al. (2004) Preserving genome integrity: The DdrA protein of Deinococcus radiodurans R1. PLoS Biol 2(10): e304. cfu : colony-forming unit *ddrA* : DNA damage response A gene DdrA : DNA damage response A protein EMSA : electrophoretic mobility-shift assay GND--HCl : guanidine hydrochloride nt : nucleotide SDS : sodium dodecyl sulfate

PubMed Central

2024-06-05T03:55:47.379848

2004-9-7

PMC515371

Defending against attack is one of the most important challenges facing any organism. But while sticks and stones may break the bones of a lion, microscopic threats such as bacteria require different weapons. And it\'s not just we humans who have this problem---insects are prey to bacterial infections too. Their immune systems, however, rely on a far simpler set of defenses than those found in mammals. Exactly how one insect immune system recognizes bacteria, and how it fights off the invader, is the subject of a new study in this issue by Johann Deisenhofer and colleagues. The fruitfly, Drosophila, has long been known to use a set of molecular sentries called "peptidoglycan recognition proteins," or PGRPs, that circulate in the fly\'s bloodstream. When a PGRP recognizes a bacterial invader, it triggers a cascade of events whose ultimate product is a group of antimicrobial compounds that attack and kill the bacteria. While the family of PGRPs has been extensively studied, exactly how they recognize their target bacteria has been less clear. At the cellular level, recognition requires contact, and the part of the bacterium the PGRP recognizes is, as its name implies, the peptidoglycan. A peptidoglycan is a special sort of molecular polymer found primarily on bacterial cell walls. Peptidoglycan forms when chains of sugar molecules (the glycans) are cross-linked by amino acids (the peptides) to form a meshwork that helps keep the bacterium from bursting under the osmotic strain of its contents. There are several types of peptidoglycans that differ in their precise sugar and amino acid constituents and in their ability to trigger the Drosophila defensive reaction. Deisenhofer and colleagues set out to determine whether this difference in triggering ability of particular peptidoglycans was linked to differences in the PGRPs that recognize them. To do this, they determined the three-dimensional structure of one PGRP, called PGRP-SA. They worked out not only the overall shape of PGRP-SA, but also which amino acids sat where on the convoluted surface of the protein. What they found on that surface was an extended groove down one entire side of the protein. To test whether this groove was indeed the recognition site for peptidoglycan, the group introduced a series of mutations to critical amino acids along the groove, testing each new form for its ability to bind peptidoglycan. Indeed, the binding and defense-triggering ability was worse for almost every mutant, demonstrating conclusively that the normal protein uses the groove to bind and recognize peptidoglycan. Then the team made a surprising discovery. They found that when PGRP-SA comes in contact with bacterial peptidoglycan, it begins to cleave the links between amino acids in the peptide portion of the peptidoglycan. This in itself is not so amazing---animals make plenty of peptide-cleaving proteins. But this protein has a difference, one which makes it unique in the animal kingdom.[](#pbio-0020313-g001){ref-type="fig"} ::: {#pbio-0020313-g001 .fig} ::: {.caption} ###### Stick model of the PGRP-SA residues chosen for mutational analysis :::  ::: To understand this difference, consider your two hands. They are mirror images of each other, alike yet not the same. No amount of twisting and turning will allow you to superimpose one exactly on the other---if you align the fingers, the knuckles will point in opposite directions, and if the knuckles point the same way, the fingers are all mismatched. This type of relationship between mirror images, called chirality (from the Greek for "hand"), is found in amino acids as well, a result of the three-dimensional geometry that radiates from their central atom. All the amino acids used by all known animal species are exclusively of the "left-handed" form, and the protein-digesting enzymes we make are designed specifically for these L-amino acids. Bacteria, however, link left-handed and right-handed amino acids together to form peptidoglycan. What Deisenhofer\'s team discovered was that unlike any other known animal enzyme, the Drosophila PGRP-SA was able to break apart this "L,D" (levo-dextro) linkage, making it, in their words, "the first eukaryotic protein exhibiting such an activity specific for peptide bonds existing only in prokaryotes." What does it all mean? Deisenhofer and colleages\' results are yet another demonstration that at the molecular level, understanding structure is the key to understanding function. They also show that when it comes to defense, it helps to be able to fight with both hands.

PubMed Central

2024-06-05T03:55:47.383828

2004-9-7

PMC515372

While the nucleus of a cell may be its command headquarters, mitochondria are equally vital---they are the power plants of the cell, and without them all cellular activity would quickly and irrevocably come to a halt. Testifying to their origins as once free-living bacteria, mitochondria have their own DNA, comprising 37 genes in humans on a single circular chromosome. Whether they invaded their ancestral hosts as parasites or were captured as subcellular collaborators, they have long since left their independent ways behind. Their meager complement of genes is far fewer than is needed to produce these complex organelles; it is clear from analyzing the nuclear genome that most of the mitochondria\'s presumed ancestral genes have been taken into the cell\'s nucleus, where they are under the strict control of their host. The transplanted mitochondrial genes have been faithfully doing their job under new management since they were first appropriated, probably hundreds of millions of years ago. However, not all of their DNA descendants have continued to make themselves so useful. For, in addition to many of the mitochondria\'s original genes, the human genome houses over 200 mitochondrial genetic fragments, useless pieces of code whose only remaining function is to be replicated generation after generation. Detritus from other sources is even more common within the genome, and most of it seems to be harmless. But in this issue, Ricchetti and colleagues show that mitochondrial fragments may not be quite so benign. They have continued to invade the human genome, even into the present day, and a large proportion of them take up residence within nuclear genes, possibly disrupting them and causing human diseases. Scanning the entire human genome, Ricchetti and colleagues found a total of 211 nuclear sequences of mitochondrial origin (NUMTs). Of these, they selected 42, which appeared to be the most recent integrations, for detailed study. Only 14 of them were also found in DNA from our closest relatives, chimpanzees, indicating that the rest arose after the human--chimp split approximately 5 million years ago. While 35 of the 42 were found in all humans tested, the rest were not, suggesting a still more recent origin for these among human populations. The authors also made two surprising discoveries about the location of these human-specific NUMTs. They were not evenly distributed across the entire genome; instead, for reasons that are unclear, there were a disproportionate number of them on two chromosomes---the Y chromosome, present only in males, and number 18. Furthermore, NUMTs were not randomly scattered among all the DNA of the chromosomes. Rather, they were much less likely to be found in non-coding "junk" DNA and much more likely to have inserted themselves within highly active genes. This phenomenon is likely to be related to the mechanism by which a NUMT enters the chromosome---it relies on the machinery that repairs breaks in the DNA, and these breaks are more common in genes that are frequently transcribed. Such insertions can cause disease, as shown by the recent discovery of a hemophilia patient with a NUMT interrupting his clotting factor gene. Much remains to be learned about the functional and temporal dynamics of NUMT insertions, but their potential for harm suggests that many NUMTS, unlike much of the rest of the flotsam that litters our genome, may be selected against quickly. Combined with their differential distribution among human ethnic groups, this may make them valuable markers for tracking both long- and short-term trends in human evolution and migration.

PubMed Central

2024-06-05T03:55:47.384328

2004-9-7

PMC515373

It is often said that after a nuclear catastrophe, cockroaches will inherit the earth, because they are so resistant to the harmful effects of ionizing radiation. But should the unthinkable come to pass, the bacterium Deinococcus radiodurans is likely to outlast even the cockroach. Its ability to endure radiation is truly impressive: it can withstand a dose a thousand times that which will kill a human. How it accomplishes this phenomenal feat of survival is the subject of a study in this issue by John Battista and colleagues at Louisiana State University in Baton Rouge and at the University of Wisconsin at Madison.[](#pbio-0020325-g001){ref-type="fig"} ::: {#pbio-0020325-g001 .fig} ::: {.caption} ###### D. radiodurans R1 :::  ::: While radiation damages many cellular components, it is the fracturing of the cell\'s DNA that is the most harmful. DNA breaks can be repaired, but in doing so, the cell is racing against time. The exposed free ends of the DNA invite digestion by the cell\'s own enzymes, called exonucleases. If the DNA is not stitched back together quickly enough, the exonucleases will degrade it past the point of repair, and the cell will ultimately succumb. Large doses of radiation can fracture a chromosome in thousands of places, far in excess of the repair ability of most cells. D. radiodurans, however, largely prevents exonuclease digestion, an ability which has previously been shown to be linked to the activity of a gene with the rather uninformative name of DR0423. But how, exactly, does this gene accomplish this life-saving feat? To answer this question, Battista and colleagues first showed that, following radiation exposure, DR0423 was upregulated 20- to 30-fold, and that deletion of the gene renders D. radiodurans susceptible to ionizing radiation. Together, these results clearly indicate that the DR0423 gene product is critical for protecting the bacterium. Based on this, they dubbed the gene *ddrA*, for "DNA damage response." They also found that DdrA, the protein encoded by *ddrA*, binds to single-stranded fragments of DNA, exactly like those found at the broken ends of the DNA double helix when damaged by radiation. Finally, they showed that when DdrA bound to these broken ends, they were protected from digestion by exonucleases. An important question about this system is what it is actually good for. Since the level of radiation tolerated by the bacterium is found nowhere on earth, of what use is such an efficient DNA protection system? The answer might be that it also protects D. radiodurans from the effects of desiccation, a condition much more common in the life of a bacterium, and one which also induces widespread DNA damage. While *ddrA* cannot prevent the damage, it can preserve the DNA from degradation until conditions once again allow the bacterium to function, and repair its DNA.

PubMed Central

2024-06-05T03:55:47.384832

2004-9-7

PMC515374

Scientists continue to argue the extent that human activities drive global warming, but few would argue that it exists. The International Panel on Climate Change predicts that greenhouse gases will increase global temperatures by 3.6 degrees F by 2100---a rise unprecedented over the past 10,000 years. What might the world look like as we approach that point? Wetlands will disappear. Floods, hurricanes, and droughts will become progressively more severe. Infectious diseases will increase in virulence and range. Montana\'s famed glaciers may all but disappear within 30 years. A quarter of species may vanish by 2050. While the effects of climate change on species\' geographic ranges and population dynamics have been studied to some extent, scientists know little about how species respond to climate change at the genetic level. In this issue, Elizabeth Hadly and colleagues analyze three different dynamic processes---environmental change, population response, and gene diversity fluctuations---and present evidence that climate change influences variation in genetic diversity. Focusing on two mammal species---the Montane vole and the northern pocket gopher---Hadly et al. asked how the two species responded to historical climate-induced habitat alterations in northwestern Wyoming. They gathered fossils from Yellowstone National Park\'s Lamar Cave, which contains a treasure trove of carbon-dated deposits that mirror the community of mammals in the area today. Comparing genetic material extracted from fossil samples from different time points over the past 3,000 years to genetic data taken from contemporary animals, Hadly\'s team tracked genetic changes in populations of the two species and used this information (along with relative fossil abundance and modern population density) to estimate changes in effective population size over time. (Effective population size refers to the number of individuals contributing genetic material to the next generation. Populations with a small effective population size, for example, would be highly vulnerable to environmental catastrophe.) The genetic and demographic data were then combined with environmental records to analyze the relationship between the factors. Studying these populations in space and time---an approach the authors call "phylochronology"---offers an opportunity to analyze the genetic diversity of a species against the backdrop of environmental fluctuation within an evolutionary time frame. It also suggests how microevolutionary forces---factors that affect genetic variation in populations over successive generations---shape genetic responses to climate change. Such evolutionary forces include mutation, genetic drift (the random gene fluctuation in small populations that stems from the vagaries of survival and reproduction), and gene flow (changes in the gene frequency of a population caused by migration).[](#pbio-0020338-g001){ref-type="fig"} ::: {#pbio-0020338-g001 .fig}  ::: The past 3,000 years includes two periods marked by dramatic climate change---the Medieval Warm Period and the Little Ice Age---that had different effects on local mammal populations depending on their habitat preferences. Habitat specialists, the vole and pocket gopher live in the wet mountain regions of western North America. Though both showed population increases during wetter climates and declines during warmer periods, Hadly et al. predicted the gene diversity fluctuations of the two species would differ based on their different ecological behaviors. And that\'s what they found: genetic response is tied to population size. Pocket gophers have low population densities, stick close to home, and are fiercely territorial, while voles live in high-density populations and range more widely. For the gophers, population declines resulted in reduced gene diversity; for the voles---which have a larger effective population size and greater dispersal between populations---population declines resulted in increased gene diversity. But what forces underlie these differences in genetic variation? A recent study suggests that migration (a primary agent of gene flow) is most common in and between low-density patches in vole populations, which implicates gene flow as the driver of gene diversity patterns. But the authors don\'t rule out selection as a possibility, and suggest how to go about resolving the question. Hadly et al. show that phylochronology opens a unique window onto the relatively recent evolutionary past and offers "the potential to separate cause from effect." They also conclude that "differences in species demography can produce differential genetic response to climate change, even when ecological response is similar." With a 3-degree temperature increase in just the past 50 years in the American West, conservation of biodiversity may well depend on such insights.

PubMed Central

2024-06-05T03:55:47.385407

2004-9-7

PMC515375

Millions of different proteins exist in nature, each with a unique structure that determines its function. Proteins can have different effects depending on where they are in the cell and which proteins or pathways they associate with. The number of proteins produced by a cell varies, but scientists estimate that the human genome produces some 100,000 proteins, and with thousands of proteins likely to be active in a single cell, it\'s inevitable that the molecular components of cellular pathways overlap. It is thought that this may be the case for a tumor suppressor named VHL (after its role in von Hippel-Lindau disease, an inherited cancer syndrome that predisposes affected individuals to kidney and vascular tumors).[](#pbio-0020339-g001){ref-type="fig"} ::: {#pbio-0020339-g001 .fig} ::: {.caption} ###### Genetic evidence for HIF-independent VHL-regulated pathways :::  ::: The broad strokes of VHL action have been outlined: VHL is a ubiquitin ligase, an enzyme that targets proteins for destruction. VHL\'s best characterized target is a transcription factor called hypoxia-inducible factor 1 (HIF-1). When oxygen levels drop below normal---a condition called hypoxia---HIF-1 proteins are not degraded and may enter the nucleus, where they trigger the transcription of roughly 100 genes whose proteins either increase oxygen delivery or engage metabolic pathways that help the cell adapt to hypoxia. Scientists have long suspected that VHL has other targets, yet only HIF-1 has been clearly established. Now Peter Ratcliffe and colleagues use genetic methods in the nematode Caenorhabditis elegans to provide direct evidence of a HIF-1-independent function of VHL. Previous studies have shown that when cells lack VHL proteins, HIF-1 is not degraded, resulting in the overexpression of HIF target genes. VHL-defective cells also show abnormalities in the extracellular matrix, the structural scaffolding that surrounds the cell. But it has not been clear whether these effects stem from HIF-1 dysregulation or something else. To disentangle the actions of the two proteins, Ratcliffe and colleagues compared the consequences of VHL protein inactivation in a variety of genetic backgrounds in C. elegans. Then they analyzed the gene expression profiles of each of these mutant strains to identify pathways that required VHL but not HIF-1. To their surprise, the authors found, "all of the VHL-regulated genes fell into one of two patterns." Their expression was either independent of HIF-1 and dependent on a range of genes associated with the extracellular matrix, or vice versa. What\'s more, these gene sets fell into distinct categories based on their chromosomal location, predicted functional similarities, and pattern of dysregulation. These results, Ratcliffe and colleagues conclude, reflect the disruption of "two discrete aspects of VHL function." One depends on HIF-1---inhibiting the transcription factor when oxygen concentrations are normal---and one doesn\'t; disruption of this HIF-1-independent function produces defects similar to those seen in mutants with defects in extracellular matrix assembly. These results also fall in line with other studies that have linked VHL deficiency to extracellular matrix defects, though the precise link remains unclear. It\'s also not clear whether this HIF-1-independent function means that VHL is still functioning as a ubiquitin ligase but targeting a different substrate or whether it represents a completely different function of VHL. For now, direct evidence of a HIF-1-independent pathway for VHL charts a clear path for researchers interested in pinning down the functions of this undoubtedly multidimensional ubiquitin ligase. It also gives VHL syndrome researchers---who have long suspected that other functions of the VHL tumor suppressor play a role in the onset of the disease---a promising lead to explore.

PubMed Central

2024-06-05T03:55:47.385978

2004-9-7

PMC515376

Milk, cheese, and yogurt are so ingrained in the diets of Europeans that it\'s easy to forget that their ancestors ever ate differently. But about 9,000 years ago, before the domestication of cows, sheep, and goats, milk was a staple only for babies. Back then---just as in most Asian and African cultures today---individuals lost their ability to digest lactose, a sugar found in milk, as they grew up. But with the domestication of animals, milk became abundant. Among herders, individuals had an advantage if they had versions of genes, also known as alleles, that allowed them to digest lactose into adulthood. They would tend to be healthier and reproduce more than those who could not digest lactose. Thus, by natural selection within herding groups, over generations those who could drink milk into adulthood became more common. Researchers have found the allele that allows adults to digest lactose, and it\'s one of the clearest signs of natural selection in humans. In groups with a history of herding, the vast majority of people have the allele, whereas in non-herding groups, most people lack it. Researchers can find such footprints of selection by comparing groups of people that have lived in different environments, for example, or have eaten different diets.[](#pbio-0020344-g001){ref-type="fig"} ::: {#pbio-0020344-g001 .fig}  ::: In this issue of *PLoS Biology*, Joshua Akey and colleagues report new-found signs of natural selection in several human genes---including a chunk of Chromosome 7 encompassing four genes, the largest footprint of selection found yet. The research group analyzed the complete sequences of 132 genes in a set of 23 European-Americans and 24 African-Americans. All these genes are involved in inflammation, blood clotting, or blood pressure regulation and were studied as part of a larger project looking for alleles that contribute to disease. In general, solid evidence of natural selection acting on genes is hard to find. The history of selection can be obscured by a variety of processes. For one, genes can undergo "neutral changes," in which some of its base pairs change, but without altering the sequence or function of the protein the gene codes for. Also, idiosyncrasies in the history of a population can leave marks on the gene pool. A lineage can go through a "bottleneck," for example, if a small group splinters off from a larger population and then later multiplies. In general, the splinter group won\'t perfectly represent the larger population, so the frequencies of alleles for many genes will be skewed in the splinter group\'s lineage. Having first ruled out irrelevant changes in genes and population history effects, Akey and colleagues found strong signs of natural selection only in the European-Americans, suggesting this group went through significant changes in climate, diet, or culture more recently than the African-American group. This fits with the well-accepted idea that European populations came from small groups that split off from the larger African population. The researchers find evidence for such an event in the European population about 40,000 years ago. They also estimate that the region of Chromosome 7 was subjected to strong selection around 10,000 years ago, roughly when European herders began drinking milk. Interestingly, two of these genes, *TRPV5* and *TRPV6*, limit the rate of calcium uptake, so selection on one or both of these genes in Europeans could have originated with herding. Recent studies also found *TRPV6* to be more active in prostate cancer cells. In addition, African-Americans suffer higher rates of prostate cancer, and Akey and colleagues found that European-Americans have alleles of *TRPV6* different from those of African-Americans. Given this evidence, the researchers suggest that this gene may be involved in susceptibility to prostate cancer. This research could therefore shed light on the evolution of complex diseases such as cancer and why different populations suffer different rates of disease.

PubMed Central

2024-06-05T03:55:47.386611

2004-9-7

PMC516016

Background ========== We have found the HotDog domain, as we suggest calling the Hotdog fold, to be widespread in eukaryotes, bacteria, and archaea and to be involved in a range of cellular processes, from thioester hydrolysis, to phenylacetic acid degradation and transcriptional regulation of fatty acid biosynthesis. We present the superfamily and its functional subfamilies here. The Hotdog fold was first observed in the structure of *Escherichia coli*β-hydroxydecanoyl thiol ester dehydratase (FabA), where Leesong *et al*. noticed that each subunit of this dimeric enzyme contained a mixed α + β \'hot dog\' fold \[[@B1]\]. They described the seven-stranded antiparallel β-sheet as the \'bun\', which wraps around a five-turn α-helical \'sausage\', see Figure [1](#F1){ref-type="fig"}. This characteristic fold has been found in a number of other enzymes, including: 4-hydroxybenzoyl-CoA thioesterase (4HBT) from *Pseudomonas*sp. strain CBS-3 \[[@B2]\] and *Arthrobacter*sp. strain SU \[[@B3]\], a novel gentisyl-CoA thioesterase from *Bacillus halodurans*\[[@B4]\] and in *Escherichia coli*thioesterase II \[[@B5]\]. Results and Discussion ====================== Although several proteins are now known to contain a Hotdog fold from structural analysis it has not to our knowledge been demonstrated that these proteins can be related to each other by sequence similarity. We have attempted to unify these structurally related proteins using a sequence analysis approach. Using sequence analysis means that we will identify additional proteins that are likely to contain a Hotdog fold. We have used the PSI-BLAST program \[[@B6]\] and used a representative of each Hotdog fold of known structure as a query against Swiss-Prot and TrEMBL protein database \[[@B7]\]. We used the sequences of the following PDB entries: 1C8U \[[@B5]\], 1IQ6 \[[@B8]\], 1LO7 \[[@B9]\], 1MKB \[[@B1]\], 1NJK \[[@B10]\], 1O0I \[[@B11]\] and 1PSU \[[@B12]\]. These searches have uncovered many novel members of this superfamily as well as finding links between the known structures with a Hotdog fold (see Table [1](#T1){ref-type="table"} and [Additional file 1](#S1){ref-type="supplementary-material"} ). The Pfam database \[[@B13]\] contains a Thioesterase superfamily with 697 members, each member containing a 4HBT domain (accession: PF03061) corresponding to the HotDog domain. The SCOP database \[[@B14]\] contains a thioesterase/thiol ester dehydrase-isomerase superfamily, divided into 5 families, namely the 4HBT-like, beta-hydroxydecanoyl thiol ester dehydrase, Thioesterase II (TesB), MaoC dehydratase and PaaI/YdiI-like families. Our searches have found a total of 1357 proteins (see [Additional file 2](#S2){ref-type="supplementary-material"} ) to be related to the known structures of HotDog domain proteins. We took these proteins and clustered them using single linkage clustering to define subfamilies with common functions. This clustering puts 1293 (95%) of the sequences into 85 clusters (see [Additional file 3](#S3){ref-type="supplementary-material"} ). The HotDog domain is found to be associated with a wide range of other domains. The various domain architectures are shown schematically in Figure [2](#F2){ref-type="fig"}. We describe the 17 subfamilies (Table [1](#T1){ref-type="table"}) that have some experimental characterisation, below. The 17 subfamilies contain 909 proteins or 67 % of the total number of HotDog domain proteins. 384 (28 %) proteins cluster into the remaining groups, which contain predominantly hypothetical proteins or proteins that have no known function. They are not discussed here but we hope that our analysis may help in identifying functions for these proteins. Finally we have generated a Hidden Markov Model (HMM) library by concatenating together the HotDog domain sequences of the 85 clusters generated in our analysis (see [Additional file 4](#S4){ref-type="supplementary-material"} ). This library can be used in conjunction with the HMMER program \[[@B15]\] to search for HotDog domain(s) in any protein of interest. Acyl-CoA thioesterase subfamily ------------------------------- The largest subfamily represents over a hundred acyl-CoA thioesterases that are widespread throughout the prokaryotic kingdom, with members also found in eukaryotes. This group of enzymes catalyze the hydrolysis of acyl-CoA thioesters to free fatty acids and coenzyme A (CoA-SH.) \[[@B16]\]. The subfamily includes thioesterases with activity towards medium and long chain acyl-CoAs (medium chain acyl-CoA hydrolase and cytosolic long-chain acyl-CoA hydrolase/brain acyl-CoA hydrolase (BACH) respectively) and also cytoplasmic acetyl-CoA hydrolase (CACH), which hydrolyzes acetyl-CoA to acetate and CoA-SH. Brown-fat-inducible thioesterase (BFIT), a cold-induced protein found in brown adipose tissue (BAT) \[[@B17]\] is also included in this group. Both BFIT and CACH possess a StAR-related lipid-transfer (START) domain \[[@B18]\] that is involved in lipid binding, consistent with the role of BFIT and CACH in lipid metabolism. Duplication of the HotDog domain and recruitment of the START domain seems to be a mammalian innovation. FabZ like dehydratase subfamily ------------------------------- Members of this subfamily are found in a wide range of bacteria and sporadically in eukaryotes. In *E. coli*the products of the fab operon catalyze the four sequential reactions necessary for each round of fatty acid elongation \[[@B19]\]. The third step in each cycle of fatty acid elongation involves the dehydration of the β-hydroxyacyl-ACP protein intermediate by β-hydroxyacyl-\[acyl carrier protein\] dehydratase (FabZ) to give *trans*-2-decenoyl-ACP. FabZ is effective at dehydrating both short-chain and long chain saturated and unsaturated pathway intermediates. This subfamily also contains a dehydratase component of the coronafacic acid (CFA) biosynthetic cluster encoded by the cfa2 gene \[[@B20],[@B21]\]. CFA is the polyketide constituent of a phytotoxin called coronatine, which is a virulence factor of *Pseudomonas syringae*, a plant pathogen that causes disease in many agriculturally important plants \[[@B20]\]. MaoC dehydratase-like subfamily ------------------------------- The *mao*C gene exists as an operon with the *maoA*gene in *E. coli*and is an enoyl-CoA hydratase involved in supplying (R)-3-hydroxyacyl-CoA from the fatty acid oxidation pathway to polyhydroxyalkanoate (PHA) biosynthetic pathways in *fadB*mutant *E. coli*strains. It was identified through its homology to *P. aeruginosa*(R)-specific enoyl-CoA hydratase (PhaJ1) \[[@B22]\]. PHAs are polyesters of (R)-hydroxyalkanoic acids, synthesized by numerous bacteria as an intracellular carbon and energy storage material in times of excess carbon sources \[[@B23]\], with intermediates of fatty acid metabolism such as enoyl-CoA, (S)-3-hydroxyacyl-CoA, and 3-ketoacyl-CoA acting as precursors for PHA biosynthesis \[[@B22]\]. The crystal structure of the (R)-specific enoyl-CoA hydratase (phaJ) from the *Aeromonas caviae*has shown that this enzyme also contains a Hotdog fold/domain \[[@B8]\]. The *E. coli*MaoC C-terminal HotDog domain is most likely responsible for its enoyl-CoA hydratase actvity. MaoC also contains an N-terminal short-chain dehydrogenase domain, involved in catalysing dehydrogenation of a variety of aliphatic and aromatic aldehydes using NADP as a cofactor. This subfamily also includes the human 17 β-hydroxysteroid dehydrogenase (17 β HSD) type 4, one of four different human 17 β HSDs that catalyze the redox reactions at position C17 of steroid molecules, one of the final steps in androgen and estrogen biosynthesis \[[@B24],[@B25]\]. We also include a NodN-like sub-subfamily here that is found in another cluster containing several other MaoC proteins. *Rhizobium*and related species form nodules on the roots of their legume hosts, a symbiotic process that requires production of Nod factors, which are signal molecules involved in root hair deformation and meristematic cell division \[[@B26]\]. The nodulation gene products, including NodN, are involved in producing the Nod factors, however the role played by NodN is unclear. YbgC-like subfamily ------------------- This subfamily contains a large number of proteins about which very little is known except for the YbgC protein. The YbgC protein of the *tol-pal*cluster in the gamma-proteobacterium *Haemophilus influenzae*\[[@B27]\] has been shown to catalyze the hydrolysis of short-chain aliphatic acyl-CoA thioesters. The *tol-pal*cluster is present in many Gram-negative bacteria and is important for the maintenance of cell envelope integrity \[[@B28]\] and this operon is well conserved across gram-negative bacteria. Therefore we hypothesize that uncharacterized members of this subfamily are thioesterases. The Asp^17^residue is conserved in YbgC from *Haemophilus influenzae*and *Pseudomonas aeruginosa*, along with the backbone amide NH of Tyr^24^, suggestive of a nucleophilic attack mechanism very similar to the *Pseudomonas*sp. strain CBS-3 thioesterase mechanism discussed below in the 4HBT class I section. FabA-like subfamily ------------------- The dehydration of the β-hydroxyacyl-ACP protein intermediate during the third step in each cycle of fatty acid elongation can be catalyzed by β-hydroxydecanoyl-ACP dehydratase/isomerase (FabA), as well as by FabZ, to give *trans*-2-decenoyl-ACP. FabA is uniquely able to isomerise trans-2-decenoyl-ACP to cis-3-decenoyl ACP, initiating unsaturated fatty acid biosynthesis \[[@B19]\] and is specific for acyl ACPs of 9--11 carbons in length. Polyketides are a large and structurally diverse class of natural products, produced mainly by soil-dwelling bacteria such as *Pseudomonas*spp. and *Streptomyces*spp. They include clinically useful drugs such as the antibiotic erythromycin A and the immunosuppressants FK506 and rapamycin. The biosythesis of polyketides is very similar to that of fatty acids \[[@B21]\] and polyketide synthases (PKSs) have been classified as type I or type II according to fatty acid synthase (FAS) similarity. Most bacteria and plants use a highly conserved type II FAS system, which uses a distinct enzyme for each reaction. This is in contrast to the mammalian type I system (also used by fungi and some mycobacteria), which uses one multifunctional polypeptide to catalyze all reactions \[[@B29],[@B30]\]. The HotDog domain is found in type II fatty acid synthesis in bacteria (FabA/FabZ), but also in a small number of bacterial polyketide synthases that are of the type I, being composed of several modules \[[@B31]\] such as β keto-acyl synthases and omega-3 polyunsaturated fatty acid synthase (PfaC). The marine bacteria *Shewanella*sp. SCRC-2738, *Moritella marina*strain MP-1 and *Photobacterium profundum*strain SS9 contain an eicosapentaenoic acid (EPA) biosynthetic cluster (*pfaA-D*), responsible for the synthesis of this omega-3 polunsaturated fatty acid (PUFA), \[[@B32],[@B33]\]. The PfaC protein contains two HotDog domains (see Figure [2](#F2){ref-type="fig"} for the domain organisation found in *P. profundum*), which are also found in the eukaryotic marine protist, *Schizochytrium*, suggesting that the PUFA synthetic cluster has undergone lateral gene transfer \[[@B32]\]. This subfamily also includes several fatty acid synthase proteins from bacteria, such as *Mycobacterium bovis*fatty acid synthase. This multifunctional protein is capable of catalysing de novo synthesis and chain elongation of fatty acids \[[@B34]\] and has a very similar domain architecture to the polyunsaturated fatty acid synthases, as it contains an acyl-transferase, β-keto acyl synthase N and C-terminal domains (see Figure [2](#F2){ref-type="fig"}). The catalytic residues of FabA\'s bifunctional active site are His^70^and Asp^84^, His^70^is conserved in FabZ dehydratase, but Asp^84^is replaced with Glutamate. This replacement may be responsible for FabZ\'s inability to catalyze the isomerization reaction \[[@B1]\]. Fat subfamily Acyl-ACP thioesterases ------------------------------------ In plants, fatty acid synthesis occurs in the stroma of plastids, where the acyl chains are bound to the acyl carrier protein (ACP) during extension cycles \[[@B35]\]. Acyl-ACP thioesterases terminate fatty acid synthesis in plants by hydrolysing the thioester bond existing between an acyl moiety and the ACP \[[@B36]\]. In higher plants acyl-ACP thioesterases have been classified into two gene classes, *fatA*and *fatB*, based on sequence similarity and substrate specificities \[[@B37],[@B38]\]. *Arabidopsis*FatA displays highest activity towards oleoyl-ACP whereas *Arabidopsis*FatB is most active towards palmitoyl-ACP \[[@B37]\]. This subfamily contains both FatA and FatB members \[[@B35]\]. The proteins in this subfamily range in length from 240 to 400 amino acids and therefore we hypothesized that they might contain two HotDog domains, located at the N and C teminal halves. By splitting the sequence of proteins from this subfamily into an N-terminal half and C-terminal half we were readily able to detect the relationship to other subfamilies using PSI-blast with query proteins such as Q899Q1 and Q42714, confirming our hypothesis. TesB-like subfamily ------------------- This subfamily contains the *E. coli*medium chain length acyl-CoA thioesterase II \[[@B5]\] encoded by the *tesB*gene \[[@B38]\], which is a close homolog of the human thioesterase II (hTE) enzyme. hTE catalyzes the hydrolysis of palmitoyl-CoA to CoA and palmitate and was identified as a human T cell protein that binds to the myristoylated HIV-1 Nef protein, correlating with Nef-mediated CD4 down regulation \[[@B39]\]. hTE could regulate targeting of the cytoplasmic Nef protein to the plasma membrane, which is dependent on a lipid modification, i.e. a myristoylation anchor and recombinant hTE shows maximal activity with myristoyl-CoA \[[@B39]\]. However further studies have shown that hTE localizes to peroxisomes \[[@B40],[@B41]\], dependent on a C-terminal peroxisomal targeting sequence, SKL, and coexpression of Nef and hTE results in relocation of Nef to peroxisomes, so the role of Nef and hTE during HIV infection remains unsolved. The catalytic site of *E. coli*thioesterase II was identified by site directed mutagenesis and involves a hydrogen-bonded triad of Asp^204^, Thr ^228^, and Gln ^278^, which synergistically activate a water molecule for nucleophilic attack of the carbonyl thioester carbon of medium chain length acyl-CoA substrates \[[@B5]\]. This is a novel reaction mechanism for a thioesterase and differs from the nucleophilic mechanisms used by β-hydroxydecanoyl dehydratase and 4HBT thioesterase in both *Pseudomonas*and *Arthrobacter*discussed below. This subfamily is found in bacteria and eukaryotes. 4HBT class II subfamily ----------------------- This subfamily includes 4-hydroxybenzoyl CoA thioesterase (4HBT) from *Arthrobacter*sp. strains SU and TM1 encoded by the *fcbC*gene \[[@B3]\]. The *Pseudomonas*thioesterase uses the Asp^17^residue to mediate the hydrolysis reaction as discussed below in the 4HBT class I section. Gln^58^from *Arthrobacter*corresponds to the Asp^17^residue in *Pseudomonas*but inspection of the *Arthrobacter*strain SU active site has revealed the catalytic base (or nucleophile) to be Glu^73^, on the opposite side of the substrate binding pocket to Asp^17^\[[@B3]\]. Also the *Pseudomonas*thioesterase dimers form a tetramer with their long α-helices facing inwards, in contrast to *Arthrobacter*thioesterase where the dimers form a tetramer with their long α-helices facing outwards \[[@B3]\]. In *Pseudomonas*and *Arthrobacter*thioesterases, the 4-hydroxyphenacyl moieties are positioned in such an orientation that the thioester C = O interacts with the α-helical N-terminus by means of hydrogen bonding to a backbone amide NH, on Tyr^24^in *Pseudomonas*and Gly^65^in *Arthrobacter*, and it is this contact that results in polarization of the C = O for nucleophilic attack \[[@B3]\]. While the structure of *Arthrobacter*sp. strain SU thioesterase displays a similar Hotdog-fold topology to the 4HBT class I *Pseudomonas*enzyme, the enzymes differ at the level of catalytic platform, CoA binding site and quaternary structure \[[@B3],[@B42]\]. This is not an unexpected finding as Todd *et al.*have found that 12 of the 31 superfamilies they analyzed displayed positional variation for residues playing equivalent catalytic roles \[[@B43]\]. A surprising inclusion in this subfamily is the ComA2 protein from *Bacillus subtilis*. ComA is a response regulator and transcription factor \[[@B44]\] that together with the histidine kinase, ComP, constitutes a two-component signal transduction system required for the development of competence. The *com*A locus is composed of two ORFs. ComA2 is cotranscribed with ComA1, which is required for competence while ComA2 is not \[[@B45]\], and so the role of the HotDog domain in this protein remains a mystery. PaaI subfamily -------------- The phenylacetic acid (PA) catabolic pathway in *E. coli*has been characterised and found to contain 14 genes, allowing catabolism of this aromatic compound into likely Krebs cycle intermediates \[[@B46]\]. The *paa*operon in *E. coli*encodes PaaI, which is probably a thioesterase involved in the catabolism of PA. The catabolism of phenylacetic acid (PA) in *E. coli*begins with an activation step where Phenylacetyl-CoA ligase, PaaK, converts phenylacetate into Phenylacetyl-CoA. 4-chlorobenzoate-CoA ligase catalyzes a similar reaction at the first step of the 4-chlorobenzoate-degradation pathway. The thioesterase, PaaI, may be involved in a reaction similar to the last step in the degradation of 4-chlorobenzoate (see 4HBT class I below), however this remains to be demonstrated. FapR subfamily -------------- This small subfamily is restricted to firmicutes. FapR is a highly conserved transcriptional regulator found in many gram-positive organisms, including all species of *Bacillus*\[[@B47]\]. It controls expression of genes involved in type II fatty acid and phospholipid biosynthesis, by binding to a consensus promoter sequence of the fap regulon and acting as a negative regulator. Malonyl-CoA, an intermediate in the lipid biosynthetic pathway, controls FapR. The HotDog domain has likely retained its substrate specificity for malonyl-CoA, but appears to have lost its catalytic ability, in common with the ligand binding domain of other transcriptional regulators. FapR contains a helix-turn-helix motif at the N-terminus (see Figure [2](#F2){ref-type="fig"}), which is similar to the DeoR transcriptional regulator family (data not shown), consistent with its role as a DNA binding protein. 4HBT class I subfamily ---------------------- The crystal structure of 4HBT from the soil-inhabiting bacterium *Pseudomonas*sp. strain CBS-3 \[[@B2]\] has helped define the HotDog domain. A lot of attention has been focused on this microorganism because of its ability to survive on 4-chlorobenzoate (4CBA) as its only source of carbon \[[@B48]\]. 4CBA is a by-product of microbial degradation of industrial pollutants such as DDT and polychlorinated biphenyl herbicides \[[@B49]\] and this strain of *Pseudomonas*may be used as a bioremediation agent for degrading 4CBA. *Pseudomonas*sp. strain CBS-3 contains an *fcb*operon responsible for hydrolytic dechlorination of 4CBA, with 4CBA-CoA ligase (FcbA), 4CBA-CoA dehalogenase (FcbB), and 4HBT (FcbC) catalyzing sequential reactions that result in the degradation of 4CBA to 4-hydroxybenzoate. The thioesterase catalyzes the third step in the degradation pathway, which is the hydrolysis of the 4-hydroxybenzoyl-CoA thioester moiety to give 4-hydroxybenzoate and CoA \[[@B50]\]. 4HBT from *Pseudomonas*sp. strain DJ-12 \[[@B51]\] is also found in this subfamily. The organization of the *fcb*operon in strain DJ-12 is different from that observed in strain CBS-3. The *fcb*genes are organised as B-A-C in both strains but strain DJ-12 has three ORFs between A and C called T1, T2, and T3 that are unique to this strain. These three genes are similar to the C4-dicarboxylate transport system in *Rhodobacter capsulatus*, suggesting that they may encode membrane proteins involved in the uptake of 4CBA \[[@B51]\]. This is in contrast to the gene organisation observed in the 4HBT class II, where *Arthrobacter*sp. strain SU and strain TM1 have an A-B-C order \[[@B51]\]. There is a duplication of the cluster in strain SU, where it is found on a plasmid, whereas only one copy exists in strain TM1, where it is located chromosomally. Both operons contain a T gene located at the end of the cluster, possibly involved in 4CBA uptake. *Bacillus halodurans*C-125 contains a gene called BH1999, encoding a novel gentisyl-CoA thioesterase, which catalyzes the hydrolysis of gentisyl-CoA (2,5-dihydroxybenzoyl-CoA)\[[@B4],[@B52]\] to yield gentisate (2,5 dihydroxybenzoate). BH1999 is found in a gentisate oxidation pathway gene cluster in *B. halodurans*. Gentisate has been implicated as an intermediate in the degradation of several industrial aromatic compounds \[[@B4]\]. Gentisyl-CoA thioesterase and 4HBT from *Pseudomonas*perform different physiological functions but remain in the same subfamily because they are highly related. The active site residues Asp^16^and Asp^31^of gentisyl-CoA thioesterase align with Asp^17^and Asp^32^of 4HBT. These are crucial residues that are proposed to function in nucleophilic catalysis and substrate binding respectively. Loss of Asp^17^in the *Pseudomonas*enzyme effectively halts catalysis, while loss of the corresponding Asp^16^residue to the *Bacillus halodurans*enzyme only reduces its catalytic rate by 230-fold, perhaps indicating that the hydrolysis reaction does not proceed through an Asp^16^-mediated nucleophilic attack mechanism previously proposed for Asp^17^\[[@B53],[@B4]\]. Asp^17^in *Pseudomonas*strain CBS-3 has been suggested to participate in nucleophilic catalysis rather than general base catalysis based on the following observations. The Asp^17^carboxylate is located at a distance of 3.2 Å from the substrate C = O thioester bond, its aligned trajectory and the absence of a water molecule near the reaction centre are all suggestive of a role for Asp^17^as a catalytic nucleophile \[[@B9],[@B53]\]. Asp^32^in Pseudomonas interacts with the benzoyl OH of 4-hydroxybenzoyl-CoA \[[@B9]\] and perhaps Asp^31^plays a similar role. Other subfamilies/ members -------------------------- In the above sections we have described the 11 subfamilies that have some functional characterization. In this section we describe the other 6 subfamilies that have no functional characterization, except they are associated with other domains or have been structurally characterized. The CBS associated subfamily contains the hypothetical protein BH3175 from *Bacillus halodurans*. The BH3175 protein contains two homologous copies of the CBS domain \[[@B54]\]. Scott *et al*. have recently shown that tandem pairs of CBS domains act as sensors of cellular energy status by binding AMP, ATP, or S-adenosyl methionine and mutations in CBS domains impair this binding in several hereditary disorders \[[@B55]\]. Although we do not know the substrate or activity of this subfamily of the HotDog superfamily, we can suggest that this step is regulated in an energy dependent manner by the CBS domains. 3-hydroxyacyl-CoA dehydrogenase is an enzyme involved in fatty acid metabolism, catalyzing the reduction of 3-hydroxyacyl-CoA to 3-oxoacyl-CoA \[[@B56]\]. The hydroxyacyl-CoA dehydrogenase-associated subfamily includes 3-hydroxyacyl-CoA dehydrogenase from *Agrobacterium tumefaciens*strain C58, which contains a HotDog domain at its C-terminus and the two domains (3HCDH\_N and 3HCDH) associated with 3-hydroxyacyl-CoA dehydrogenase activity are located at the N-terminus and central portion of this protein. The combination of activities may allow substrate to be passed from one domain to the next. Other subfamilies in the superfamily include the YiiD protein from *E. coli*, where an acetyltransferase domain is fused. The human mesenchymal stem cell protein DSCD75 and its counterpart in mouse also contain a HotDog domain. A Structural proteomics project has shown that the conserved hypothetical *E. coli*protein YbaW contains a Hotdog fold \[[@B10]\]. Finally the *Ralstonia solanacearum*hypothetical protein RSp0367, containing a HotDog domain and two AMP-binding domains, found in proteins involved in ATP-dependent covalent binding of AMP to their substrate, is a member of another subfamily. Domain fusion events -------------------- It has been shown that proteins that are functionally linked are occasionally found to be fused in various genomes. These fusion proteins have been termed Rosetta proteins \[[@B57],[@B58]\] and can be used to predict the functional linkages of proteins with each other. The HotDog domain superfamily contains several rosetta proteins where the fused proteins are also found unfused in other genomes. In these cases they are adjacent to each other in known operons. The examples found in the HotDog superfamily are shown in Figure [3](#F3){ref-type="fig"} and are described briefly here. Within the FabZ subfamily the LpxC deacetylase domain (UDP-3-O-acyl N-acetylglucosamine deacetylase) is fused to the FabZ-like HotDog domain in *Chlorobium tepidium*(see Figure [3a](#F3){ref-type="fig"}). LpxC catalyzes the N-deacetylation of UDP-3-O-acyl N-acetylglucosamine deacetylase, the second and committed step in the biosynthesis of lipid A, which anchors lipopolysaccharide (LPS) in the outer membranes of most gram-negative bacteria \[[@B59]\]. The unfused proteins are found adjacent in operons from several species of chlamydia and cyanobacteria. In the 4HBT class II subfamily we observed the order of the operon is ligase(A)-dehalogenase(B)-thioesterase(C). In *Bacteroides thetaiotaomicron*there is a Rosetta protein that contans a haloacid dehalogenase-like hydrolase domain (see Figure [3b](#F3){ref-type="fig"}). This domain architecture is similar to the *fcb*operon structure in *Arthrobacter*, with a dehalogenase-like hydrolase (HAD) domain and a HotDog domain (see Figure [3](#F3){ref-type="fig"}) i.e. it represents a fusion of the *fcbB*and *fcbC*gene products to form a novel protein in *B. thetaiotaomicron*. The final domain fusion is in the 3-hydroxyacyl-CoA dehydrogenase from *Agrobacterium tumefaciens*strain C58, which possesses the HotDog domain, 3HCDH\_N domain (3-hydroxyacyl-CoA dehydrogenase, NAD binding) and 3HCDH (3-hydroxyacyl-CoA dehydrogenase, C-terminal domain) domain (see Figure [3c](#F3){ref-type="fig"}). This may represent a fusion of the PaaC and PP3281 proteins in the gamma-proteobacterium *Pseudomonas putida*2440 phenylacetic acid degradation operon. These fusion events suggest that the domain fusion process can occur in a simple scheme with two distinct phases. Firstly, two proteins are recombined into adjacent positions in an operon. Secondly, the two genes are then fused by a process of mutation that removes the stop codon at the end of the first gene and maintains reading frame through the second gene \[[@B60],[@B61]\]. Sequence motifs --------------- The MASIA program \[[@B62]\] was used to search for HotDog domain motifs in the aligned sequences of the 17 subfamilies. The various motifs are found in [Additional file 5](#S5){ref-type="supplementary-material"} . It must also be noted that the PROSITE database release 18.29 \[[@B63]\] contains a consensus sequence motif (PS01328), called the 4-hydroxybenzoyl-CoA thioesterase family active site, and this is found in 29 Swiss-Prot, TrEMBL and TrEMBL-NEW entries cross-referenced with PS01328. This consensus pattern, \[QR\]-\[IV\]-x(4)-\[TC\]-**D**-x(2)-G \[IV\]-V-x-\[HF\]-x(2)-\[FY\], where D is the active site residue, is found in the YbgC-like subfamily and in the 4HBT-I subfamily. 19 of the 29 members are found in the YbgC-like group and 3 in the smaller 4HBT-I group. The remaining 7 proteins are scattered in various clusters consisting of hypothetical or unknown proteins. We have found, using MASIA, that this motif is found in the entire YbgC and 4HBT-I subfamilies, extending the number of proteins containing this motif to 107. We have also identified a HGG motif in the 4HBT-II and PaaI subfamilies. This motif is HGGAS-x-ALA**E**in the 4HBT-II subfamily and HGG-x-IF-x-LA**D**in PaaI members. The active site residue, Glu^73^, is known for 4-hydroxybenzoyl-CoA thioesterase from *Arthrobacter*sp. Strain SU, however the active site for *E. coli*PaaI is not known and we suggest that it is Asp^61^in the HGG motif above, which is 100 % conserved in all members of this subfamily (see [Additional file 6](#S6){ref-type="supplementary-material"} ). Conclusions =========== We have defined and analyzed the HotDog domain superfamily and in our analysis of this superfamily we have found 18 different domain architectures and defined 17 subfamiles. We have also investigated the domain organisation and the role that this plays in generating functionally diverse enzymatic and nonenzymatic functions based on the HotDog fold. Domain duplication, domain recruitment and incremental mutation have been key to the evolution of this superfamily. We have also looked at gene context and operon structures and found many examples of fusion proteins, in which the HotDog domain has been fused to another protein to generate functional diversity. The large number of subfamilies we have found, the diverse range of activities these proteins participate in and the taxonomic distribution of the HotDog domain indicates an ancient superfamily that has diverged substantially to fulfil numerous roles in the cell. Our analysis may help with further experimental investigation of members of this superfamily. Some members of this superfamily, such as the *P. falciparium*FabZ enzyme have been proposed as a target for new anti-malarial drugs \[[@B64]\] as FabZ homologues are not found in humans. Finally our analysis identified hundreds of novel proteins such as human mesenchymal stem cell protein DSCD75 and the *Ralstonia solanacearum*hypothetical protein RSp0367 as probable enzymes potentially involved in lipid metabolism. Given that the large majority of proteins in this family are involved in bacterial lipid metabolism we suggest that the HotDog domain evolved in bacteria first and may then have been transferred to eukaryotes and archaea on several occasions. Since this time duplication and mutation has allowed it to fill a variety of roles. Methods ======= Sequence analysis ----------------- All PSI-BLAST searches were carried out using default inclusion thresholds and searched against the Swiss-Prot and TrEMBL sequence database (SWISS-PROT release 42.12 and TrEMBL release 25.12). To define subfamilies we clustered the results of an all-against-all search of the 1357 HotDog domain proteins using NCBI BLASTP and single linkage clustering at an E-value of 10^-15^. Operon analysis --------------- Gene context/operon analysis was carried out with the GeConT tool ([Ge]{.underline}ne [Con]{.underline}text [T]{.underline}ool) \[[@B65]\] available at the GeConT Home Page \[[@B66]\]. Domain analysis --------------- Protein domain analysis was carried out using Pfam \[[@B13]\] (release 12.0) available at the Pfam Home Page \[[@B67]\]. Motif analysis -------------- Consensus motif sequences were identified in the subfamily alignments using the MASIA program \[[@B62]\] available at the MASIA 2.0 Home Page \[[@B68]\]. Authors\' contributions ======================= AB conceived the study and was involved in all stages of the manuscript preparation including figure preparation. SCD conducted the PSI-BLAST searches, prepared the manuscript and some of the figures. Both authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 **A network showing the unification of the HotDog superfamily using PSI-BLAST searches.**Each yellow oval represents a query sequence used to seed a PSI-blast search. We compared each PSI-BLAST output to all the others and connected them with a line if they shared any sequences in common. There are only two connections in the graph that were not made (Image is in EPS format). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 **A complete list of the 1357 HotDog domain containing proteins.** ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 3 **The HotDog domain superfamily.**This xml-like file contains 1293 (95%) of the HotDog domain containing sequences, grouped into 85 clusters, which permits investigators to immediately identify HotDog domain(s) in their \'unknown\' protein of interest and allow them to infer some functionality. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 4 **A HotDog domain HMM library.**This library can be used in conjunction with the HMMER program to search for HotDog domains in any protein sequence. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 5 **A list of motifs identified in each subfamily.**Motifs were identified using the MASIA program \[62\]. A motif starts when at least 3 of 4 consecutive positions are more than 40% conserved and extend until at least 2 amino acids in a row are less than 40% conserved \[[@B71]\]. Motifs corresponding to PROSITE motif PS01328, \[QR\]-\[IV\]-x(4)-\[TC\]-D-x(2)-G \[IV\]-V-x-\[HF\]-x(2)-\[FY\] are underlined and in bold. Motifs highlighted in red and green are conserved between the respective subfamilies. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 6 **Subfamily alignments.**These alignments were constructed using the MAFFT alignment program \[72\] and rendered using the CHROMA software package \[73\]. Known active site residues are indicated below the subfamily alignments. The highly conserved Asp residue in the PaaI subfamily is proposed as an active site residue based on motif similarities between the 4HBT-II subfamily and the PaaI subfamily. Jpred predicted consensus secondary structures are indicated above the alignments \[74\]. ::: ::: {.caption} ###### Click here for file ::: Acknowledgments =============== We would like to acknowledge Robert Finn\'s help in generating Figure [1](#F1){ref-type="fig"} and Alexey Murzin and Chris Ponting for helpful discussions. We would also like to thank David J. Studholme for critical reading of the manuscript. Finally we would like to thank the referee\'s for their helpful comments. AB and SCD are funded by the Wellcome Trust. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **The structure of the active HotDog domain dimer.**(A) A ribbon representation of the *Escherichia coli*FabA dimer (PDB code: 1MKB), viewed along the dyad axis. Each 171-residue subunit contains a Hotdog fold/ domain, consisting of a seven-stranded antiparallel b-sheet \'bun\', coloured magenta and green, and a five-turn a-helical \'sausage\' coloured blue and purple in the respective subunits. The Hotdog fold is best observed in Figure B. There are two independent active sites located between the dimers, the active site residues of His^70^from one subunit and Asp^84^from the other subunit, represented as a *ball-and-stick model*with CPK colouring (carbon, black; hydrogen, white; oxygen, red; nitrogen, blue), constitute the potential reactive protein groups in the active sites \[1\]. (B) A view of FabA rotated 90° along the dyad axis. The figures were generated with MOLSCRIPT \[69\] and rendered with RASTER3D \[70\]. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **A schematic Figure showing the various domain organizations of proteins with a HotDog domain**. For each distinct architecture we show an example protein, the species it has come from and its length is shown in parentheses. The key identifies all the domains in the Figure and also includes the Pfam accession number or numbers describing each domain. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Rosetta fusion proteins in the HotDog domain superfamily.**For each fusion event we show an example operon containing the two proteins separate and an example of the fused rosetta protein. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Classifying the HotDog superfamily into subfamilies. ::: **Number of Members** **Phyletic distribution** **Subfamily Name** **Representative member(s)** **Accession Number** **E.C. number** **Associated Domain(s)** **General Function** ----------------------- --------------------------------- ------------------------------------------ ---------------------------------------------------------------------------------------------------- ---------------------- ----------------- ------------------------------------------------------------------------------------------------------ -------------------------------------------------------- 135 Proteobacteria : 67 Acyl-CoA thioesterases Cytosolic long-chain acyl-CoA thioester hydrolase/ Brain acyl-CoA hydrolase Q64559 3.1.2.2 2 × HotDog Fatty acid metabolism Metazoa: 25 Firmicutes: 16 Archaea: 9 Actinobacteria: 7 Cytoplasmic acetyl-CoA hydrolase (including brown fat inducible thioesterase) Q8WYK0 3.1.2.1 2 × HotDog, START Fatty acid metabolism Chlamydiae: 4 Viridiplantae: 2 Fungi: 2 Deinococcus-Thermus: 2 Acyl-CoA hydrolase Q81EE4 3.1.2.20 None Fatty acid metabolism Chlorobi: 2 Probable medium chain acyl-CoA hydrolase Q7NUH6 3.1.2.19 None Fatty acid metabolism 130 Proteobacteria: 61 FabZ-like Dehydratases (3R)-hydroxymyristoyl -- (acyl-carrier-protein) dehydratase/ β-hydroxy-acyl ACP dehydratase (FabZ) P94584 4.2.1.58 Lpx C in Q8KBX0 Fatty acid biosynthesis, lipid A biosynthesis (Q8KBX0) Firmicutes: 36 Cyanobacteria: 9 Chlamydiae: 4 Coronafacic acid (CFA) dehydratase P72238 4.2.1- None CFA biosynthesis Alveolata: 4 Viridiplantae: 4 Planctomycetes: 3 Bacteroidetes: 2 Fusobacteria: 2 Aquificae: 1 Chlorobi: 1 Deinococcus-Thermus: 1 Metazoa: 1 Thermotogae: 1 122 Proteobacteria: 81 MaoC dehydratase-like \(R) specific enoyl-CoA hydratase (phaJ) Q8KRE2 4.2.1.- None PHA biosynthesis Firmicutes: 13 PHA synthase (phaC) O32472 Archaea: 11 MaoC protein P77455 None PHA biosynthesis Actinobacteria: 8 17-beta-hydroxysteroid dehydrogenase P51659 Aldehyde Dehydrog. PHA biosynthesis Deinococcus-Thermus: 2 type 4 ADH short, SCP Hormone biosynthesis Fusobacteria: 2 Spirochaetes: 2 Acidobacteria: 1 Planctomycetes: 1 Viridiplantae: 1 32 Proteobacteria: 23 NodN-like Nodulation protein N P25200 None Nodule formation Actinobacteria: 5 Bacteroidetes: 2 Deinococcus-Thermus: 1 Spirochaetes: 1 102 Proteobacteria: 70 YbgC-like YbgC protein P44679 None Cell envelope maintenance? Firmicutes: 14 Cyanobacteria: 6 Archaea: 3 Actinobacteria: 2 Planctomycetes: 2 Spirochaetes: 2 Aquificae: 1 Deinococcus-Thermus: 1 Metazoa: 1 77 Proteobacteria: 36 FabA-like dehydratases/ synthases β-hydroxydecanoyl ACP dehydratase (FabA) P18391 4.2.1.60 None Unsaturated fatty acid biosynthesis Actinobacteria: 26 Omega-3 polyunsaturated fatty acid synthase (pfaC) Q93CG6 2 × HotDog, various numbers of BKAS N-term, BKAS-C-term & Acyl-transf. in Q93CG6, Q8YWH0, and Q9S1Z9 Polyunsaturated fatty acid biosynthesis Fungi: 13 Fatty acid synthase Q48926 Acyl-transf., BKAS-N & BKAS-C Fatty acid biosynthesis Cyanobacteria: 1 Stramenopiles: 1 73 Viridiplantae: 57 Fat subfamily FatA acyl-ACP-thioesterase Q43718 3.1.2.14 None Fatty acid synthesis Firmicutes: 15 FatB acyl-ACP-thioesterase Q41634 3.1.2.14 None Fatty acid synthesis Bacteroidetes: 1 66 Proteobacteria: 31 TesB-like *E. coli*acyl-CoA thioesterase II (tesB) P23911 3.1.2.- 2 × HotDog, cNMP in Q8GYW7 Fatty acid metabolism Actinobacteria: 14 Metazoa: 14 Human thioesterase II/ Peroxisomal acyl-CoA thioesterase O15261 3.1.2.2 2 × HotDog Fatty acid metabolism. Role in HIV infection? Viridiplantae: 4 Fungi: 3 59 Proteobacteria: 30 4HBT-II 4-hydroxybenzoyl-CoA thioesterase from *Arthrobacter*sp. strain SU and TM1 Q04416 HAD domain in Q89YN2 4-chlorobenzoate degradation Firmicutes: 12 Actinobacteria: 11 ComA2 protein P14205 None Unknown Bacteroidetes: 2 Viridiplantae: 2 Chlorobi: 1 Deinococcus-Thermus: 1 21 Firmicutes: 21 CBS-associated Hypothetical protein bh3175 Q9K832 DRTGG, 2 × CBS Unknown 19 Proteobacteria: 10 PaaI Phenylacetic acid degradation protein I P76084 None Phenylacetic acid metabolism Archaea: 5 Actinobacteria: 2 Bacteroidetes: 1 Firmicutes: 1 14 Proteobacteria: 8 Hydroxyacyl-CoA dehydrogenase-associated *A. tumefaciens*C58 3-hydroxyacyl-CoA dehydrogenase Q8UJY0 3HCDH\_N, 3HCDH in Q8UJY0 and Q92NF5 Fatty acid metabolism Firmicutes: 5 Actinobacteria: 1 13 Proteobacteria: 10 Acetyltransferase *E. coli*YiiD protein P32148 acetyltransf Putative acetyltransferase Chlorobi: 1 Cyanobacteria: 1 Metazoa: 1 11 Firmicutes: 11 FapR Transcription factor FapR Q9KA00 HTH Transcriptional regulation of fatty acid metabolism 11 Metazoa: 10 Proteobacteria: 1 MSCP Mesenchymal stem cell protein DSCD75 Q9NYI2 None Unknown 10 Proteobacteria: 10 YbaW Hypothetical protein ybaW P77712 None Unknown 9 Proteobacteria: 9 AMP-binding subfamily Hypothetical protein RSp0367 Q8XSV0 2 × AMP-bind Unknown 5 Proteobacteria: 4 Firmicutes: 1 4HBT-I 4-hydroxybenzoyl-CoA thioesterase from *Pseudomonas*sp. strain CBS-3 and DJ-12 P56653 3.1.2.23 None 4-chlorobenzoate degradation Gentisyl-CoA thioesterase Q9KBC9 None Degradation of aromatic compounds? :::

PubMed Central

2024-06-05T03:55:47.387225

2004-8-12

PMC516017

Background ========== Allelic loss of heterozygosity (LOH) of the human chromosome 16q in several sporadic cancer types, including breast, prostate and ovary cancers, suggests this chromosome arm harbors tumor suppressive loci \[[@B1]-[@B3]\]. The most frequent region of allelic loss occurs within 3-megabases (Mb) at 16q24.3 between the marker D16S498 and the telomere \[[@B1],[@B4]\]. Sequencing of the 3-Mb region has identified approximately 100 genes \[[@B4]\]. Eight of these have been excluded as potential tumor suppressors for breast cancer based on mutation analysis in tumor DNA \[[@B5]\]. Recently, a messenger RNA (mRNA) expression survey was completed within 2.4-Mb of this region examining the expression profiles of over 75 genes in a panel of breast tumor cell lines \[[@B6]\]. It was found that only three genes exhibited significantly aberrant expression profiles. These genes were highly expressed in some cell lines and lowly expressed in others. This led to the hypothesis that this aberrant expression may reflect a role for these genes as tumor suppressors in determining cancer phenotype and behavior. One of these genes was the core-binding factor, alpha subunit 2, translocation to 3; termed CBFA2T3. CBFA2T3 encodes for a protein that belongs to the eight-twenty-one (ETO) family, which also includes the genes CBFA2T1, CBFA2T2 in mammalian cells and nervy in Drosophila \[[@B7]\]. The mammalian members of this protein family are involved in therapy-related chromosomal translocations causing acute myeloid leukemia \[[@B8]\]. The CBFA2T3 gene encodes two alternative transcripts, CBFA2T3A (NM\_005187) and CBFA2T3B (NM\_175931) (Figure [1A](#F1){ref-type="fig"}). CBFA2T3A reportedly functions as a nuclear transcriptional co-repressor via its interaction with histone deacetylase (HDAC) complexes \[[@B9]\]. CBFA2T3B functions as a kinase anchorage protein in T lymphocytes and may play a role in inflammatory response \[[@B10]\]. Recently, it was demonstrated that CBFA2T3B functions as a transcriptional repressor and exhibits in vitro characteristics consistent with tumor suppressor activity \[[@B11]\]. This gene was found to be lowly expressed in a number of breast tumor cell lines and upon re-introduction it reduced their growth on plastic and in soft agar. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **CBFA2T3 gene structure and promoter methylation patterns.**(A) CBFA2T3 encodes two alternative transcripts, CBFA2T3A and CBFA2T3B. CBFA2T3A is encoded by exons 1A and 2--12. CBFA2T3B is encoded by exons 1B-12 splicing out exon 3. Relative exon sizes are shown. The exon 1A start site contains no CpG island. The black box marks the location of a high-density CpG island located five prime to the exon 1B start site. The black arrowheads mark the primers used for real-time RT-PCR. (B) CBFA2T3B contains a CpG island of approximately 160 CpG sites spanning 1-kb of sequence. The single black bars represent CpG sites scaled relative to each other. CAT ELISA promoter constructs and primers used for MSP, real-time MSP and bisulfite sequencing are shown. The asterisk marks the location of the amplicon and internal primers used for second-round real-time MSP. (C) CBFA2T3B promoter methylation patterns were examined in hypermethylated cell lines using sodium bisulfite sequencing. A characteristic sinusoidal pattern of approximately six high to low frequency methylation levels every 40--150 bp was detected. The high-frequency cytosine methylation levels residing within a consensus Sp1 binding site located approximately minus 450 bp from the transcriptional start of exon 1B are shown. :::  ::: Mutation analysis in breast tumor cell lines and primary breast tumor samples failed to identify any CBFA2T3 sequence aberrations \[[@B11]\]. It was recognized that aberrant promoter methylation might be the mechanism responsible for the altered expression of CBFA2T3 in breast tumors. The expression of several tumor suppressor genes has been found inactivated or reduced in tumors in association with promoter hypermethylation \[[@B12]\]. Promoter hypermethylation can occur in conjunction with allelic loss and or mutation and is regarded as an alternative form of \'knockout\' in biallelic inactivation. Accumulating evidence now suggests that promoter hypermethylation may affect genes that reside within regions of frequent allelic loss more often than mutation \[[@B13]\]. Alternatively, several oncogenes have been found to be over-expressed in tumors in association with promoter hypomethylation \[[@B14]\]. As such, aberrant promoter methylation is considered a fundamental process in developing cancers and has recently received considerable interest as a rapid non-invasive molecular screening tool for the early detection of tumor cells in a range of bodily fluids and biopsy specimens \[[@B15]\]. In this study, the methylation status of a high-density CpG island promoter region located five prime to the exon 1B sequence of the CBFA2T3B transcript is described (Figure [1A](#F1){ref-type="fig"}). We explored whether this region is aberrantly methylated in breast tumor cell lines and primary breast tumor samples and whether a correlation exists between methylation and gene expression. Both aberrant hypo and hypermethylation levels were detected in breast tumors in correlation with elevated and reduced expression. The phenomenon of hypo and hypermethylation relates to the amount of DNA used in the methylation methodology as discussed. Results ======= Isoform-specific analysis of CBFA2T3 gene expression levels ----------------------------------------------------------- CBFA2T3 encodes two alternative transcripts, CBFA2T3A and CBFA2T3B (Figure [1A](#F1){ref-type="fig"}). It was recently demonstrated that CBFA2T3 expression levels are aberrant in breast tumor cell lines. In this study, the total expression levels were assayed using real-time RT-PCR and primers that span exons 4--5 of the CBFA2T3 gene. The tumor suppressor activity previously shown for CBFA2T3B, led us to examine the expression profile of this transcript in breast tumor cell lines using a TaqMan probe specific for exon 1B. CBFA2T3B displayed aberrant expression similar to the total (Figure [2A](#F2){ref-type="fig"}). This aberrant expression was found to be endogenously low in all samples examined. The expression of this gene was so low that it could not be reliably detected using Northern Blots or RNase protection (data not shown). In addition, expression levels of the CBFA2T3A transcript were assayed in breast tumor cell lines using a TaqMan probe specific for exon 1A. CBFA2T3A expressed at lower levels than CBFA2T3B, but due to the identification of several complex splice variants between exon 1A and exons 1B, 2, 3 and 4, further analysis is required (data not shown). An example of the raw data real-time RT-PCR expression analysis of the CBFA2T3 transcripts is shown ([Additional file 1](#S1){ref-type="supplementary-material"}). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **CBFA2T3 gene expression levels, 5-Aza-dC re-expression and promoter activity.**(A) CBFA2T3 gene expression levels were assayed using real-time RT-PCR. Several breast tumor cell line and normal tissue sample expression levels are shown. The y-axis represents mRNA mlcls expressed per 10^4^cells shown on a log scale (mean ± SD, n = 6). Fold changes in expression relative to normal breast expression are shown above each sample. The white diamonds and white squares represent expression levels of the housekeeping genes cyclophilin A (CYPA) and ATPase coupling factor 6 subunit (ATP5A), respectively. CBFA2T3B and CBFA2T3 expressed at endogenously low yet aberrant levels in breast tumor cell lines. Using normal breast as a reference, CBFA2T3B (600 mRNA mlcls per 10^4^cells) and CBFA2T3 (1,800) were low compared to ATP5A (600,000) and CYPA (1,600,000). CBFA2T3 expression ranged 30,000-fold from 4 to 120,000 mRNA mlcls per 10^4^cells in MDA-MB-231 and BT-483, respectively. In contrast, CYPA and ATP5A expression ranged 2-fold and 20-fold, respectively (100--200 and 15--300 mRNA mlcls per cell). Expression levels were also examined in several primary breast tumor samples for which total RNA was available (see Figure 6). (B) CBFA2T3 re-expression levels were examined in MDA-MB-231 cells using 5-Aza-dC. Fold changes in CBFA2T3B, CBFA2T3 and SYK expression levels are shown upon exposure to 5-Aza-dC relative to control untreated cells. \> 100-fold re-expression was detected in CBFA2T3 and CBFA2T3B. 5-Aza-dC had no affect on CBFA2T3A expression (data not shown). \> 1,000-fold re-expression was also detected in SYK, a control gene known to be hypermethylated and down-regulated in MDA-MB-231 cells. (C) CBFA2T3B promoter activity was assayed using CAT ELISA. Promoter constructs labeled A to D (see Figure 1B) were inserted upstream of the CAT reporter in pBLCAT3 (Boehringer Mannheim). pBLCAT2 driven by the tyrosine kinase promoter was used as a positive control. 2.0 × 10^6^293T cells were transfected in triplicate using Lipofectamine 2000 (Invitrogen) with 1.5 μg of construct and control vector and 0.3 μg of the internal pSVβ-galactosidase control vector (Stratagene). Cells were lysed after 24 h and CAT concentrations determined using ELISA. Of the four constructs labeled A to D, the 1-kb B construct promoted a 30-fold increase in CAT expression (mean ± SD are triplicates, n = 3). :::  ::: Qualitative analysis of CBFA2T3B promoter methylation levels ------------------------------------------------------------ Based on the idea that this altered expression might develop via aberrant methylation, the promoter activity and methylation status of a high-density CpG island located five prime to the exon 1B sequence of CBFA2T3B was examined (Figures [1A](#F1){ref-type="fig"} and [1B](#F1){ref-type="fig"}). The promoter activity was assayed using chloramphenicol acetyl-transferase (CAT) ELISA and confirmed that a 1-kb region spanning the island was capable of promoting a 30-fold increase in CAT expression (Figure [2C](#F2){ref-type="fig"}). To determine if this promoter region is aberrantly methylated in breast tumors, 24 breast tumor cell lines, 20 primary breast tumors, 20 normal breast counterparts and 24 normal whole blood samples were screened using methylation-specific PCR (MSP). Four separate 100--200 bp regions spanning the promoter were amplified using primers specific for either unmethylated or methylated cytosines at the CpG sites shown (Figure [1B](#F1){ref-type="fig"}). The full results of this analysis are summarized in [Additional file 2](#S2){ref-type="supplementary-material"}. In general, a low \'basal\' level of methylation was detected in all samples at the various regions examined. An example of this basal methylation is shown for the normal blood samples at region two (Figure [3A](#F3){ref-type="fig"}). Unlike the bloods, several breast tumor cell lines, primary breast tumors and normal breast counterpart samples displayed complex high to low methylation levels. An example of this complex methylation is shown for the primary breast tumors and their normal counterparts at region four (Figure [3B](#F3){ref-type="fig"}). This analysis revealed that only few cell lines displayed clear hypermethylation (e.g. MDA-MB-231) or hypomethylation (e.g. BT-483) in association with reduced and elevated expression. To support this association, it was found that treatment of MDA-MB-231 cells with the demethylating agent 5-aza-2\'-deoxycytidine (5-Aza-dC) was capable of increasing both CBFA2T3 and CBFA2T3B expression levels by \> 100-fold relative to controls (Figure [2B](#F2){ref-type="fig"}). Overall, however, it was difficult to comprehend this complex promoter methylation and an association with expression thus suggesting that further analysis was required. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **CBFA2T3B promoter methylation levels examined using MSP.**The full results of this analysis in breast tumor cell lines, primary breast tumors, normal breast counterparts and normal whole blood samples are summarized in Additional file 2. (A) An example of the characteristic basal methylation levels detected in all samples is shown for normal blood samples examined at region 2. (B) An example of the complex high to low methylation levels is shown for 20 primary breast tumor samples with adjacent normal breast counterparts. The pUC19 DNA/MspI marker concentrations reflect an approximate concentration of 100 to 1 unmethylated to methylated mlcls. The asterisks indicate the samples examined by bisulfite sequencing. :::  ::: Bisulfite sequence analysis of CBFA2T3B promoter methylation patterns --------------------------------------------------------------------- To understand this complex promoter methylation, sodium bisulfite sequencing of the hypermethylated breast tumor samples was used to examine the pattern and frequency of methylation with this region. This analysis revealed that specific cytosines appeared more susceptible to methylation compared to others (Figure [4](#F4){ref-type="fig"}). A characteristic sinusoidal pattern of approximately six high to low frequency methylation levels every 40--150 bp was detected in cell lines and tumor samples (Figures [1C](#F1){ref-type="fig"} and [4](#F4){ref-type="fig"}). Based on this pattern, primers were designed specific for the real-time MSP quantitation of high-frequency cytosine methylation levels residing within a consensus Specificity protein (Sp1) binding site located approximately minus 450 bp from the transcriptional start of exon 1B (Figures [1B](#F1){ref-type="fig"} and [1C](#F1){ref-type="fig"}). As Sp1 proteins are commonly known to regulate gene transcription, it was considered that variable methylation at this site may be reflective of elevated or reduced expression and thus suitable for resolving and correlating the complex promoter methylation levels. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **CBFA2T3B promoter methylation patterns examined using bisulfite sequencing.**The methylation patterns in breast tumor cell lines, primary breast tumors, normal breast counterparts and normal whole blood samples are shown. The y-axis represents protected 5-methylcytosines scored as percent cytosines methylated per 5--10 clones. The complete methylation maps displaying 160 CpG sites spanning 1-kb of sequence are shown for breast tumor cell lines only. The white bars indicate the Sp1 sites targeted by second-round real-time MSP. Bisulfite sequencing of the CBFA2T3B promoter region in hypermethylated breast tumor samples MDA-MB-231, MDA-MB-468 and 14T revealed a characteristic sinusoidal methylation pattern. This sinusoidal pattern was also detected in samples SK-BR-3, 14N and 17T at levels approximately 1 tenth of the hypermethylated samples. No methylation was detected in the normal blood samples 4B and 5B or the breast tumor samples 3T and MCF-7. :::  ::: Quantitative analysis of CBFA2T3B promoter methylation levels ------------------------------------------------------------- To assay CBFA2T3B promoter methylation levels at this Sp1 site, a bisulfite sequencing amplicon spanning this region was initially PCR amplified and column purified from 24 breast tumor cell lines, 55 primary breast tumors, 22 normal breast counterparts and 46 normal whole blood samples. Second-round real-time MSP was performed on the amplicons using internal forward primers to detect for either unmethylated or methylated cytosines at the Sp1 site. Standard curve dilutions of previously prepared internally primed clones representative of either unmethylated or methylated sequence were used to extrapolate methylation levels and normalize for differences in amplification efficiencies. The methylated cytosines were expressed as a fractional ratio of unmethylated cytosines to determine the methylation indices \[mi = m/(m + u)\]. On plotting the indices, a clear difference between the tumor, normal groups and complex promoter methylation levels was revealed (Figure [5](#F5){ref-type="fig"}). The normal blood samples maintained a specific basal methylation level and were similar to normal breast counterparts with methylation indices ranging from 0.006--0.09 and 0.002--0.08, respectively. Median methylation index levels were 0.02 in normal bloods and 0.01 in normal breast counterparts. In contrast, several breast tumor cell lines and primary breast tumors were highly variable relative to the normal samples with methylation indices ranging from 0.0002--0.8 and 0.0001--0.9, respectively. A Levene\'s test revealed a statistically significance difference in variances of methylation indices with the tumor groups being more varied than the normals (P = .001). It was predicted that 83--75% of breast tumor cell lines and 78--69% of primary breast tumors displayed aberrant methylation levels outside the methylation index interquartile ranges of the normal blood (0.02--0.04) and normal breast counterpart samples (0.008--0.03), respectively. Half of the aberrations detected in breast tumor cell lines were either hypo or hypermethylated relative to the normal breast counterpart interquartile ranges. Up to 22% of the primary breast tumors were hypomethylated and 47% hypermethylated relative to the normal breast counterpart interquartile ranges. An example of the raw data real-time MSP and melt curve analysis showing the aberrant methylation levels in breast tumor cell lines compared to normal whole blood samples is shown ([Additional files 3](#S3){ref-type="supplementary-material"} and [4](#S4){ref-type="supplementary-material"}). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **CBFA2T3B promoter methylation levels assayed using second-round real-time MSP.**The methylation levels in normal whole blood samples (nwb), normal breast counterparts (nbc), primary breast tumors (pbt) and breast tumor cell lines (btcl) were assayed at the Sp1 site shown in Figure 1C using real-time MSP. The y-axis represents methylation levels plotted as methylation indices \[mi = m/(m + u)\] on a log scale. Each white circle represents a different sample. The breast tumor cell lines examined are shown in descending order from high to low methylation. The horizontal bars mark the median methylation indices calculated for each group. The asterisks mark the interquartile ranges for normal groups. The median methylation indices and (interquartile ranges) were 0.02 (0.02--0.04) for nwb, 0.01 (0.008--0.03) for nbc, 0.03 (0.009--0.08) for pbt and 0.02 (0.002--0.3) for btcl. The median methylation index variance in each tumor group was statistically significantly different than the normal groups (P = .001); nwb/btcl (P \< .0001), nwb/pbt (P = .009), nbc/btcl (P = .01), nbc/pbt (P = .05). The normal group median methylation index variances were not significantly different; nwb/nbc (P = 0.6). :::  ::: Correlation of CBFA2T3B promoter methylation and gene expression levels ----------------------------------------------------------------------- To correlate CBFA2T3B promoter methylation levels with gene expression, the methylation indices from 24 breast tumor cell lines and 20 primary breast tumor samples were plotted against their expression. A statistically significant inverse correlation between CBFA2T3B promoter methylation and exon 1B specific expression was established (r^2^= 0.63; r = -0.8, P = .0002). Based on the possibility that five prime RNA degradation and secondary structures may have affected the exon 1B complementary DNA (cDNA) synthesis, a correlation between methylation and the total expression is alternatively shown (Figure [6](#F6){ref-type="fig"}). CBFA2T3B promoter hypermethylation and reduced expression inversely correlated with hypomethylation and elevated expression (r^2^= 0.72; r = -0.9, P \< .0001). At a hypermethylated index of around 0.9, approximately 4 mRNA molecules (mlcls) per 10^4^cells were detected compared to a hypomethylated index of 0.0001 and 120,000 mRNA mlcls per 10^4^cells. The number of CBFA2T3B promoter mlcls methylated per cell for each breast tumor cell line was also calculated by multiplying the methylation index values by the number of 16q24.3 DNA mlcls per cell as previously determined by FISH \[[@B16]\]. These values were plotted against expression in aim to improve the original correlation (r^2^= 0.77; r = -0.9, P \< .0001) ([Additional file 5](#S5){ref-type="supplementary-material"}). By plotting these data sets a power regression was derived which could be used to solve unknown x or y values. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **CBFA2T3B promoter methylation levels versus gene expression.**The data for 24 breast tumor cell lines and 20 primary breast tumor samples are shown. The y-axis represents methylation levels assayed using real-time MSP and plotted as methylation indices \[mi = m/(m + u)\] on a log scale. The x-axis represents total expression levels assayed using real-time RT-PCR and plotted as mRNA mlcls per 10^4^cells on a log scale. Each white circle represents a different primary breast tumor sample and the black circles represent the breast tumor cell lines. The breast tumor cell lines examined are shown in descending order from high to low methylation. A statistically significant inverse correlation was established between promoter hypermethylation (mi = 0.9) and reduced expression (4 mRNA mlcls per 10^4^cells) versus hypomethylation (0.0001) and elevated expression (120,000) (r^2^= 0.72; r = -0.9, P \< .0001). A power regression (y = cx^b^) describes the relationship between methylation and expression. :::  ::: Discussion ========== In this study, it has been demonstrated that expression of the CBFA2T3B isoform is altered in breast tumors and that this correlates strongly with aberrant CpG island promoter methylation. Moreover, a comprehensive method for the detection, quantitation and correlation of promoter methylation and gene expression levels has been developed. MSP was used in combination with sodium bisulfite sequencing to identify sites within the CBFA2T3B promoter region displaying high-frequency methylation in breast tumors. Second-round real-time MSP was used to quantitate methylation levels at these sites in breast tumor cell lines, primary breast tumors, normal breast counterparts and normal whole blood samples. The CBFA2T3B promoter methylation levels were calculated as methylation indices and the indices from breast tumors were plotted against their expression. Validated methylation detection using the MSP and real-time MSP methods ----------------------------------------------------------------------- Throughout the development of this method, it was recognized that a pre-requisite for valid MSP amplification required that sufficient amounts of bisulfite modified DNA were used in order to detect methylation and avoid stochastic effects when quantitating low target mlcl numbers and parameters relating to particle distribution statistics. The amount of DNA modified and subsequently amplified is an important parameter in terms of actually detecting potential CpG methylation in a given specimen as discussed. Most MSP studies use 50 ng of bisulfite modified DNA for amplification. As a normal diploid cell contains an average 6.6 ρg of DNA, this equates that approximately 15,000 alleles or mlcls are made available as starting templates for MSP. If for example, 1 mlcl in 1,000 of these is methylated (mi = 0.001) for a given gene at a specific CpG site(s), then approximately 15 mlcls will act as potential starting templates. Routinely, in unmodified DNA specimens a \'purified\' 100--300 bp amplicon diluted to this mlcl number should be detected at around 28--30 cycles of real-time PCR under standard primer efficiencies. Initially, this is a late cycle threshold (C~T~) for the detection of amplification and is prone to stochastic effects. Under MSP conditions, the DNA has been bisulfite modified which introduces numerous variables that further reduce the overall probability of detection. Upon modification multiple sequence permutations at the CpG site(s) of interest may arise (e.g. from 3 CpG sites a total 8 possible C to TpG permutations exist as 2^3^= 8). If primers are designed to detect a permutation present at say 4 mlcls in the above example of 15 mlcls (i.e. at attogram starting amounts), then under MSP, a 100--300 bp amplicon at this mlcl number will not amplify until after 35 cycles or may not amplify at all. Moreover, because the initial cycles of MSP are asymmetric the starting template mlcl numbers are further reduced. In addition, it is well known that the DNA is substantially degraded following bisulfite modification with studies demonstrating up to 84--96% degradation \[[@B17]\]. If 1 μg of DNA is modified, as routinely reported, and say 80% is degraded, then based on an unlikely recovery rate of 100%, only 20% of the original pool of templates used in the 50 ng will be available for amplification. Thus, in the above example if only 20% of the 15 mlcls and its possible permutations are available then they may not amplify at all. The pool of primer specific methylated templates could in fact be instantly diminished. When considering the levels of bisulfite-mediated degradation in 1 μg of DNA, even alleles methylated at levels as high as 1 in 10 might be difficult to detect or become spurious. For instance, if assuming that a 100 times more template is available (i.e. 1,500 mlcls) yet permutations exist (i.e. 400 mlcls) and the recover rate is only 4%, then as little as 16 mlcls will be available as starting templates for amplification. Thus, several technical thresholds exist for the detection of methylation using MSP. The use of 1 μg of DNA for modification, 50 ng for amplification, sequence permutations, PCR efficiency and bisulfite-mediated DNA degradation all reduce the overall probability and validity of detection. To warrant potential detection of methylation in a given specimen at specific CpG sites above a threshold of say 1 in 1,000, it was found necessary to use at least 10X coverage of the 1 μg amount or 10 μg of DNA for modification and 400--500 ng for MSP amplification. Taking into account the above example, this should provide that at least 30--40 mlcls are made available as starting templates, which should amplify within 28 cycles. Specimens with methylation levels lower than this, such as 1 in 50,000, will not be reliably detected and require the use of second-round nested MSP amplification. In this case, the first-round will still require sufficient amounts as any initial stochastic effects may result in poor reproducibility for the second-round. Moreover, as demonstrated in this study, bisulfite sequence analysis of the CBFA2T3B promoter methylation patterns in the hypermethylated MSP samples indicated that the potential for variable methylation frequencies do exist. In this case, it was found that even with sufficient amounts of DNA the qualitative MSP was unreliable and presented complex methylation data. The use of MSP primers for sites that are not methylated or methylated at low to high frequencies, in combination with these other technical thresholds, created complex stochastic methylation data merely decipherable by examining several regions. As a result, this concealed the true methylation status in a majority of the samples under investigation. Only those samples with clear hypermethylation (e.g. MDA-MB-231) or hypomethylation (e.g. BT-483) were greatly reproducible and associated with reduced and elevated expression (Figure [2A](#F2){ref-type="fig"} and [Additional file 2](#S2){ref-type="supplementary-material"}). To overcome these technical thresholds, a second-round real-time MSP assay was developed. A sequencing amplicon displaying high-frequency methylation sites was initially PCR amplified under 10X coverage conditions and column purified from all samples. Second-round real-time MSP was performed on approximately 100 ρg of the amplicon using internal forward primers to detect for either unmethylated or methylated cytosines at the high-frequency sites. Using this approach, methylation was detectable in every sample examined. All samples amplified with 28 cycles and a methylation index was calculated for these. In comparison, when the standard MSP conditions were assayed by real-time using SYBR Green I, it was found that the C~T~of methylation detection was below 28 cycles for most samples (data not shown). Extremely low methylation levels, such as those in BT-483, were undetectable using this method but were detected using the second-round. In the latter case, up to 1 methylated mlcl in 10,000 (mi = 0.0001) could be reliably detected within 25 cycles of amplification (Figure [5](#F5){ref-type="fig"} and [Additional file 3](#S3){ref-type="supplementary-material"}). Notably, no wild-type, unmethylated or methylated cross-amplification was detected when using more DNA. If this occurs then there may be a problem with primer design or modification as the conversion should be complete. In fact, with more DNA a bias in amplification for the target sequence should be created. In addition, several other technicalities were overcome. The use of only a single second-round primer to quantitate the methylation levels of the high-frequency sites limited the number of possible 5-methyl-CpG sequence permutations that could perturb the accuracy of quantitation. This coupled with the quantitation of methylation levels from absolute standard curves enabled the normalization of reaction efficiencies and calculation of absolute methylation indices \[mi = m/(m + u)\] and absolute methylation ratios (u/m). The methylation ratios are simply an alternative way of expressing the methylation data ([Additional file 6](#S6){ref-type="supplementary-material"}). The calculation of these single methylation values based on the quantitative normalization discriminates against biased amplifications and comparisons of unmethylated to methylated \'band\' intensities when using MSP. Moreover, the absolute quantitation of indices and ratios offers improvement over the relative methods, such as comparative cycle thresholds (ΔΔC~T~), as they have biological significance, are less consuming, more accurate, and do not require the dynamic range in amplification efficiencies of target and references to be similar to enable valid quantitation. The concept of gene promoter hypo and hypermethylation ------------------------------------------------------ Thus, by using this method it was possible to detect CBFA2T3B promoter methylation in all samples. This phenomenon may actually be widespread for genes under the control of methylation as in accordance with the replication model of maintenance methylation \[[@B18]-[@B20]\]. This model might suggest that the CpG island is not just randomly methylated; the CpG island is \'always\' methylated by memory yet propagated at variable levels for cell type-specific expression or at aberrant levels in association with cancers. Remarkably, in this study it was found that the methylated CBFA2T3B CpG island is propagated as a sinusoidal pattern at aberrant levels in both permanent cell lines and recently resected tumor specimens suggesting that a methylation memory does exist. In fact, it was recognized that the bisulfite sequencing detection levels of this sinusoidal pattern complemented the second real-time MSP quantitations. For example, a methylation ratio in SK-BR-3 of 10:1 (mi = 0.1) was concordant with a 1 in 10 or 10% 5-methylcytosine per 10 clone frequency (Figure [4](#F4){ref-type="fig"} and [Additional file 6](#S6){ref-type="supplementary-material"}). If by comparing the other bisulfite sequencing levels with their methylation ratios this might suggest that up to 30, 60, 90 or 5,000 clones of 4B, 5B, 3T and MCF-7, respectively, would require sequencing in order to detect 1 methylated sinusoidal mlcl. Based on the idea that promoter methylation may occur ubiquitously, the concept arises; what is hypo and hypermethylation. In this study, it was demonstrated that hypo and hypermethylation are merely a prediction of levels outside a majority or \'interquartile range\' of methylation levels found in the normal samples. For instance, of all the methylation levels detected in the cell lines, only 37.5% of these were predicted as either hypo or hypermethylated relative to the interquartile ranges of normal breast methylation. Moreover, the number of predicted hypo and hypermethylated samples are much lower when compared to the \'full range\' of normal methylation ratios as shown in [Additional file 6](#S6){ref-type="supplementary-material"}. This is considerably lower for the primary breast tumor samples (i.e. only 24% hypermethylated relative to normal breast) and is likely due to the heterogeneity of breast tumors in concealing the true tumor-related methylation levels. Regardless, by using a more sensitive detection method, the phenomenon of hypo and hypermethylation has appeared. The phenomenon hypomethylation is a reflection of the low methylation levels that can be detected in the \'hypomethylated\' samples relative to the methylated and hypermethylated samples. Alternatively, the phenomenon of hypermethylation is readily observed yet indicates that the detection of methylation alone does not simply represent hypermethylation. The CBFA2T3B gene is methylated in all samples and according to MSP in a high-percentage of samples. In effect, several studies which have examined the methylation status of potential tumor suppressors and cancer-related genes often demonstrate methylation in a high-percentage of samples and state hypermethylation based on the detection of methylation. Examples include, ras-effector nore1A (RASSF1A) \[[@B21],[@B22]\], stratifin (14-3-3σ) \[[@B23]\], p15 and p16 \[[@B24]\], O6-methylguanine-DNA methyltransferase (MGMT) \[[@B25]\], mismatch repair gene (hMLH1) \[[@B26]\] and hyperplastic colon polyps gene (HPP1) \[[@B27]\]. In these cases, as with CBFA2T3B, it is probable that only a small percentage of this methylation has cancerous significance in terms of \'methylation-induced silencing\', which is routinely confirmed by correlation with \'absence\' of expression. Notably, several studies use insufficient amounts of RNA and cDNA for expression analysis and state the absence of amplification to represent inactivation rather than a level of reduction, particularly in cases of low endogenous expression. The need for quantitation to classify methylation levels has been recognized for genes such as glutathione S-transferase P1 (GSTP1) \[[@B28]\] and adenomatous polyposis coli (APC) \[[@B29]\], although in these cases the use of low DNA may nevertheless affect intra and interassay reproducibility. In biological terms, several scenarios exist as to how these aberrant methylation levels might develop. In the case of hypomethylation, a correlation could be made with severely duplicated chromosome copies of 16q24.3 (e.g. BT-483, MCF-7). A possible scenario here is that the hypomethylation is apparent because of a duplicated copy number and thus, for example, not a direct cause of reduced DNA methyltransferase activity or over-expression per se. In the case of hypermethylation, a correlation could be made with 16q24.3 LOH (e.g. MDA-MB-231, MDA-MB-468). An emerging scenario here is that the hypomethylation induces 16q LOH to promote aberrant DNA methyltransferase activity and hypermethylation \[[@B13]\]. Accumulating evidence suggests that the hypermethylation itself, or \'aberrant\' methylation, may be targeted to constantly methylated CpG islands (i.e. methylation induces methylation), and in addition targeted to transcriptionally inert CpG islands \[[@B19]\]. In this study, it was found that the CBFA2T3B promoter region is in fact constantly methylated at a median methylation index of 0.02 (i.e. 2 mlcls in 100 are methylated). When comparing this median methylation index with a median gene expression index from the breast tumor cell lines, it was calculated that only approximately 20 mRNA mlcls per 98 \'active\' alleles are transcribed. These calculations are shown in [Additional file 5](#S5){ref-type="supplementary-material"}. This suggests that the CBFA2T3B gene is not only constantly methylated but also largely transcriptionally inert with the remaining active alleles possibly trans-factor dependent for expression. Although the nature of such targeted aberrations are unknown, several studies demonstrating that altered maintenance and or de novo DNA methyltransferase activities can induce tumorigenesis, clearly demonstrates that the controlled methylation of CpG islands is crucial for normal cell development \[[@B30]-[@B34]\]. Accordingly, an in silico prediction of several Sp1, homeotic, epidermal and insulin growth factor recognition sites within the CBFA2T3B promoter region may implicate a role for this element in epithelial development. How the CBFA2T3B CpG island is maintained and dispersed at specified levels within a population of cells is unknown but likely relates to the methylation machinery in controlling distribution within a cell type-specific population (i.e. methylation memory and or phenotype). In this study, it was shown that aberrant deviations outside these specified levels occur profoundly in a majority of breast tumors, particularly the pure tumorigenic cell lines, and as such might comprise an element in tumor formation. Conclusions =========== Overall, this study lends further support to the idea that CBFA2T3B is aberrantly regulated in breast cancer. Additional clues supporting a role for this gene in tumor suppression may reside within its protein structure. CBFA2T3B contains a characteristic zinc finger myeloid-nervy-DEAF-1 (zf-MYND) domain and nervy homology regions \[[@B7]\]. Several studies demonstrate that these regions function in transcriptional co-repression via interaction with HDAC and or nuclear co-repressor complexes \[[@B9],[@B35]-[@B38]\]. Accumulating evidence suggests the zf-MYND domain, which is also common to developmental proteins RP-8, DEAF-1, suppressin, Blu, BS69, PDCD2 and Bop, may interact with co-repression complexes to regulate cell-cycle transcription during cell type-specific differentiation \[[@B39]-[@B46]\]. Abnormal regulation may be central to tumor formation as supported by reports that several zf-MYND-like proteins display tumor suppressive activity. Recently, much emphasis has been placed on the development of methylation-based tumor biomarkers for early breast cancer detection to predict disease outcome and strategies for therapy \[[@B15]\]. Interesting data indicates that the use of nipple aspirate fluids may provide for a rapid non-invasive source of screening material for methylation biomarker analysis \[[@B47]\]. Similar studies are underway to evaluate if the methylation status of CBFA2T3B presents biomarker utility. Although in this case, because the normal breast fluids and breast tumors themselves are histologically complex tissues containing a variety of cell types, it is recognized that further studies are required to refine the methylation index interquartile ranges. This may involve large-scale comparisons between normal and tumor cells captured using laser microdissection. Ideally, such normal controls would be resected at autopsy or reduction mammoplasty from non-risk category individuals. Methods ======= Sample collection and nucleic acid isolation -------------------------------------------- 24 breast tumor cell lines were obtained from the American Type Culture Collection and cultured under recommended conditions. 46 normal whole blood samples, 55 primary breast tumor samples with pathologically classified grade III lesions and 22 adjacent normal breast counterparts samples were obtained with clinical research approval from the Flinders Medical Centre, Department of Surgery. Breast tumor cell line, primary breast tumor and normal breast counterpart genomic DNA was isolated using GenElute for Mammalian Tissues (Sigma). Whole blood genomic DNA was isolated using the QIAamp DNA Blood Kit (Qiagen). Breast tumor cell line and primary breast tumor total RNA was isolated using RNAqueous-4PCR (Ambion). Nucleic acid concentrations were determined using RiboGreen (Molecular Probes). Breast tumor cell line, primary breast tumor and normal tissue (Clontech) total RNA extracts were DNase I treated (Ambion). Real-time reverse transcription-PCR (RT-PCR) -------------------------------------------- 10--20 μg of total RNA was oligo(dT)~16~reverse transcribed at 55°C for 2 h using MMLV (Promega) with addition of RNAguard (Promega) and 5% DMSO. CBFA2T3, CYPA, ATP5A and SYK expression levels were assayed by real-time RT-PCR using SYBR Green I (BMA). CBFA2T3 isoform expression levels were assayed using TaqMan probes specific for exons 1A and 1B (GeneWorks). Real-time RT-PCR was performed on a Rotor-Gene 2000 (Corbett Research) using standard 25 μl HotStar Taq conditions (Qiagen) on cDNA equivalent to 100--1,500 ng total RNA. 0.35X final SYBR Green I or 200 nM probe was used for detection. Amplifications were at 95°C for 10 min, 45 cycles at 94°C for 20 s, annealing temperature for 30 s and 72°C for 30 s. Primer sequences and annealing temperatures are shown ([Additional file 7](#S7){ref-type="supplementary-material"}). Unknown expression levels were extrapolated from standard curve dilutions of column purified cDNA amplicons. Replicate standard curve assays (n ≥ 2) were used with C~T~coefficient variations averaging \< 15% over six orders within replicates and between dilutions. mRNA mlcl numbers were quantitated from samples at the parameter C~T~from standard curves with known mlcls/μl calculated from the dilution mass in μg/μl ÷ M.W. of ssRNA transcript (× 6.02 × 10^17^mlcls/μmole). mRNA mlcls per cell were calculated at the C~T~concentration ÷ amount of total RNA (ρg) per reaction multiplied by 4 based on a 4 ρg total RNA per cell estimation. mRNA was shown as raw expression data (n \> 4) and or normalized against CYPA or ATP5A. Several other housekeeping and cancer-related gene expression levels were quantitated to ensure the mRNA mlcl per cell estimations were compatible with other methods (data not shown). Sodium bisulfite modification ----------------------------- 10--20 μg of genomic DNA was digested overnight at 37°C with restriction enzymes XbaI, XhoI, HindIII and EcoRI and cleaned using nucleotide purification columns (QIAvac 24, Qiagen). The digested DNA was pooled and bisulfite modified \[[@B17]\]. 10 μg of DNA was diluted in 500 μl of water and denatured with 55 μl of 2 M NaOH for 20 min at 37°C. DNA was mixed with 300 μl of 10 mM hydroquinone (Sigma), 5.2 ml of 3.6 M NaHSO~3~(pH 5.0) (Sigma), overlaid with paraffin oil and deaminated in the dark for 16 h at 55°C. DNA was desalted using Qiagen purification columns, eluted in 500 μl water and desulfonated with 55 μl of 3 M NaOH for 20 min at 37°C. DNA was neutralized and precipitated with 800 μl of 10 M ammonium acetate (pH 7.0), 20 μl linear acrylamide and 5 ml cold 100% EtOH. Modified DNA was pelleted, resuspended in 40 μl 1 mM Tris-Cl (pH 8.0) and the concentrate stored at -80°C. Methylation-specific PCR (MSP) and sodium bisulfite sequencing -------------------------------------------------------------- Four 100--200 bp regions spanning approximately 1-kb of the CBFA2T3B promoter region were amplified from ≥ 300 ng of bisulfite modified DNA under standard HotStar Taq conditions using primers specific for either unmethylated or methylated cytosines. Hypermethylated breast tumor samples were used for bisulfite sequencing to generate a CBFA2T3B promoter region methylation map. Four 300--400 bp regions spanning the promoter were amplified under standard conditions as described above using primers simultaneous for both unmethylated and methylated cytosines. The amplicons were sub-cloned into pGEM (Promega) and 5 to 10 clones sequenced using BigDye (Applied Biosystems). Primer sequences and annealing temperatures are shown ([Additional file 7](#S7){ref-type="supplementary-material"}). Demethylation assay ------------------- MDA-MB-231 was treated with the demethylating agent 5-Aza-dC (Sigma). Approximately 2.0 × 10^5^cells per T75 flask were seeded in 10 ml RPMI-1640 supplemented with 10% FCS, 15 mM HEPES, 10 mg/liter PGS and cultured for 48 h at 37°C with 5% CO~2~. Cells were treated with 50 μm 5-Aza-dC for 120 h and replenished with fresh medium and 5-Aza-dC every 24 h. These concentrations are not inhibitory to cell growth \[[@B48]\]. Concentrations ranging from 1--5 μM 5-Aza-dC had no affect on expression levels (data not shown). Control untreated cells were cultured in parallel and supplemented with DMSO. Replicate T75 flasks for both treatments and controls were performed (n = 4). Total RNA was isolated and analyzed for CBFA2T3 isoform and SYK expression levels using real-time RT-PCR. MDA-MB-231 cells were also treated with the HDAC inhibiting agents trichostatin A (Sigma), sodium butyrate (Sigma) and apicidin (Calbiochem). Only trichostatin A elicited re-expression levels similar to 5-Aza-dC (data not shown). Real-time methylation-specific PCR (MSP) ---------------------------------------- A sequencing amplicon was initially tested for bisulfite PCR amplification and cloning bias by scoring percent methylation frequencies of overlapping amplicons. Bisulfite PCR bias was also tested by amplification on proportional mixtures of hypo and hypermethylated bisulfite treated DNA. This amplicon was PCR amplified and column purified from all samples. Second-round real-time MSP was performed on the amplicon using internal forward primers to detect for either unmethylated or methylated cytosines at the Sp1 site displaying high-frequency cytosine methylation. The methylated and unmethylated primer specificities were tested by real-time amplification on serial dilutions of unmethylated and methylated clones, respectively. Second-round amplicons were also sequenced to ensure specificity. Second real-time MSP was performed on a Rotor-Gene 2000 using internal forward primers under standard 25 μl HotStar Taq conditions with approximately 100 ρg of first-round cleaned amplicon and 0.35X SYBR Green I. Amplifications were at 95°C for 10 min, 45 cycles at 94°C for 20 s, annealing temperature for 30 s and 72°C for 30 s. Unknown unmethylated and methylated cytosine levels were extrapolated at the Sp1 site from standard curve dilutions of internally PCR amplified and column purified cloned amplicons originally sequenced and found to be representative of either unmethylated (e.g. BT-483) or methylated (e.g. MDA-MB-231) sequence. Unmethylated and methylated mlcl numbers were quantitated from all samples at the parameter C~T~from standard curves with known mlcls/μl calculated from the dilution mass in μg/μl ÷ M.W. of dsDNA clone sequence (× 6.02 × 10^17^mlcls/μmole). CBFA2T3B promoter methylation levels were expressed as methylation indices \[mi = m/(m + u)\] and ratios (u/m) \[[@B49],[@B50]\]. Statistical analysis -------------------- CBFA2T3B promoter methylation index medians and interquartile ranges were determined for each group of tissue samples. Statistical comparisons between groups were performed using the two-way Levene\'s test for Equality of Variances, i.e. H0: median methylation index variances are the same in each group; HA: median methylation index variances are not the same in each group. Variances in the median methylation indices between each group were considered statistically significant when P ≤ .05. Correlations between CBFA2T3B promoter methylation and gene expression levels were determined by calculating a Spearman\'s rank coefficient. All statistical analyses were performed using GraphPad Prism Version 4.0. Authors\' contributions ======================= AJB completed the work and manuscript. AEG aided in bisulfite sequencing. OLDM performed promoter analysis. DFC, GRS and GK supervised the work. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 **CBFA2T3 gene expression levels assayed using real-time RT-PCR**The raw data CBFA2T3B expression levels and preliminary CBFA2T3A expression levels in breast tumor cell lines are shown. The y-axis represents the fluorescence detection scale and the x-axis represents the C~T~of amplification. The CBFA2T3 gene expresses at endogenously low levels and requires at least 500 ng of reverse transcribed total RNA to cDNA template for reproducible detection. In contrast, the housekeeping genes such as CYPA require only 100 ng of template. When 500 ng is used, the CYPA expression levels are off the fluorescence scale. Note the C~T~values are below 35 cycles for the down-regulated cell lines such as MDA-MB-231. This low-level of expression is undetectable by conventional RT-PCR. Moreover, because of this low endogenous expression the CBFA2T3 mRNA could not be reliably detected using Northern Blots or RNase protection (pdf file). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 **CBFA2T3B promoter methylation levels examined using MSP**The methylation levels in 24 breast tumor cell lines (labeled), 20 primary breast tumors (1T-20T), 20 normal breast counterparts (1N-20N) and 24 normal whole blood samples (1--24) are shown. Four separate regions (1--4) spanning 1-kb of CBFA2T3B promoter sequence were amplified using primers to detect for either unmethylated or methylated cytosines (see Figure 1B for primer locations and Additional file 7 for primer sequences and annealing temperatures). The asterisks indicate the samples examined by bisulfite sequencing. Unmethylated (U) and methylated (M) band intensities were scored as high (black dot), low (gray dot) or negative (white dot). Low-level basal methylation defined by high unmethylated to low methylated band intensities was detected in all normal bloods in ≥ 2/4 regions. High methylation was also detected in 21% of bloods in region four. Samples were considered hypermethylated when high intensity bands amplified in all four regions or hypomethylated when no methylation was detected. Based on this, 21% of breast tumor cell lines were hypermethylated (e.g. MDA-MB-231 and MDA-MB-468), 16% hypomethylated (e.g. BT-483 and MDA-MB-361) and 63% displayed a combination of basal to high methylation. Primary breast tumor samples were complex. 57% displayed basal methylation in ≥ 2/4 regions, 14% displayed high methylation in ≥ 2/4 regions and 29% tested negative in ≥ 3/4 regions. Normal breast counterpart samples were similar with 62% displaying basal methylation in ≥ 2/4 regions, 14% displaying high methylation in ≥ 2/4 regions and 24% testing negative in ≥ 3/4 regions. The detection of high-methylation in MCF-7 is due to spurious amplification events. Several overlapping primers spanning this region in this cell line were methylation negative (data not shown). This cell line is hypomethylated according to real-time MSP (pdf file). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 3 **CBFA2T3B promoter methylation levels assayed using second-round real-time MSP**CBFA2T3B promoter methylation levels were assayed using second-round real-time MSP. The raw data methylation levels in normal whole blood samples and breast tumor cell lines are shown. The y-axis represents the fluorescence detection scale and the x-axis represents the C~T~of amplification. Second round real-time MSP was performed on a bisulfite sequencing amplicon using internal forward primers to detect for either unmethylated (uF) or methylated (mF) cytosines at the Sp1 CpG sites shown in Figure 1C. The C~T~for methylated amplification in breast tumor cell lines was highly aberrant compared to normal blood samples. Note the late unmethylated C~T~obtained in MDA-MB-231 compared the early hypermethylated C~T~. In contrast, BT-483 shows an early unmethylated and late hypomethylated C~T~(pdf file). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 4 **CBFA2T3B promoter methylation melt curve analysis**CBFA2T3B promoter methylation melt curves were examined following second-round real-time MSP. Raw data melt curves of second round amplicons in normal blood samples and breast tumor cell lines are shown. Curves were calculated from the negative derivative in fluorescence over temperature versus temperature (-dF/dT~m~versus T~m~). Normal blood samples displayed consistent peak levels for both the unmethylated and methylated mlcls. In contrast, breast tumor cell lines displayed highly aberrant peak levels depicted by broad melt transitions and heterogeneous melt curves reflective of the aberrant concentration and composition of 5-methylcytosines (pdf file). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 5 **CBFA2T3B promoter methylation levels versus gene expression**Methylation indices were calculated as methylated promoter mlcls per 10^4^cells for breast tumor cell lines with pre-determined 16q24.3 DNA mlcls per cell and plotted against CBFA2T3 mRNA mlcls per 10^4^cells. The y-axis represents methylation levels assayed using real-time MSP and the x-axis represents expression levels assayed using real-time RT-PCR. Both data sets are shown on a log scale. Each black circle represents a different breast tumor cell line. The asterisk marks the median methylation and median gene expression levels. The median methylation index of 0.02 (i.e. 2 mlcls or alleles methylated in 100) is calculated from the median methylation level of 450 methylated alleles per 10^4^cells divided by the number of unmethylated \'active\' alleles in the 10^4^cells or 20,000 alleles, i.e. \[450 ÷ (20,000 - 450) = 0.02\]. The median gene expression index of 0.2 (i.e. 20 mRNA mlcls expressed in 98 \'active\' alleles) is calculated from the median expression level of 4,500 mRNA mlcls per 10^4^cells divided by the number of unmethylated \'active\' mlcls in the 10^4^cells, i.e. (4,500 ÷ 19,550 = 0.2). This calculation equates to approximately 4--5 mRNA mlcls expressed per 10 cells and suggests that the CBFA2T3B gene is largely transcriptionally inert. The remaining active alleles may be trans-factor dependent for expression. An inverse correlation between promoter methylation and expression levels per population of cells was established (r^2^= 0.77; r = -0.9, P \< .0001). In hypermethylated MDA-MB-231, approximately 17,000 promoter mlcls were methylated (i.e. mi = 0.85 as 17,000 in 20,000 are methylated) and 4 mRNA mlcls expressed per 10^4^cells. In hypomethylated BT-483, approximately 5 promoter mlcls were methylated (mi = 0.0002) and 120,000 (± 40,000) mRNA mlcls expressed per 10^4^cells. This elevated expression equates that 12 (± 4) mRNA mlcls per cell are expressed from an estimated four-promoter mlcls per cell (i.e. four-16q24.3 DNA mlcls per cell) or that one \'active\' unmethylated CBFA2T3B promoter mlcl per cell transcribes 2--4 mRNA mlcls. From a methylation index of approximately 2% and greater, a large increase in y may lead to a small decrease in x. Under this condition, the relationship may be asymptotic. The regression equation y = 25,801 x^-0.58^(r^2^= 0.7669) describes the methylation and expression relationship. When plotted as methylation index values the equation was y = 1.05 x^-0.61^(r^2^= 0.7665) (pdf file). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 6 **CBFA2T3B promoter methylation levels assayed using second-round real-time MSP**The methylation levels in normal whole blood samples, normal breast counterparts, primary breast tumors and breast tumor cell lines were assayed at the Sp1 site shown in Figure 1C and plotted as absolute methylation ratios (u/m). All samples are labeled corresponding in part to Additional file 2. The asterisks indicate the samples examined by bisulfite sequencing. The gray highlights indicate the normal blood and normal breast counterpart \'full\' methylation ranges. Basal methylation ratios in normal bloods averaged 60:1 (cumulative mean) unmethylated to methylated CBFA2T3B promoter mlcls. This average coincided with conventional MSP band intensities of 100:1 based on pUC19 DNA/MspI marker concentrations (see Figure 3) and the median methylation index of 0.02 (i.e. 2 mlcls in 100 are methylated). Normal blood ratios ranged 20:1 to 160:1 unmethylated to methylated mlcls. Normal breast counterparts were similar to normal bloods averaging 100:1 but with a larger range of 10:1 to 350:1. Relative to the normal samples, breast tumors displayed highly aberrant methylation ratios clearly resolved by second-round real-time MSP. 75% of breast tumor cell lines were aberrantly methylated outside the full range of normal blood basal methylation. 58% were outside the range of normal breast. Half of the aberrations were either hypo or hypermethylated relative to both normal blood and normal breast. Similar to cell lines, 51% of primary breast tumors were aberrantly methylated relative to normal blood. 35% were aberrant relative to normal breast (i.e. 24% were hypermethylated and 11% hypomethylated). Aberrant methylation ratios ranged from 1:10 in hypermethylated MDA-MB-231 to 7,000:1 in hypomethylated BT-483 (pdf file). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 7 **Primers, probes and annealing temperatures used**(pdf file). ::: ::: {.caption} ###### Click here for file :::

PubMed Central

2024-06-05T03:55:47.393224

2004-8-10

PMC516018

Background ========== Intermediate filaments (IFs) with microtubules and actin microfilaments are the major cytoskeletal components of most vertebrate cells \[[@B1]-[@B4]\]. IF proteins constitute a large family of proteins that is divided into five types \[[@B1],[@B2]\]. The expression of the different IF proteins is differentiation and tissue specific \[[@B1],[@B5]\]. Keratins expressed in epithelial cells, represent the largest and most complex subtype of IF proteins (more than 20 proteins)\[[@B2]\]. They are classified into two groups, the type I (acidic K9 to K20) and the type II (neutral-basic, K1 to K8), which form obligate heteropolymers composed of equimolar amounts of type I and type II keratins \[[@B2],[@B6]\]. It is now generally accepted that, in multilayered epithelia, one of the function for keratins IFs is to protect the tissue from mechanical stress \[[@B7]-[@B9]\]. The first evidences for this function came from studies on epidermis, which showed that transgenic mice lacking epidermal keratins, or expressing mutated keratins, displayed blistering skin disease phenotypes, similar to human skin diseases such as epidermolysis bullosa simplex or epidermolytic hyperkeratosis \[[@B7],[@B10],[@B11]\]. As for epidermal keratins, the production of transgenic mice targeting K8 or K18 has been necessary to unravel the role of IFs in simple epithelium such as in the liver. In hepatocytes, K8/18 is the only keratin pair and thus both keratins are necessary to form an IF network. Transgenic mice expressing K8 or K18 carrying mutations that affect filament formation, develop mild hepatitis and display greater liver sensitivity to mechanical and toxic stress than wild type animals \[[@B12],[@B13]\]. Recent studies from Ku et al. \[[@B14]-[@B16]\] have shown that mutations on K8/18 predispose to the development of liver disease in humans. Moreover, modifications in IF organization and the formation of keratin containing aggregates, named Mallory bodies (MBs), are observed in different liver diseases such as alcoholic hepatitis, Wilson\'s disease, Indian childhood cirrhosis and liver steatosis in obesity \[[@B17]-[@B21]\]. Other proteins, such as ubiquitin and the heat shock protein 70 kDa (HSP70), are also present in MBs and could play a role in their formation \[[@B22]-[@B24]\]. Taken together, these results support the hypothesis that keratins are necessary to preserve the hepatocytes integrity upon stressful conditions. It is still unclear how keratins accomplish these protective roles. Previous studies have shown that modifications in keratin phosphorylation are associated with various conditions such as mitosis, apoptosis and stress, suggesting a role for this post-translational modification in the modulation of keratin-related functions \[[@B25]-[@B27]\]. Long-term treatment of mice with a diet containing griseofulvin (GF) induces the development of an hepatitis associated with the formation of MBs, which are biochemically and morphologically similar to those found in humans \[[@B19],[@B28]\]. This animal model constitutes a useful tool to investigate the keratin dynamics in the response of hepatocytes to the presence of a hepatotoxic agent. In the present study, we investigated the effect of chronic GF intoxication on hepatocytes from C3H and FVB/n mouse strains. We monitored the expression of the stress inducible form of the heat shock protein 70 kDa (HSP70i) and the induction of K8/18 phosphorylation at specific sites: K8 on serine 79 (K8 S79), K8 on serine 436 (K8 S436), K18 on serine 52 (K18 S52) and K18 on serine 33 (K18 S33) (reviewed in \[[@B26],[@B29]\]). We also examined the possible relationship between HSP70i expression and K8/18 phosphorylation, during the development of hepatitis and in MB formation. Results ======= Induction of HSP70i and K8/18 phosphorylation upon GF-treatment in C3H and FVB/n mouse livers --------------------------------------------------------------------------------------------- The modifications in the amount of HSP70i, K8/18, and phosphorylated keratins (K8 pS79, K8 pS436 and K18 pS33) were analyzed by Western Blotting of total proteins from control and GF-treated C3H and FVB/n mouse livers (2 weeks, 6 weeks and 5 months). GF intoxication induced an increase in keratin levels in livers from both mouse strains (Fig. [1A,1B](#F1){ref-type="fig"}). HSP70i, which was present in control livers of both mouse strains, was also increased by the treatment (Fig. [1A](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Biochemical analysis of livers from C3H and FVB/n mice.**Western blots from C3H mouse livers; A, K8 and HSP70i; C, K8 pS79; E, K18 pS33; G, K8 pS436. Western blots from FVB/n mouse livers; B, K8; D, K8 pS79; F, K18 pS33. :::  ::: Total proteins from C3H and FVB/n livers were probed with antibodies against K8 pS79, K8 pS436 and K18 pS33 (Fig. [1C,1D,1E,1F,1G](#F1){ref-type="fig"}). Significant changes in K8 and K18 phosphorylation occurred after GF-treatment in both mouse strains (Fig. [1C,1D,1E,1F,1G](#F1){ref-type="fig"}). Small amounts of K8 pS79 and K18 pS33 were found in control livers (Fig. [1C,1D](#F1){ref-type="fig"} and [1E,1F](#F1){ref-type="fig"}), whereas K8 pS436 was not detected (Fig. [1G](#F1){ref-type="fig"}). After 2 weeks of treatment, an increase in the amount of all phosphokeratin species studied was observed. The phosphorylation levels of K8 S436 and K18 S33 remained higher than control values in both mouse strains for the entire treatment (Fig. [1E,1F,1G](#F1){ref-type="fig"}). However, when compared with 2 week treatment, a decrease in K8 pS436 and K18 pS33 was noted after 5 months of treatment (Fig. [1E,1F,1G](#F1){ref-type="fig"}). Similarly, a decrease in K8 S79 phosphorylation was observed after 5 months of treatment, in C3H mice (Fig. [1C](#F1){ref-type="fig"}). However in FVB/n mice, K8 pS79 was not detected after the same period of treatment (Fig. [1D](#F1){ref-type="fig"}). Localization of HSP70i and K8/18 during GF intoxication ------------------------------------------------------- We analyzed at the cellular level, by double immunofluorescence staining, the distribution of HSP70i and IFs on liver sections of control and GF-treated C3H and FVB/n mouse livers. In control hepatocytes, IFs formed a complex cytoplasmic network that was denser at the cell membrane and particularly around the bile canaliculi (Fig. [2A](#F2){ref-type="fig"}). Our biochemical analysis showed that HSP70i was present in control hepatocytes. However, by immunofluorescence, we did not detect the presence of HSP70i in the cells (Fig. [2B](#F2){ref-type="fig"}). After 2 weeks of treatment, most of the hepatocytes were enlarged and the bile canaliculi were dilated. IF network was denser around dilated bile canaliculi (Fig. [2C](#F2){ref-type="fig"}). All hepatocytes contained a very dense cytoplasmic IF network. These modifications were accompanied by an increase in the amount of HSP70i in hepatocytes and a granular staining was detectable at the membrane and in the nuclei (Fig. [2D](#F2){ref-type="fig"}). A few cells showed a high level of HSP70i. After 5 months of treatment, there was a mosaic pattern of cells with and without IF staining (Fig. [2E](#F2){ref-type="fig"}). HSP70i showed a granular staining pattern in many hepatocytes and was also present in MBs (Fig. [2F](#F2){ref-type="fig"}). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Distribution of keratin IFs and HSP70i in hepatocytes from control and GF-fed C3H mice.**A, C, E keratin IFs; B, D, F HSP70i; A, B) control; C, D) 2 week treatment; E, F) 5 month treatment. Arrows in E and F indicate reactive MBs with Troma 1 (anti-K8) and anti-HSP70i, respectively. Scale bar = 20 μm. :::  ::: Phosphorylation of K8 S79, K8 S436 and K18 S33 during GF intoxication --------------------------------------------------------------------- Cryosections of control and GF-treated C3H and FVB/n mouse livers were fixed with 4% paraformaldehyde and processed for double immunofluorescence staining. As mentioned above, control mice showed hepatocytes with a cytoplasmic IF network, which was denser at the cell periphery (Fig. [3](#F3){ref-type="fig"},[4](#F4){ref-type="fig"},[5A](#F5){ref-type="fig"}). K8 pS79 and K8 pS436 were not generally detected in the IF network of control hepatocytes. Only, occasionally some doublet cells, most likely representing cells in mitosis, were stained (data not shown). A basal level of phosphorylation for K18 S33 was detected at the periphery of all hepatocytes (Fig. [5B](#F5){ref-type="fig"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Distribution of keratin IFs and K8 pS79 in hepatocytes from control and GF-fed C3H mice.**A, C, F, I keratin IFs; B, D, E, G, H, J K8 pS79; A, B) control; C, D, E) 2 week treatment; F, G, H 6 week treatment; I, J 5 month treatment. Arrow in D indicates clusters of cells containing K8 pS79. Empty arrowheads in I and J indicate MBs reactive with Troma 1 but not with LJ4 (anti-K8 pS79), respectively. Scale bar = 20 μm. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Distribution of keratin IFs and K8 pS436 in hepatocytes from control and GF-fed C3H mice.**A, C, E, G keratin IFs; B, D, F, H K8 pS436; A, B) control; C, D) 2 week treatment; E, F) 6 week treatment; G, H); 5 month treatment. Arrows in D indicate clusters of cells containing K8 pS436. Filled arrowheads in G and H indicate reactive MBs with Troma 1 and 5B3 (anti-K8 pS436), respectively. Scale bar = 20 μm. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Distribution of keratin IFs and K18 pS33 in hepatocytes from control and GF-fed C3H mice.**A, C, E, G keratin IFs; B, D, F, H K18 pS33; A, B) control; C, D) 2 week treatment; E, F) 6 week treatment; G, H) 5 month treatment. Asterisk in D shows an hepatocyte containing a high level of K18 pS33; arrow indicates a dilated bile canaliculi. Filled arrowheads in G and H indicate reactive MBs with Troma 1 and Ab8250 (anti-K18 pS33), respectively. Scale bar = 20 μm. :::  ::: After 2 weeks of GF-treatment, hepatocytes were enlarged and an increase in cytoplasmic IF network was observed (Fig. [3](#F3){ref-type="fig"},[4](#F4){ref-type="fig"},[5C](#F5){ref-type="fig"}). This treatment induced the phosphorylation of K8 on S79 and S436 in some hepatocytes. K8 pS79 and K8 pS436 were present in clusters of cells scattered over the whole liver (Fig. [3](#F3){ref-type="fig"},[4](#F4){ref-type="fig"},[7D](#F7){ref-type="fig"}). The groups of cells stained with the anti-K8 pS79 or anti-K8 pS436 usually surrounded damaged cells (Fig. [3](#F3){ref-type="fig"},[7D](#F7){ref-type="fig"}). In the case of K8 pS79, IFs located in the cytoplasm and at the periphery of the cells were highly stained (Fig. [3D,3E](#F3){ref-type="fig"}). For K8 pS436, the staining was stronger at the cell periphery and around the dilated bile canaliculi (Fig. [4D](#F4){ref-type="fig"}). In addition to their presence in clusters of cells, K8 pS79 and K8 pS436 displayed an intense cytoplasmic staining in some isolated cells or cell doublets (Fig. [3E](#F3){ref-type="fig"}). Since both epitopes showed similarities in their patterns of distribution, we asked whether they were present in the same hepatocytes. Immunostaining for the detection of K8 pS79 and K8 pS436 were performed on serial liver sections of GF-treated mouse liver. Our results showed that the groups of hepatocytes positives for K8 pS79 were also positive for K8 pS436 (Fig. [6](#F6){ref-type="fig"}). ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Distribution of keratin IFs and K8 pS79 in hepatocytes from control and GF-fed C3H mice.**A, C, E, G keratin IFs; B, D, F, H K8 pS79; A, B) control; C, D) 2 week treatment; E, F) 6 week treatment; G, H) 5 month treatment. Arrow in D indicates clusters of cells containing K8 pS79; asterisk shows a damaged hepatocyte. Filled arrowheads in G and H indicate MBs reactive with Troma 1 and LJ4 (anti-K8 pS79), respectively; empty arrowheads indicate MBs reactive with Troma 1 but not with LJ4 (anti-K8 pS79), respectively. Scale bar = 20 μm. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Colocalization of K8 pS79 and K8 pS436 in hepatocytes from GF-fed C3H dmice.**A, C K8 pS79; B, D K8 pS436; A, B, C, D 2 week treatment. Arrows in A and B indicate clusters of hepatocytes containing both K8 pS79 and K8 pS436. Note: reactivity in nuclei observed in A, B, C and D represents non-specific staining due to the secondary antibody. Scale bar = 20 μm. :::  ::: In the case of K18 S33, its phosphorylation was increased in most (if not all) hepatocytes. Most of the staining was observed at the periphery of the cells, delimitating clearly the bile canaliculi. A few hepatocytes showed high levels of cytoplasmic K18 pS33 (Fig. [5D](#F5){ref-type="fig"}). After 6 weeks of GF-treatment, the distribution of hepatocytes containing K8 pS79 and K8 pS436 was different from the one observed after 2 weeks of treatment. Clusters of labeled cells were smaller, whereas labeled isolated cells became more prominent (Fig. [3G](#F3){ref-type="fig"},[4F](#F4){ref-type="fig"}). Singlet and doublet cell(s) highly labeled with K8 pS79 and K8 pS436 were also present (Fig. [3H](#F3){ref-type="fig"}). K18 pS33 was present in most hepatocytes and showed a similar pattern as the one observed after staining with Troma1 (Fig. [5F](#F5){ref-type="fig"}). After 5 months of GF-treatment, MBs were present in some hepatocytes in both mouse strains (Fig. [3I](#F3){ref-type="fig"},[4G](#F4){ref-type="fig"},[5G](#F5){ref-type="fig"}). MBs had variable size and different positions depending on the cell and were observed in cells with or without a visible intracytoplasmic IF network, as detected with Troma 1. In both mouse strains, K8 pS436 and K18 pS33 were present in MBs (Fig. [4](#F4){ref-type="fig"},[5H](#F5){ref-type="fig"}), whereas K8 pS79 seemed to be absent (Fig. [3J](#F3){ref-type="fig"}). Experiments described above were also performed using cold acetone instead of 4% paraformaldehyde. After acetone fixation, no difference in the staining pattern was observed for GF-treatment of 2 and 6 weeks in both mouse strains (Fig. [7](#F7){ref-type="fig"}). However, differences were observed for MB staining in the 5 months GF-treated mouse liver. In C3H mouse strain livers, K8 pS79 was present in many MBs although some of them showed no staining (Fig. [7H](#F7){ref-type="fig"}). In FVB/n mouse strain livers, K8 pS79 was not present in most MBs (Fig. [8](#F8){ref-type="fig"}). No difference in staining of MBs was observed for K8 pS436 and K18 pS33. ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **Distribution of keratin IFs and K8 pS79 in hepatocytes from GF-fed FVB/n mice.**A, C keratin IFs; B, D K8 pS79; A, B, C, D 5 month treatment. Asterisks in A and B indicate MBs reactive with Troma 1 but not with LJ4; arrows in C and D indicate MBs reactive with Troma 1 and LJ4, respectively. Scale bar = 20 μm. :::  ::: Localization of phosphorylated K8 species and HSP70i during GF intoxication --------------------------------------------------------------------------- Double immunostaining with anti-HSP70i and anti-phosphorylated keratins (K8 pS79 or K8 pS436) was performed for studying the localization of HSP70i in relation to keratin phosphorylation. The results showed that HSP70i and phosphorylated K8 species colocalized in some cells (Fig. [9](#F9){ref-type="fig"}). However, in most of the cells, the colocalization was not observed. ::: {#F9 .fig} Figure 9 ::: {.caption} ###### **Distribution of phosphorylated keratin IFs and HSP70i in hepatocytes from GF-fed C3H mice.**A K8 pS79; B, D HSP70i; C K8 pS436; A, B, C, D 2 week treatment. Arrows in A and B indicate cells in which HSP70i and K8 pS79 colocalized. Arrows in C and D indicate cells in which HSP70i and K8 pS436 colocalized. Scale bar = 20 μm. :::  ::: Discussion ========== The functional significance of K8/18 in simple epithelium has been the subject of numerous studies over the last decade \[[@B29]-[@B35]\]. Although most of these reports lead towards roles for K8/18 in the resistance of cells to mechanical and toxic stress, the molecular mechanisms underlying these phenomena remain to be elucidated. To date, most of our understanding of the pathways involving keratins in the response of hepatocytes to toxic stress comes from the analyses of various cell lines \[[@B36]-[@B38]\]. K8/18 phosphorylation at specific sites has been proposed to be a key factor in the regulation of those keratin functions. In this regard, K8 pS79, K8 pS436, K18 pS52 and K18 pS33 are the most studied phosphorylation sites \[[@B39]\]. *In vivo*, K8/18 are also subjected to phosphorylation and, as suggested by *in vitro*studies, it is proposed to help hepatocytes to cope with toxic stress \[[@B26],[@B29],[@B35]\]. For instance, transgenic mice expressing human K18 S52 mutated in alanine mutant are more susceptible to drug-induced liver injury than transgenic mice over expressing wild type human K18 \[[@B40]\]. In the present study, we showed that the chronic intoxication of mice with GF, which is known to induce modifications in keratin organization and formation of MBs \[[@B33]\], was associated with increased expression of the stress protein HSP70i. GF-treatment resulted in a rapid increase in the expression of HSP70i. This modification was already perceptible after 2 weeks of treatment and was maintained for the whole period of treatment. This result provides direct evidence that GF-treatment, which has been proposed to constitute an oxidative stress for hepatocytes \[[@B41]\], triggers signaling pathways involved in cellular protection \[[@B33]\]. This interpretation of our biochemical data is in agreement with our immunofluorescence study, which showed that HSP70i partly relocalized to the nucleus during the treatment. This distribution pattern is typical of the distribution of HSP70i in stressed cells \[[@B42],[@B43]\]. We have previously shown that GF intoxication induced an overall increase in K8/18 phosphorylation \[[@B44],[@B45]\]. Here, we show that GF-treatment is associated with modifications in K8/K18 phosphorylation at specific sites such as: K8 S79, K8 S436 and K18 S33. Among the studied phosphorylation sites, and because it was present and was increased in all treated hepatocytes, K18 pS33 was the keratin phosphoepitope the tissue distribution of which resembled that of HSP70i. The phosphorylation of K18 S33 has been shown to play a role in keratin reorganization during mitosis and by linking 14-3-3 proteins, to modulate their function \[[@B46],[@B47]\]. Hence, we propose that K18 S33 phosphorylation could be linked to IF reorganization during GF intoxication. Moreover, because K18 pS33 is increased in all hepatocytes, it could be implicated in the stress response by participating in the relocalization and/or the recruitment of molecules or factors implicated in stress-induced cell signaling. In contrast to K18 pS33, phosphorylated K8 species, K8 pS79 and K8 pS436 were not present in control mice hepatocytes. After 2 and 6 weeks of treatment, we observed an increase in the level of phosphorylation on these sites. However, contrary to HSP70i and K18 pS33, these phosphorylation sites were only present in isolated cells (singlet or doublets) or clusters of cells. Labeled singlet or doublet cells were more numerous after staining with the anti-K8 pS79 than after staining with the anti-K8 pS436. These cells could correspond to cells that are undergoing mitosis. This is supported by previous studies which showed that the phosphorylation of K8 on S79 and S436 occurs during mitosis \[[@B25],[@B48]\]. This interpretation is also in agreement with the work of Stumptner et al. \[[@B49]\], which showed the presence of cell doublets reactive with the anti-K8 pS79 after a short treatment with DDC that induces on the long term MB formation. The discrepancy in the number of cells stained for K8 pS79 and K8 pS436, both in the singlet and doublet cells, suggests that different kinases are involved in the phosphorylation of those sites. The presence of K8 pS79 and K8 pS436 was also detected in islets of cells. Interestingly, both antigens were present in the same clusters of cells surrounding unstained cells that were most likely undergoing apoptosis. These unstained cells are evocative of detached cells during anoikis, an apoptotic process that can be induced by loss of cell-cell anchorage. Stress and apoptosis has been shown to modulate K8 S79 and K8 S436 phosphorylation \[[@B25],[@B48]\]. The observed phosphorylation could indicate that these hepatocytes are stressed hepatocytes intended to apoptosis. However, analysis of the livers for the presence of apoptosis showed that only a few hepatocytes are going through programmed cell death and groups of cells in apoptosis were never observed (data not shown). We propose that the apoptotic cell could represent the starting point of a signal transduction pathway to neighboring cells. The activation of specific kinases that would phosphorylate keratins could provide those cells a resistance to apoptosis. This latter interpretation is in agreement with the notion that K8/18 intermediate filaments play a key role in the protection of cells against apoptosis \[[@B26],[@B35]\]. Liao et al. \[[@B50]\] have shown that HSP70 associates with K8/18 *via*K8. Our study show that colocalization of HSP70i and IFs occurs only in a few hepatocytes. Since the hepatocytes, in which colocalization was observed, contained K8 pS79 or K8 pS436, HSP70i binding to IFs in these cells may be related to the presence of keratin phosphorylation and participates to cellular pathways involving phosphorylated K8/18 on specific sites. Ku et al. \[[@B51]\] have shown that phosphorylation could modulate K8/18 ubiquitination and ensuing turnover. Knowing that binding of HSP70 to a protein can affect its targeting by kinases or phosphatases \[[@B52]\], HSP70i could bind to phosphorylated K8 species, prevent dephosphorylation by specific phosphatases, and thereby enhance phosphorylation-mediated K8/18 protection from degradation by the ubiquitin pathway \[[@B51],[@B53]\]. However, since HSP70i and phosphorylated K8 species colocalized only in a few cells over the whole tissue, the relevance of this phenomenon in the response to the presence of the hepatotoxin needs to be addressed and further investigations will be necessary to confirm that hypothesis. Chronic intoxication of mice with GF induces the formation of MBs. Numerous studies have demonstrated the presence of different phosphorylated K8/18 species within MBs, suggesting that K8/18 phosphorylation could participate in the MB formation processes \[[@B49],[@B54]\]. In our experiments, we showed that K8 pS436 and K18 pS33 were present in all observed MBs, whereas K8 pS79 was present in MBs in C3H mice hepatocytes but not in FVB/n mice. The difference in the presence of K8 pS79 phosphoepitope within MBs suggests that phosphorylation at that specific site is not essential for MB formation. However, as suggested by Stumptner et al. \[[@B49]\], because K8 pS436 and K18 pS33 are always detected in MBs, phosphorylation on these sites could be implicated in the processes of MB formation. Taken together, those results indicate that in the context of MB formation, K8/18 phosphorylation should not be considered as a general phenomenon but as specific events that affect precise sites on K8 or K18. The difference observed between keratin phosphorylation in C3H and FVB/n mice indicates that the genetic background influences the response of hepatocytes to toxic stress. This interpretation is in agreement with the results obtained with K8-null mice which displayed variable phenotypes depending on the genetic background \[[@B30],[@B31]\]. The treatment with GF that represents a toxic stress, most likely, involves the activation of stress activated protein kinases (SAPKs) in some hepatocytes. SAPKs p38 and JNK are physiologic kinases for K8 S79 and K8 S436 \[[@B37],[@B55]\]. We postulate that p38 kinase and/or JNK are activated by GF-treatment in some hepatocytes and are responsible for the modifications in K8 phosphorylation we observed. K8 and K18 give different patterns of phosphorylated cells indicating that, under the same conditions, K8 and K18 phosphorylation is regulated differently. Conclusions =========== Our results show that increases in HSP70i, K8/18 expression and K8/18 phosphorylation constitute early events in the response of hepatocytes to the presence of GF. These observations support a role for keratins in preserving cellular integrity during stress conditions induced by the presence of a chemical agent \[[@B33],[@B35]\]. HSP70i expression in hepatocytes after GF-treatment is not directly related to K8/18 phosphorylation at the studied sites: K8 S79, K8 S436 and K18 S33. With regard to MB formation, it appears that both HSP70i and K8/18 phosphorylation might contribute to the IF aggregation processes. The involvement of K8/18 phosphorylation in MB formation seems to be related only to specific sites and dependent on mouse genetic inheritance. Methods ======= Experimental design ------------------- Experiments were performed with adult C3H mice (Charles River Canada, St-Constant, QC) and FVB/n mice (Baribault et al. 1994) weighing 25 to 30 g. Two mouse strains were used to minimize the potential effect of different genetic background on the response of hepatocytes and to facilitate the interpretation of the data. All animals were housed with a 12-hour light-dark cycle and allowed the consumption of water and of a standard mouse semi-synthetic diet (Texlad Test Diet, Madison, WI), both *ad libitum*. GF-treated mice were fed a diet containing 2.5% (w/w) GF (Schering Corp., Kenilworth, NJ) for different periods of time: 2 weeks, 6 weeks and 5 months according to the method of Denk et al. \[[@B28]\]. Control mice were fed the same diet without GF. For control and each period of GF-treatment, experimental groups included 3 animals. Mice were sacrificed by cervical dislocation and livers were snap frozen in methylbutane precooled with liquid nitrogen and stored at -70°C before use. All experiments were conducted according to the requirements of Canadian Council Animal Care and the \"Université du Québec à Trois-Rivières\" Animal Welfare Committee. For microscopical studies, paraformaldehyde and cold acetone were routinely used, as fixatives, to ensure that the staining patterns were not a consequence of the fixative used. Reagents -------- The antibodies used were as following: Troma 1, a rat monoclonal antibody (rAb) that recognizes K8 \[[@B56]\]; LJ4, a mouse monoclonal antibody (mAb) that recognizes human K8 pS73 equivalent to mouse K8 pS79 \[[@B25]\]; mAb 5B3 that recognizes K8 pS431 equivalent to mouse K8 pS436 \[[@B48]\]; 8250, a rabbit polyclonal antibody (pAb) that recognizes K18 pS33 \[[@B46]\] and a pAb that recognizes the stress inducible form of HSP70, HSP70i (Stressgen, Victoria, BC). The secondary antibodies for fluorescence microscopy were as follows: tetramethylisothiocyanate (TRITC) or fluorescein isothiocyanate (FITC) conjugated goat anti-rat IgG, FITC-conjugated donkey anti-rabbit IgG (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON). The M.O.M. kit and Avidin/Biotin blocking kit (Vector^®^Laboratories Canada, Burlington, ON) were used to perform immunolabelling with mAbs LJ4 and 5B3. The secondary antibodies used for Western blotting were as follows: biotinylated goat anti-rat IgG, biotinylated donkey anti-mouse IgG and peroxydase donkey anti-rabbit IgG (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON). Other reagents used were: Horseradish Streptavidin Peroxydase-conjugated (SPC) (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON), Bovine Serum Albumin (BSA) (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON), Leupeptin (Sigma-Aldrich Canada, Oakville, ON), Pepstatin (Sigma-Aldrich Canada, Oakville, ON), Aprotinin (Sigma-Aldrich Canada, Oakville, ON), Normal Horse Serum (NHS) (Vector^®^Laboratories Canada, Burlington, ON), Luminol (Amersham Pharmacia Biotech, Oakville, ON). Gel electrophoresis and immunoblotting -------------------------------------- Livers were homogenized in 62.5 mM Tris-HCl, pH 6.8 containing 2.3 % (w/v) SDS, 50 mM sodium fluoride, 10 mM EDTA, 1 mM sodium pyrophosphate, 1 mM DTT, 1 mM PMSF, 1 μM leupeptin (Sigma-Aldrich Canada, Oakville, ON), 1 μM pepstatin (Sigma-Aldrich Canada, Oakville, ON), 2.5 μg/ml aprotinin (Sigma-Aldrich Canada, Oakville, ON). Proteins were separated by electrophoresis on 10% SDS-polyacrylamide gels \[[@B57]\]. Protein concentration was determined by the Lowry method, modified for the presence of SDS \[[@B58]\], and equal amounts of proteins (5 to 12.5 μg) were loaded on each well. Gels were stained with 0.1% Coomassie Blue, or transferred onto nitrocellulose membranes (Biorad laboratories Canada, Mississauga, ON), and processed for immunodetection. Membranes were blocked overnight with 5% (w/v) non-fat dry milk (Carnation, Nestlé^®^) in PBS (Phosphate Buffer Saline, 0.137 M NaCl, 2.7 mM KCl, 4.3 mM Na~2~HPO~4~, 14.7 mM KH~2~PO~4,~pH 7.2), incubated with the primary antibodies for 45 min, at room temperature, washed in PBS containing 0.2% (v/v) Tween 20 and incubated for 45 min with the appropriate secondary antibody: biotinylated goat anti-rat IgG (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON), biotinylated donkey anti-mouse IgG (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON) and horseradish peroxydase donkey anti-rabbit IgG (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON). When biotinylated secondary antibodies were used, membranes were washed with PBS-Tween 20 and incubated with streptavidin conjugated with horseradish peroxydase (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON) for 30 min and washed with PBS-Tween 20. The chemiluminescent horseradish peroxydase substrate Luminol (Amersham Pharmacia Biotech, Oakville, ON) was added to the membranes according to recommendations of the company, and membranes were exposed to Blue X-Omat X-ray film sheets (Mandel Scientific Company, Guelph, ON) to localize antibody binding. Fluorescence microscopy ----------------------- Cryosections (4 μm) of fresh liver were fixed with 4% (w/v) paraformaldehyde in PBS pH 7.2 for 20 min, at room temperature, and rinsed in PBS or TBS (Tris Buffer Saline, 10 mM Tris-HCl, 0.138 M NaCl, 2.7 mM KCl, pH 7.4) upon staining protocols. Since fixation can affect antibody-binding capacity, cryosections were also fixed with cold acetone (-20°C) for 10 min. For the detection of K8, sections were incubated with rAb Troma 1 at room temperature, washed in PBS and incubated with a FITC or a TRITC conjugated goat anti-rat IgG (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON) for 45 min, at room temperature. For immunostaining of K18 pS33, sections were incubated for 1 hour, at room temperature with anti-K18 pS33 (8250) diluted in PBS containing 10% (w/v) BSA, washed in PBS and incubated for 45 min with a FITC conjugated donkey anti-rabbit IgG in PBS containing 10% BSA (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON). Immunostaining with anti-K8 pS79 (LJ4) and anti-K8 pS436 (5B3) mAbs, was done using the M.O.M. (mouse on mouse) detection kit (Vector^®^Laboratories Canada, Burlington, ON) and an Avidin/Biotin blocking kit (Vector^®^Laboratories Canada, Burlington, ON) according to recommendations of the company. Normal horse serum (Vector^®^Laboratories Canada, Burlington, ON) was added to solution during incubation step with secondary antibody. For heat shock proteins staining, liver sections were incubated with anti-HSP 70i diluted in TBS containing 10% BSA for 45 min at 37°C, washed in TBS and incubated for 45 min at 37°C with a FITC conjugated donkey anti-rabbit IgG (Jackson Immunoresearch, Bio/Can Scientific, Mississauga, ON) diluted in TBS containing 10% BSA. For detection of HSP70i, K8 pS79 and K8 pS436, the sections were treated with 1% (v/v) Nonidet P-40 (Sigma-Aldrich Canada, Oakville, ON) following fixing step with 4% paraformaldehyde. The tissues were mounted in P-phenylene diamine diluted in 50% (v/v) glycerol. The slides were kept at -20°C and photomicrographs were collected using an Olympus^®^BX60 photomicroscope. List of abbreviations ===================== HSP70i -- inducible form of 70 kDa Heat shock protein. GF -- griseofulvin. IFs -- intermediate filaments. K8 -- keratin 8. K8/18 -- keratin 8 and keratin 18. K8 S79 -- serine 79 on keratin 8. K8 pS79 -- phosphorylated serine 79 on keratin 8. MBs -- Mallory bodies. Authors\' contributions ======================= MF carried out all western blotting analyses, performed the immunofluorescence studies and participated in drafting the manuscript. LV participated in the design of the study. MC participated in the design of study, its coordination and drafting the manuscript. All authors read and approved the final manuscript. Acknowledgements ================ We thank Dr B. Omary from the Department of Medicine, Palo Alto VA Medical Center and Stanford University, Palo Alto, California, for providing antibodies directed against phosphorylated keratins. We also thank Dr N. Marceau from the \"Centre de recherche de L\'Hôtel-Dieu de Québec (CHUQ)\", Québec, for providing Troma 1 and FVB/n mice. This work was supported by a grant from NSERC.

PubMed Central

2024-06-05T03:55:47.397056

2004-8-12

PMC516019

Background ========== The biological effects of PAHs are often mediated by oxidative metabolism of the parent hydrocarbon to reactive intermediates that adduct DNA and induce oxidative stress \[[@B1]\]. In the kidney, PAHs elicit cell type-specific effects that differentially influence glomerular versus tubular epithelial cell structure and function. BaP and ANTH selectively injure glomerular mesangial cells, while CHRY preferentially targets cortico-tubular epithelial cells \[[@B2],[@B3]\]. The study of single chemical effects has provided fundamental information on the nephrotoxic potential of specific PAHs, but human exposure to these group of chemicals is rarely limited to a single agent, and most often involves exposure to PAH mixtures \[[@B4]\]. Thus, a more realistic approach is to evaluate the cellular, biochemical, and molecular mechanisms by which PAHs interact to produce additive, synergistic or antagonistic interactions. Such studies have demonstrated that binary and ternary mixtures of PAHs yield paradoxical antagonistic interactions *in vitro*\[[@B3]\]. A toxicological interaction is a circumstance in which exposure to two or more chemicals results in qualitative or quantitative modulation of the biological response elicited by individual agents. Toxicological interactions may be mediated by changes in the absorption, distribution, metabolism and excretion of one or more of the chemicals present in the mixture. Since the ability of PAHs to compromise cellular and genomic integrity often requires bioactivation by cytochrome P-450 enzymes (*CYPs*) to reactive intermediates, their role in PAH-induced environmental diseases is profound \[[@B5]\]. The interaction of PAHs with *CYPs*is unique in that the expression of genes that encode for *CYP*-associated activities is itself regulated by the PAH substrates they metabolize. Shimada et al. \[[@B6]\] have shown that BaP and CHRY induce *Cyp1a1*and *1b1*through the aryl hydrocarbon receptor (Ahr), and that the enzymes encoded by these genes mediate toxicity and tumorigenicity. The Ahr belongs to the basic helix loop helix/PAS family of proteins \[[@B7]\]. The activation of cytoplasmic complexes containing the Ahr depends on ligand binding to the receptor, nuclear translocation and formation of active heterodimers with a nuclear protein called Arnt \[[@B8]\]. The AhR-Arnt complex binds to s pecific *cis*-acting responsive elements known as xenobiotic responsive elements located in the promoters and enhancers of target genes, including *CYPs*themselves \[[@B7]\]. PAHs or halogenated aromatic hydrocarbons function as ligands of the Ahr. The present studies were conducted to evaluate profiles of *Cyp1a1*and *Cyp1b1*inducibility in binary PAH mixtures, and their impact on BaP bioactivation. Evidence is presented that chemical-specific differences in the regulation of *Cyp1a1*and *Cyp1b1*contribute to differential metabolic activation of PAHs in binary mixture. On the basis of these findings it is concluded that interactions between BaP, ANTH and CHRY involve altered expression of cytochrome P450s that modulate bioactivation profiles and nephrotoxic/ nephrocarcinogenic potential. Materials and Methods ===================== Materials --------- BaP, ANTH and CHRY were purchased from Sigma Chemical Co. (St. Louis, MO). RPMI 1640 and M199 were purchased from GIBCO-BRL (Grand Island, NY, USA). All other chemicals were from Sigma Chemical Co. Cell culture/chemical treatments -------------------------------- Rat glomerular mesangial cells in serial culture were seeded on 6-well plates at a density of 200 cells/mm^2^. At least three replicates were used for each chemical concentration tested in multiple experiments. The concentrations examined are similar to those used in previous studies and representative of those encountered in the environment. Cultures were challenged with selected PAHs for 24 hr at concentrations ranging from 0.03 to 30 μM. Stock solutions of PAHs were dissolved in DMSO with final DMSO concentrations never exceeding 0.1%. Cells, RNA, or protein were harvested after chemical challenge and processed for biochemical measurements. RNA extraction and analysis --------------------------- Total RNA was extracted using Tri reagent (Molecular Research Center, Inc., Cincinnati, OH) according to manufacturer\'s specifications. Cells were scraped using 1.0 ml of Tri reagent and allowed to sit at room temperature for 5 minutes to dissociate nucleoprotein complexes and then combined with 0.2 ml chloroform, vortexed and allowed to sit at room temperature for 2 minutes. After centrifugation at 12,000 × g (4°C) for 15 minutes, the upper aqueous layer was mixed with and equal volume of isopropanol and stored at -20°C overnight. This solution was then centrifuged for 15 minutes at 12,000 × g (4°C) and the pellet washed with 70% ethanol, dried, and resuspended in 20 μl of RNase free water. RNA concentration was determined spectrophotometrically at 260 nm. Northern analysis ----------------- Ten μg of total RNA were dissolved in RNase free water, mixed with 3.5 μl formamide, 1 μl of 37% formaldehyde, 1.0 μl of MOPS buffer and 15% 6X gel loading buffer, and denatured by heating at 55°C for 10 minutes. Total RNA was separated by electrophoresis on a formaldehyde denaturing gel (1.2% agarose, 1 M formaldehyde and 10X MOPS) in 1X MOPS buffer and transferred onto a nylon membrane by capillary transfer. Membranes were dried at room temperature, UV crosslinked and hybridized with 32P labeled cDNA probes synthesized using High Prime (Boehringer Mannheim, Germany). The β-tubulin probe was obtained from a 1.6 Kb fragment cloned into a pBluescript plasmid at the EcorI site. The *Cyp1a1*probe was obtained from a 1.2 kb Pst1 fragment from a pUC18 vector (ATCC) and the 1 kb *Cyp1b1*probe was kindly provided by Dr. Colin Jefcoate (University of Wisconsin, Madison, WI). Following hybridization the blots were subjected to stringent washes, dried at room temperature, and exposed to x-ray film in -80°C for 24 hours. Cyp-related enzyme activities ----------------------------- Confluent subcultures of mesangial cells were grown in 100 mm dishes and treated with selected PAHs or their mixtures for 24 hr. Cells were then scraped and collected after the addition of 5 ml of ice-cold Tris-sucrose buffer (pH 8.0), then centrifuged for 5 min at 50 g (4°C). The supernatants were removed and the pellet resuspended in 300 μL of Tris-sucrose buffer. Two 100 μl aliquots containing cellular protein were processed for fluorometric enzyme analysis, while 50 μL of sample was used to measure protein concentration \[[@B9]\]. Aryl hydrocarbon hydroxylase (AHH) assay ---------------------------------------- Cultures were processed for measurements of AHH activity as described by Nebert and Gelboin \[[@B10]\]. A 100 μl aliquot was combined with 1 ml of reaction mixture containing 0.1 M HEPES (pH 8.0) and 0.4 mM NADPH. Samples were pre-incubated at 37°C for 2 minutes, and the reaction initiated by addition of 80 μM BaP dissolved in 40 μM of methanol. Samples were incubated for 15 minutes and the reaction terminated by addition of 1 ml of ice cold acetone and 3.25 ml of hexane. After vortexing, 2 ml of the organic layer was collected and extracted with 5 ml of 1 N NaOH. Samples were vortexed and the NaOH fraction read on a spectroflurometer at a wavelength of 396 nm excitation and 522 nm emissions. The spectroflurometer was calibrated using authentic 3-OH BaP standards. Ethoxyresorufin-O-deethylase activity ------------------------------------- Cultures were processed for measurements of EROD activity as described by Burke and Mayer \[[@B11]\] with modifications. Briefly, 1.2 ml of 0.1 M HEPES buffer (pH 7.5) containing 0.1 mg of NADH, 0.1 mg of NADPH, 1.5 mg of magnesium sulfate and 1.1 mg of BSA was added to 100 μl of sample. The tubes were incubated for 2 minutes at 37°C prior to addition of 100 mM ethoxyresorufin. 50 ml of ethoxyresorufin was added to each tube and allowed to incubate at 37°C for 15 minutes. The reaction was terminated by addition of 2.5 ml of methanol. Samples were incubated for an additional 2 minutes to allow for protein flocculation and centrifuged for 10 minutes at 1500 × g. EROD activity in the supernatant was measured fluorometrically at a wavelength of 550 nm excitation and 585 nm emission as described by pohl and Fouts \[[@B12]\]. The spectrofluorimeter was calibrated using authentic resorufin standards. HPLC analysis ------------- PAH metabolism was analyzed according to the method of Selkirk et al. \[[@B13]\]. Chemical separation, identification, and quantification was performed on a high-performance liquid chromatograph (Beckman model 334) fitted with a Rainin Microsorb C18 reverse-phase column (4.6 × 250 mm), using a 22.5 min linear gradient of 75--100% methanol at a flow rate of 1.0 ml/min. A 20 μl sample was injected into the column. The PAH metabolites were monitored by ultraviolet absorption at 254 nm. Identification of metabolites was made by comparison of retention times to known standards. Results ======= *CYP*expression profiles and toxicological interactions in renal cells after challenge with PAH mixtures -------------------------------------------------------------------------------------------------------- Antagonistic interactions among nephrotoxic/nephrocarcingenic PAHs may be mediated by modulation of metabolic activation profiles in mesangial cells. To test this hypothesis, steady state levels of *Cyp1a1*and *Cyp1b1*mRNAs were evaluated by Northern analysis in cultured mesangial cells treated with 0.03 -- 30 μM BaP alone, or in combination with 3 μM ANTH and CHRY (Figs [1](#F1){ref-type="fig"} and [2](#F2){ref-type="fig"}). *Cyp1a1*mRNA was not expressed constitutively, but was highly inducible by BaP in a concentration-dependent manner (Fig [1A](#F1){ref-type="fig"}). ANTH alone (0.03 -- 30 μM) did not induce *Cyp1a1*at any concentration, but modestly enhanced (3 μM) the mesangial cell response to BaP (3 or 30 μM). In contrast to the *Cyp1a1*gene, *Cyp1b1*was constitutively expressed in mesangial cells (Fig [1B](#F1){ref-type="fig"}). Treatment with BaP induced concentration-dependent increases in steady state *Cyp1b1*mRNA levels. ANTH alone, induced *Cyp1b1*mRNA by 5--6 fold at the higher concentrations examined. Combined treatment of glomerular mesangial cells with BaP (0.03 μM) and ANTH (3 μM) slightly enhanced the response to individual hydrocarbons at the 0.03 μM concentration, but this enhancement was dissipated at the higher concentrations. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Northern analysis of P-450 induction in rGMCs treated with BaP, ANTH and their binary mixtures.**mRNA expression of *Cyp1a1*in rGMCs challenged with BaP, ANTH and their binary mixtures for 24 hr (A). mRNA expression of *Cyp1b1*in rGMCs challenged with BaP and ANTH and their binary mixtures for 24 hr (B). RNA extraction and analysis were performed as described in methodology. β-tubulin was analyzed to asses loading and transfer efficiency. These results shown are representative of three separate experiments. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Northern analysis of P-450 induction in rGMCs treated with BaP, CHRY and their binary mixtures.**mRNA expression of *Cyp1a1*in rGMCs challenged with BaP and CHRY and their binary mixtures for 24 hr (A). mRNA expression of *Cyp1b1*in rGMCs challenged with BaP and CHRY and their binary mixtures for 24 hr (B). RNA extraction and analysis were performed as described in methodology. β-tubulin was analyzed to asses loading and transfer efficiency. These results shown are representative of three separate experiments. :::  ::: The metabolic interaction between BaP and CHRY was examined next. As expected, BaP induced *Cyp1a1*and *Cyp1b1*mRNA levels in a concentration-dependent manner (Figure [2](#F2){ref-type="fig"}, panels A and B). CHRY was also a potent *Cyp*inducer, and in fact elicited greater induction of *Cyp1a1*and *Cyp1b1*mRNA than BaP (Fig [2](#F2){ref-type="fig"}, panels A and B). CHRY induction of *Cyp1a1*at the 3 and 30 μM concentrations was 2-fold higher than the response elicited by BaP at the same concentration (Figure [2A](#F2){ref-type="fig"}). As with *Cyp1a1*, CHRY markedly induced *Cyp1b1*steady state mRNA levels at all concentrations examined (Fig [2B](#F2){ref-type="fig"}). Combined treatment of mesangial cells with BaP (0.03 -- 30 μM) and CHRY (3 μM) yielded modest antagonistic responses for both *Cyp1a1*and *Cyp1b1*, particularly at the highest concentrations examined. For *Cyp1b1*, the induction response in cells treated with 30 μM BaP and 3 μM CHRY was reduced by 30% relative to either hydrocarbon alone. CYP enzymatic activities in renal cells after challenge with PAH mixtures ------------------------------------------------------------------------- Because mRNA expression does not always correlate with changes in protein levels, measurements of EROD and AHH activity were completed in mesangial cells treated with PAHs alone, or in binary mixture. EROD, an enzyme activity encoded by both *Cyp1a1*and *Cyp1b1*, was inducible by all hydrocarbons in a concentration-dependent manner (Fig [3](#F3){ref-type="fig"}). Basal and inducible EROD activities in mesangial cells were considerably lower than those in cultured rat hepatocytes (not shown). A greater than 6-fold enhancement of EROD activity was observed in cells treated with 0.3 μM of BaP, with significant decreases observed as the BaP concentration increased. ANTH and CHRY also induced EROD, but induction patterns for these hydrocarbons were remarkably different. ANTH was a weak inducer of EROD, with only a 3-fold induction observed at 30 μM, while CHRY elicited a greater than 14-fold increase in enzymatic activity. As with BaP, reductions in activity were observed at the highest CHRY concentrations. Different profiles were observed in cells treated with BaP in combination with either ANTH or CHRY. Combined treatment of mesangial cells with BaP and ANTH completely inhibited EROD inducibility. In the case of CHRY, co-treatment with BaP yielded an erratic response, with less than additive interactions observed at the lowest concentrations, and modest inhibition observed as hydrocarbon concentrations increased. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **EROD activity in rGMCs treated with PAHs.**EROD activity in rGMCs challenged with BaP, ANTH and CHRY alone or the binary mixtures of BaP with ANTH or CHRY. These results shown are representative of three separate experiments. :::  ::: AHH activity is also encoded by the *Cyp1a1*and *Cyp1b1*genes. BaP was a potent inducer of AHH activity in mesangial cells, with concentration-dependent increases observed over the full concentration range examined. A greater than 40-fold induction in enzymatic activity was observed at 30 μM BaP (Fig [4](#F4){ref-type="fig"}). Individual treatment with ANTH or CHRY did not modulate AHH activity. Combined treatment of mesangial cells with BaP and ANTH repressed AHH activity at the lower BaP concentrations, but the negative interaction was dissipated at higher concentrations. Likewise, CHRY inhibited AHH inducibility by BaP, with greater than 50% reduction observed when compared to cells treated with BaP alone. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **AHH activity in rGMCs treated with PAHs.**AHH activity in rGMCs challenged with CHRY, ANTH and BaP alone or the binary mixtures of CHRY with ANTH or BaP. These results shown are representative of three separate experiments. :::  ::: Cellular hydrocarbon uptake and metabolic profiles of PAHs ---------------------------------------------------------- To further evaluate cellular mechanisms of toxicological interactions in PAH mixtures, BaP metabolism was examined in mesangial cells treated with 30 μM BaP alone, or in combination with 3 μM ANTH or 3 μM CHRY. At least 50% of the parent compound was reproducibly detected in mesangial cells treated with BaP (Table [1](#T1){ref-type="table"}). ANTH was readily taken up by mesangial cells, and did not influence the cellular uptake of BaP. 3-hydroxy-BaP was the primary oxidative metabolite detected in mesangial cells treated with BaP, with other metabolites representing less than 1% of the total metabolite pool detected (not shown). Combined treatment of cells with ANTH significantly reduced BaP metabolism, with greater than 80% reduction in detectable metabolite levels (Table [1A](#T1){ref-type="table"}). Mesangial cells readily took up CHRY, with greater than 77% of the parent compound detected at the end of the treatment. As with ANTH, CHRY did not influence the cellular uptake of BaP. However, a 46% reduction in measurable 3-hydroxy-BaP levels was observed in cells subjected to binary hydrocarbon treatment (Table [1B](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Oxidative metabolism of benzo(a)pyrene (BaP) alone or in combination with CHRY (A) and ANTH (B) in rat glomerular mesangial cells (rGMCs). ::: **A** -------------- -------------- -------------------- --------------------- -------------------- BaP 30 14.39 ± 0.18 \_ 0.026 ± 0.0 BaP/ANTH 30/3 13.72 ± 1.41 3.7 ± 0.49 0.005 ± 0.003 **B** **Chemical** **Con (μM)** **Calc. BaP (μM)** **Calc. CHRY (μM)** **BaP 3(OH) (μM)** BaP 30 17.3 ± 0.29 \_ 0.013 ± 0.002 BaP/CHRY 30/3 17.4 ± 1.2 2.31 ± 0.12 0.007 ± 0.002 ::: Discussion ========== Cyp mRNA expression profiles and toxicological interactions ----------------------------------------------------------- PAHs elicit a broad spectrum of toxic and carcinogenic effects in multiple organ systems, including the kidney \[[@B14]\]. To study the complexity of chemico-biological interactions following exposures to multiple PAH carcinogens, we evaluated the renal cell-specific response to binary and ternary mixtures of BaP, ANTH and CHRY \[[@B3]\]. Challenge of renal mesangial and cortico-tubular epithelial cells with BaP in combination with ANTH or CHRY yielded unexpected antagonistic interactions that may be partly explained by differential regulation of enzymes involved in PAH metabolism. Renal *Cyp1a1*and *Cyp1b1*are particularly relevant since these enzymes mediate the conversion of PAHs to intermediates that induce oxidative stress and bind covalently to DNA in the kidney, and are the primary enzymes responsible for bioactivation of carcinogenic PAHs in other tissues \[[@B6]\]. In mesangial cells,*Cyp1a1*mRNA was undetectable under constitutive conditions, highly inducible by BaP and CHRY, and refractory to ANTH (Figs [1](#F1){ref-type="fig"} and [2](#F2){ref-type="fig"}). In contrast, *Cyp1b1*mRNA was constitutively expressed and highly inducible by all three hydrocarbons. BaP and CHRY were more potent inducers of *Cyp1a1*and *Cyp1b1*than ANTH (Figs [1](#F1){ref-type="fig"} and [2](#F2){ref-type="fig"}), a profile consistent with their relative abilities to activate Ahr signaling in mammalian cells (15). Shimada et al. \[[@B16]\] have shown that liver and lung *Cyp1a1*and *Cyp1b1*mRNAs are highly induced in AhR(+/+) mice by a single intraperitoneal injection of carcinogenic PAHs, and that 6-aminochrysene, chrysene, benzo \[e\]pyrene, and 1-nitropyrene weakly induced *Cyp1a1*and *Cyp1b1*mRNAs, while non-carcinogenic hydrocarbons, such as anthracene, pyrene, and fluoranthene, were poor or inactive enzyme inducers. These findings indicate that the toxicity and carcinogenicity profiles of PAHs may be defined on the basis of their ability to regulate *Cyp1a1*and *Cyp1b1*at either the mRNA or protein level. The tissue-specific induction of *Cyp1a1*and *Cyp1b1*mRNAs by PAHs and polychlorinated biphenyls (PCBs) has been investigated in wildtype and AhR-deficient C57BL/6J mice \[[@B17]\]. While expression of *Cyp1a1*is AhR-dependent, *Cyp1b1*is constitutively expressed in various organs in male and female Ahr (+/+) and Ahr (-/-) mice. *Cyp1b1*is of interest because it encodes for AHH and EROD, the predominant PAH-metabolizing activities in rodent embryos \[[@B17]\]. Important roles of *Cyp1b1*in PAH carcinogenesis have been proposed by Gonzalez and co-workers \[[@B18]-[@B20]\] who observed that *Cyp1b1*knock-out mice expressing significant levels of *Cyp1a1*are highly resistant to lymphoma formation by 7,12-DMBA. *In vitro*human studies with recombinant enzymes have shown that *Cyp1b1*is more active (about 10-fold) than *Cyp1a1*in the formation of BaP-7,8-diol \[[@B21]\]. CYP enzymatic activities and toxicological interactions ------------------------------------------------------- Measurements of EROD and AHH were used to monitor kidney microsomal catalytic activities (Figs [3](#F3){ref-type="fig"} and [4](#F4){ref-type="fig"}). The profile of enzyme induction was significantly different between BaP, ANTH and CHRY. The basal activities of EROD and AHH were very low in control cultures, but highly inducible in response to BaP and CHRY. BaP was considerably more potent than ANTH as an inducer of EROD and AHH, while CHRY only induced EROD activity. Interestingly, CHRY was a better inducer of *Cyp1a1*mRNA, but induction at the mRNA level was not associated with corresponding increases of enzymatic activity. The greater potency of BaP as an inducer of enzymatic activities compared to ANTH and CHRY is consistent with previous observations showing that BaP is the most toxic PAH to renal mesangial cells \[[@B3]\]. The profiles of mRNA and protein induction elicited by single exposures to BaP, ANTH, and CHRY alone were markedly different from those following challenge in binary mixture. ANTH enhanced both *Cyp1a1*and *Cyp1b1*inducibility by BaP, but EROD and AHH activities were reduced in binary mixture. CHRY, on the other hand, modestly inhibited *Cyp1a1*and *Cyp1b1*mRNA inducibility by BaP and also reduced AHH activity. Because both ANTH and CHRY antagonize the mesangial cell response to BaP (3), these findings implicate AHH as the primary enzymatic target for antagonistic interactions among nephrotoxic PAHs. Oxidative metabolic profiles and toxicological interactions ----------------------------------------------------------- The above interpretation is supported by the marked inhibition of 3-OH BaP formation seen in cells co-treated with BaP and ANTH or CHRY (Table [1](#T1){ref-type="table"}). A role for EROD in mesangial cell injury cannot be ruled out, however, since ANTH selectively antagonized EROD activity and inhibited hydrocarbon metabolism. Thus, patterns of Cyp inducibility and oxidative metabolism may account for differences in the responses to BaP, ANTH and CHRY in kidney cells, and their interactions in simple and complex mixtures. Conclusion ========== A major implication of our findings is that despite similarities in gene and protein inducibility among structurally-related PAHs, their behavior may be influenced by the reactivity of oxidative intermediates generated during the course of cellular metabolism. As such, PAH substrates in complex mixtures may compete for, and inhibit, the same metabolizing enzymes that act upon them to give rise to antagonistic interactions that protect against further chemical toxicity. Interactions of this nature are among the most commonly encountered in environmental mixtures \[[@B22]\], but the magnitude and capacity of these interactions for most relevant environmental nephrocarcinogens is unknown. Acknowledgments =============== The authors wish to thank Dr. Colin Jefcoate (University of Wisconsin, Madison, WI) for providing *Cyp1b1*probe and Dr. Lingyu He for assistance with HPLC analysis. This study was supported by NIEHS Grants ES04917 and ES09106.

PubMed Central

2024-06-05T03:55:47.399654

2004-8-17

PMC516020

Background ========== Viral promoters are commonly used as regulatory elements in gene therapy vectors due to their strong activity in various cell lines in vitro. Probably the most widely used promoter in mammalian expression systems is the human cytomegalovirus immediate-early gene (CMV) promoter. The CMV promoter induces high-level constitutive expression in a variety of mammalian cell lines \[[@B1]\]. In many gene therapy applications, however, an inducible or cell specific promoter would be more appropriate. A regulated transgene expression system in mammalian cells is preferable for effective and safe gene therapy and for the study of gene function in cell biology. The most important features of an inducible promoter would be 1) low basal expression levels; 2) high induced expression; and 3) inducer-specific, modulated expression \[[@B2]\]. Our need for a mammalian expression vector promoter for preventative gene therapy that would be induced by bovine leukemia virus (BLV) infection or BLV Tax protein expression led us to investigate the use of the BLV promoter for gene therapy. The U3 region of the BLV promoter, located in the 5\' long terminal repeat (LTR), contains several important *cis*-acting elements in addition to the CAAT box, TATA box, and transcription start site \[[@B3],[@B4]\]. The major regulatory elements are three copies of an imperfectly conserved 21-bp sequence called the tax responsive element (TxRE). The TxREs are essential for the promoter\'s responsiveness to the Tax transactivator protein encoded by the 3\' end of the proviral genome \[[@B5]\]. These *cis*-elements contain motifs resembling the cyclic AMP-responsive element (CRE) as well as an E box sequence \[[@B6]\]. Tax does not bind directly to the TxRE but interacts with cellular proteins that recognize the CRE including the transcription factors CREB, ATF-1, and ATF-2 \[[@B7]-[@B9]\]. The transcription factor AP4 can potentially bind to the E box sequence and is important in Tax activation \[[@B10]\]. There is a glucocorticoid responsive element (GRE) that responds to dexamethasone in the presence of glucocorticoid receptors and Tax \[[@B11]\]. A nuclear factor κB (NFκB) binding site responds to phorbol 12-myristate 13-acetate (PMA) treatment \[[@B12]\]. Finally, there is a Tax transactivator independent site specific for the B cell transcription factors PU.1 and Spi-B \[[@B13]\]. This PU.1/Spi-B binding site may be involved in the B lymphocyte tropism of BLV. We hypothesized that the BLV promoter could be used in mammalian expression vectors for regulated high-level gene expression. Our approach was to compare reporter gene expression driven by either the BLV promoter or CMV promoter in different cell types with or without BLV infection or Tax induction. Our results demonstrate that the BLV promoter can be induced to express the reporter gene to levels as great as the constitutive CMV promoter. Methods ======= Cell culture ------------ All cells used in these studies were maintained in RPMI 1640 medium (Invitrogen) supplemented with 10% fetal calf serum (FCS), 4.5% dextrose, 1 mM sodium pyruvate, and antibiotic-antimycotic solution (100 μg/ml penicillin G sodium, 100 μg/ml streptomycin sulfate, 0.25 μg/ml amphotericin B). In addition, the following concentrations of drugs were added for each selective media: Blasticidin-S (Invivogen) 10 μg/ml; G418-sulfate (Alexis Biochemical) 400 μg/ml. The following cell lines were used: D17 \[dog osteosarcoma; ATCC CCL-183; \[[@B14]\]\], FLK \[sheep kidney; BLV expresser; \[[@B15]\]\], BL3.1 \[bovine B-lymphosarcoma; BLV expresser; ATCC CRL-2306; \[[@B16]\]\]. Primary bovine B cells were supplemented with 10 ng/ml each of recombinant human interleukin-4 and interleukin-7 (Peprotech, Inc.), and gamma-irradiated (4,000 R) murine CD40L-expressing L cells (J558L; a gift from Philip Griebel) as described elsewhere \[[@B17]\]. For the TSA experiments, Trichostatin A (Sigma) was supplemented at 500 nM for 48 h. Cells were cultured at 37°C in a 5% CO2 humidified atmosphere. Viable cells were identified by trypan blue dye exclusion, and cell number was counted with a hemacytometer. Primary bovine B cells were purified as follows. Peripheral blood mononuclear cells (PBMC) were isolated from heparinized cow blood through a ficoll density gradient as previously described \[[@B18]\]. B cells were separated from the PBMCs using the MiniMACS system following the manufacturer\'s (Miltenyi Biotec) protocol. Briefly, 1 × 10^7^cells were stained for 15 min at 6° -- 12°C with 10 μg/100 μl total volume anti-IgM (PIG45A; VMRD, Inc.). After washing, 20 μl/100 μl total volume of MACS rat anti-mouse IgG2a+b microbeads were mixed with the cells and incubated for 15 min at 6° -- 12°C. Cells were thoroughly washed, and magnetically separated. These IgM+ cells were considered primary B cells. Microfluorimetry using anti-IgM (PIG45A; VMRD, Inc.) indicated 90% purity. Stably transduced cell lines were generated after one week in selective media. Primary B cells were analyzed after one week in selective media since they began to die out after two weeks in culture. Vector construction ------------------- The plasmid pBLV913 (a gift from David Derse), coding for an infectious molecular clone of BLV \[[@B5]\] was used as the source for the BLV promoter and BLV Tax sequences. Briefly, the BLV promoter from the U3 region of 3\' LTR of BLV was isolated from plasmid pBLV913 (Derse) as a 345 bp fragment (GenBank LOCUS BLVCG, ACCESSION K02120 bp 8096 -- 8440) and cloned in place of the CMV promoter fragment into pLNCX (Clontech; Genebank LOCUS SYNMMLPLN3 ACCESSION M28247 -- CMV promoter removed as *Bam*HI-*Hind*III fragment) to create the vector pLNBlv. The pLNBlv and pLNCX retrovector plasmids were modified to place the Gateway Rfa cassette (1.7 kb; Invitrogen) downstream of the internal promoters (BLV or CMV) in order to simplify further cloning, to create retrovector plasmids pLNBlv-G or pLNC-G. For enhanced protein expression, the WPRE element (from plasmid BluescriptII SK+ WPRE-B11 (a gift from Tom Hope--the same as bp 2717--3309 of Genbank Locus OHVHEPBA ACCESSION J04514) was cloned downstream of the Gateway Rfa cassette with standard cloning methods to create vectors pLNC-GW and pLNBlv-GW. The source for firefly luciferase encoding sequence was pGEM-luc (Promega). The luciferase coding sequence was subcloned into pENTR1A (Invitrogen) to engineer the Gateway entry vector pENTR1A/luc. The Luciferase gene was recombined into pLNC-GW or pLNBlv-GW using LR Clonase (Invitrogen) per manufacturer\'s instructions. The promoter-less luciferase expression control vector pLN\[\]W/luc was engineered by removing the BLV promoter (*Bam*HI digest) from pLNBlv/luc. BLV Tax (Genbank Locus AAF97920) was isolated by reverse transcription PCR from FLK cells and subcloned into pENTR1A (Invitrogen) to engineer the Gateway entry vector pENTR1A/Tax. The Tax gene was recombined into pLBC-GW where the neomycin resistance gene of pLNC-GW was replaced with the blasticidin resistance gene. Throughout these studies we assayed expression vectors with and without the WPRE. WPRE enhanced transgene expression in all cell lines used, and in a promoter-independent fashion (about 2-fold greater for BLVp and CMVp in D17 cells). Subsequently, all data shown in this report are only with vectors containing WPRE. Cell transfection and transduction ---------------------------------- Retrovirus-mediated gene transfer was accomplished using the BD Retro-X System (BD Biosciences Clontech) following the manufacturer\'s suggested protocol. Briefly, 100 mm × 20 mm tissue culture dishes (Falcon) were seeded with the packaging cell line GP2-293 at 70--90% confluency. Each dish of GP2-293 cells was co-transfected with 5 μg each of retroviral vector and the envelope glycoprotein expression vector pVSV-G using 15 μl/transfection of Lipofectamine 2000 (Invitrogen) cationic lipid reagent for 3 h in a total volume of 5 ml medium/dish. Subsequently, transfection medium was replaced with 10 ml growth medium, and the cells were incubated for 72 h. Retrovirus-containing supernatant was then harvested and passed through a 0.45 μm cellulose acetate filter, then concentrated by ultracentrifugation at 50,000 × g for 30 min at 4°C. Supernatant was carefully poured off and virus was resuspended in the residue (\~200 μl) and frozen (-70°C) for future use. Cells for transduction were plated on 6-well tissue culture plates (Falcon) at 50% confluency. Concentrated retrovirus (titer unknown) along with polybrene (8 μg/ml) were added to one ml/well cells (in a 6-well plate) and incubated overnight. Transduction medium was replaced with fresh growth medium, and the following day cells were split into appropriate selective medium. BLV was harvested from supernatant of FLK cells, concentrated, and used to transduce cells in a similar fashion. Luciferase assay ---------------- Luciferase assays (Promega) were performed using a single-tube luminometer (Pharmingen) to measure relative light units (RLU) on a linear scale. Cells to be assayed were counted using a hemacytometer, and 1 × 10^6^cells were aliquoted to 1.5 ml microcentrifuge tubes. Then, cells were pelleted at 300 × g for 10 min, washed once with PBS, and lysed with 200 μl reporter lysis buffer (Promega). Lysate was stored at -20°C until assayed. Lysate was thawed and pelleted (300 × g for 10 min), and luciferase was measured with the luminometer using 10 μl lysate/50 μl reagent for 10 s. Linear range was under 1 × 10^7^RLU. Statistical analysis -------------------- Student\'s t-test was performed for statistical evaluation of the results. Results are expressed as the arithmetic mean with the variance of the mean (mean ± SE). Results ======= The BLV promoter was engineered to drive reporter genes ------------------------------------------------------- Our studies utilized a commercially available retroviral system with its standard CMV promoter (CMVp) or replaced with the BLV promoter (BLVp). Figure [1](#F1){ref-type="fig"} shows a schematic of the BLV promoter used in these studies with its unique regulatory elements. The luciferase reporter gene was used to compare promoter expression strength within different cell lines and treatments. The Woodchuck Hepatitis Virus posttranscriptional regulatory element (WPRE) was also incorporated to enhance transgene expression within these retroviral vectors \[[@B19],[@B20]\]. WPRE has been reported to significantly stimulate expression of transgenes in a promoter-independent fashion \[[@B19]\]. Retroviral vectors were used because of the ease of stable cell line establishment, and because of their prominent use in transgenics and gene therapy. The commercially available retroviral vector used in these studies contained the CMV IE promoter for transgene expression. We modified this retrovector for comparison studies replacing the CMV IE promoter with the BLV promoter (see methods). Cells of several different tissues and species were used in our studies. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Schematic representation of BLV promoter used in comparison studies. The BLV promoter (BLVp) consisting of the U3 region of the 5\'LTR of BLV includes the basic elements of transcription start site (+1), CAAT (nt -97/-92) and TATA (nt -43/-37) boxes as shown. Unique to the BLVp are the three imperfectly conserved 21 bp sequences known as the Tax Responsive Elements (TxRE). The numbers following the TxRE designation represent its position relative to the transcription start site. Each TxRE contains a consensus E box-binding motif overlapping an imperfect cyclic AMP responsive element motif (CRE/Ebox). Additionally, the BLVp contains a glucocorticoid responsive element (GRE), Nuclear Factor Kappa Binding motif (NFkB), and B cell specific PU.1 or Spi-B transactivator binding motif (PU.1/Spi-B). The transcription elements are not drawn to scale. :::  ::: The BLV promoter can be as strong as the CMV promoter depending on the host cell -------------------------------------------------------------------------------- In contrast to the constitutive expression of the CMV promoter, the BLV promoter has *cis*elements that are dependent on BLV Tax for transgene expression \[[@B5],[@B21],[@B22]\]. We hypothesized therefore that in a cell line such as D17, the BLV promoter would have little or no activity compared to the CMV promoter. Conversely, in a cell line expressing the BLV Tax transactivator such as the BLV-producing FLK cell line, the BLV promoter would have similar activity compared to the CMV promoter. We tested this assumption with luciferase as the transgene and found indeed, BLV promoter activity was about 50-fold less than CMV promoter activity in D17 cells but was about equal in FLK cells (Fig. [2](#F2){ref-type="fig"}). As shown in Figure [1](#F1){ref-type="fig"}, the BLVp also has a *cis*element that is B cell specific (PU.1/Spi-B). We therefore compared the strengths of BLV and CMV promoters in primary B cells and a BLV infected B cell line hypothesizing that BLVp expression would be comparable to CMVp activity. BLVp activity was still less than CMVp activity in primary B cells but by only about a 5-fold difference (Fig. [2](#F2){ref-type="fig"}). In the BLV infected BL3.1 cell line, BLVp activity was approximately equal to CMVp activity, analogous to results using the BLV infected FLK cell line. Thus, the BLV promoter can be as strong as the CMV promoter within a cell line under specific conditions e.g. BLV infection/Tax expression. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### BLVp and CMVp activity comparison in D17, FLK, primary cow B cells, and BL3.1. Relative light units (RLU) of luciferase activity driven by either the BLV promoter (BLVp) or CMV promoter (CMVp) of 1 × 10^6^stably transduced cells was measured during a 10 s period. Bars represent the arithmetic mean and variance of 10 experiments. \*P \< 0.05; \*\*P \< 0.001 determined by t-test. :::  ::: BLV infection enhances BLV promoter expression but has no effect on the CMV promoter ------------------------------------------------------------------------------------ Since BLV promoter activity was greater than CMV promoter activity in the BLV infected FLK cell line but minimal compared to CMV promoter activity in the non-BLV infected D17 cell line, we set out to determine whether BLV infection of D17 cells would enhance BLVp and/or suppress CMVp expression. The dog derived D17 cell line can be infected with BLV albeit not very efficiently \[[@B14]\]. D17 cells were infected with concentrated BLV from FLK cells, then clonally selected for BLV expression using *pol*RT-PCR and BLV reverse transcriptase assay of the supernatant (data not shown). Luciferase assays demonstrated that BLV promoter activity in infected D17 cells was about 10-fold greater than BLV promoter activity in non-infected D17 cells (Fig. [3](#F3){ref-type="fig"}). CMV promoter activity remained unchanged. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### BLV infection enhances BLVp activity but has no effect on CMVp activity. D17 cells or D17 cells infected with and productively expressing BLV (D17+BLV) were transduced with luciferase expression vectors. Relative light units (RLU) of luciferase activity driven by either the BLV promoter (BLVp) or CMV promoter (CMVp) of 1 × 10^6^stably trasduced cells was measured during a 10 s period. Bars represent the arithmetic mean and variance of 10 experiments. \*\*P \< 0.001 determined by t-test. :::  ::: BLV Tax enhances BLV promoter expression but has no effect on the CMV promoter ------------------------------------------------------------------------------ To assess directly the effect of constitutive Tax expression on the BLV promoter and CMV promoter, BLV Tax was provided as a transgene to cells. As expected, Tax significantly enhanced BLV promoter activity but had no effect on CMV promoter activity (Fig. [4](#F4){ref-type="fig"}) inducing BLVp activity about 48-fold in D17 cells and 4-fold in primary B cells. Interestingly, we found that when BLV infected cells were transduced with the Tax transgene, the resulting increase in BLV promoter activity was a greater-than-additive enhancement of BLV infection and Tax transgene (Table [1](#T1){ref-type="table"}). This effect could likely be caused by Tax expressed from the trangene upregulating expression of the entire BLV provirus, including Tax. The effect on the CMV promoter was not significant. Further, BLVp activity was enhanced in cell lines FLK (2-fold) and BL3.1 (4-fold) actively producing high-levels of BLV (Fig. [5](#F5){ref-type="fig"}). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### BLV Tax expression significantly enhances BLVp activity but has no effect on CMVp activity. D17 cells and primary bovine B cells (D17; B cells), or D17 cells and primary bovine B cells stably transduced with a BLV Tax expression vector (D17+TAX; B cells+TAX), were assayed. Relative light units (RLU) of luciferase activity driven by either the BLV promoter (BLVp) or CMV promoter (CMVp) of 1 × 10^6^stably transduced cells were measured during a 10 s period. Bars represent the arithmetic mean and variance of 10 experiments. \*\*P \< 0.001 determined by t-test. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Percent of Basal Luciferase Expression ::: Promoter D17+Tax D17+BLV D17+Tax+BLV ---------- --------- ----------- ------------- BLVp 115 ± 7 1226 ± 15 2038 ± 202 CMVp 96 ± 5 130 ± 23 118 ± 14 ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Tax trans-gene expression significantly enhances BLVp activity in cells producing high levels of BLV. FLK and BL3.1 cells, or FLK and BL3.1 cells stably transduced with a BLV Tax expression vector (FLK+TAX; BL3.1+TAX) were assayed. Relative light units (RLU) of luciferase activity driven by either the BLV promoter (BLVp) or CMV promoter (CMVp) of 1 × 10^6^stably transduced cells were measured during a 10 s period. Bars represent the arithmetic mean and variance of 10 experiments. \*P \< 0.05; \*\*P \< 0.001 determined by t-test. :::  ::: Trichostatin A non-specifically enhances BLV promoter and CMV promoter Activity ------------------------------------------------------------------------------- The deacetylase inhibitor trichostatin A (TSA) has been shown to be the most efficient activator of BLV expression known to date \[[@B23]\]. To determine whether increased promoter activity due to TSA was a generalized attribute applicable to the CMVp as well, D17, FLK, and BL3.1 cells possessing BLVp or CMVp driven luciferase expression were treated with TSA. However, since CMVp activity was already 50-fold greater than BLVp activity in D17 cells, comparison of TSA induced promoter activities in a cell line where BLV and CMV promoter activities were similar would permit a more effective evaluation of TSA on the two promoters. Further, TSA induced much less death within the 48 h assay period in BL3.1 cells compared to other cell lines tested (\< 10%). Using BL3.1 with counts adjusted for live cells, TSA treatment enhanced both BLVp and CMVp activity by about 40-fold (Fig. [6](#F6){ref-type="fig"}) indicating TSA was a non-specific promoter enhancer. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### Trichostatin A (TSA) enhances BLVp and CMVp activity. Relative light units (RLU) of luciferase activity driven by either the BLV promoter (BLVp) or CMV promoter (CMVp) of 1 × 10^6^stably trasduced cells was measured during a 10 s period. BL3.1 cells were either non-treated or treated with 500 nM TSA for 48 h. Bars represent the arithmetic mean and variance of 10 experiments. \*\*P \< 0.001 determined by t-test. :::  ::: Discussion ========== Viral promoters are used in mammalian expression vectors because they can have strong activity in a wide variety of cells of differing tissues and species. Probably the most employed is the CMV promoter because of its proven high-level constitutive expression in a variety of mammalian cell lines \[[@B24],[@B25]\]. While constitutive transgene expression is suitable for certain research or gene therapy applications, a strong regulated transgene expression is preferable in many other applications \[[@B26]\]. The BLV promoter, consisting of the U3 region of the LTR, is highly dependant upon Tax for activation and transgene expression. In this study, we set out to determine the strength of BLV promoter activity compared to the strength of the CMV promoter to ascertain the utility of the BLV promoter for mammalian expression vectors. Information on the BLV promoter describing the *cis*-acting elements and the dependence upon Tax using reporter vectors in mammalian cell lines has been published \[[@B13],[@B27],[@B28]\]. However, a direct comparison of promoter strength of the BLV promoter and the standard of mammalian expression vectors, the CMV promoter, has not been performed. Several attributes are important in developing a mammalian expression vector. Probably the most important attribute of a mammalian expression vector promoter is its ability to accomplish high-level transcriptional activity in a large variety of cell types of different tissues and species. Our studies showed that the BLV promoter could achieve similar high-level activity to the CMV promoter in cells expressing BLV Tax or infected with BLV. This comparatively high BLV promoter activity was demonstrated in D17 cells which we have found to be the highest expresser of CMV promoter driven transgenes of all cell lines tested in our laboratory. The CMV promoter activity was still about 5-fold greater than BLV promoter activity in the BLV infected D17 cells compared to the relatively equal activity of the CMV promoter versus BLV promoter in BLV infected FLK cells. However, FLK cells contain four copies of the BLV provirus \[[@B29]\] whereas BLV infected D17 cells contain a single copy of the provirus (data not shown). Thus there may be relatively greater expression of Tax in FLK cells effecting greater activity of the BLV promoter. Quantitative levels of Tax in BLV infected D17 or FLK cells were not measured. In this study, we showed relative to the CMV promoter high levels of induced BLV promoter activity in cell lines of canine osteosarcoma (D17), fetal lamb kidney (FLK), bovine B-lymphosarcoma (BL3.1), and bovine primary B cell origin. We also have data (not shown) demonstrating high BLV promoter driven transcriptional levels in cell lines derived from bat lung (TB1Lu), monkey kidney (Vero), and human kidney (HEK-293). Other researchers have also shown high BLV promoter activity using reporter gene assays in cell lines of various tissues derived from cow, dog, cat, mouse, human, monkey, sheep, and hamster \[[@B5],[@B6],[@B13],[@B22],[@B23],[@B27],[@B28]\]. Clearly the BLV promoter possesses the significant trait of high-induced expression in a wide variety of cell types. A second important attribute of an inducible promoter apart from high-induced expression is low basal expression. Researchers have reported barely detectable BLV promoter Tax-independent activity through luciferase assays of COS-1, C8, and KU-1 cell transient transfections \[[@B28]\]. Our results using reporter vector stable D17 cell lines showed low but definite BLV promoter basal activity. Others measuring BLV promoter-driven luciferase activity in transiently transfected D17 cells reported an above background activity of the BLV promoter, but the basal activity seemed much closer to background than we report here \[[@B6],[@B23]\]. The difference could be due to vectors employed (the commercial retrovector we used had weak promoter activity from the 5\'LTR (data not shown) and our vectors contained the WPRE), or that we used stable cell lines versus transient transfections. Researchers using B cell lines (Raji, Daudi, DG75, A20) also showed low, but definite BLV promoter activity in transient and stable transfected cells similar to our results using primary B cells \[[@B6],[@B13],[@B27]\]. Nevertheless, in all of these studies Tax addition was able to induce expression ranging from 50 to 800-fold over basal expression. Our data showed Tax enhanced BLV promoter activity to levels comparable to the CMV promoter. A low but significant BLV promoter Tax-independent activity is not surprising considering the E boxes, CRE, GRE, NFkB and PU.1/Spi-B binding sites are available for cellular transactivating factors (Fig. [1](#F1){ref-type="fig"}). In fact, mutation studies of these *cis*-elements have demonstrated significant decreases in basal level activity, as with the mutation of the GRE site \[[@B30]\], significant increases in basal level activity, as with the mutation of the CRE sites \[[@B6]\], or either decrease or increase in basal activity, depending on the cell line assayed, as with mutations of the E box \[[@B30]\]. Still, compared to CMV promoter activity, or Tax-induced activity, BLV promoter basal activity is very low. A third important attribute for an inducible promoter would be a sensitive modulated response to a specific inducer. Enhancement of the BLV promoter can occur independent of Tax by the addition of activating agents. Phorbol esters, phytohemaglutinin, and lipopolysaccharides have all been shown to enhance BLV promoter expression \[[@B31]\]. However, all of these agents are non-specific activators and upregulate many promoters within the cell \[[@B32]\]. The most efficient activator of BLV expression is the deacetylase inhibitor, trichostatin A (TSA). Addition of TSA to D17 cells enhanced luciferase expression driven by the BLV promoter 11-fold over basal expression \[[@B23]\]. In BL3.1 cells, less variability occurred from TSA induced cell death and basal BLVp and CMVp activity was relatively the same. TSA upregulated activity of both BLV and CMV promoters within BL3.1 by about 40-fold. In contrast, the BLV promoter was specific to Tax activation, while CMV promoter expression was not affected by Tax. For example in D17 cells, Tax specifically increased BLVp activity 48-fold. Nevertheless, the transactivating properties of BLV Tax are not limited to activation of the BLV promoter. Tax has been shown to upregulate Bcl-2 and increase nuclear NFkB activity \[[@B17]\]. Tax expression induces immortalization of primary rat embryo fibroblasts and causes cytokine-independent B cell growth \[[@B17],[@B33]\]. These \"side effects\" of Tax may deter the use of BLV promoter for mammalian expression vectors. However, studies have demonstrated that the BLV promoter transactivation and immortalization activities of wild-type Tax can be dissociated by mutations within specific regions of the protein \[[@B9]\]. In fact, phosphorylation of Tax serines 106 and 293 are required for in vitro cell transformation but not BLV LTR transactivation \[[@B34]\]. Tax transcriptional activity requires an amino-terminal zinc finger and an internal leucine-rich activation domain \[[@B9]\]. Phosphorylation-deficient Tax mutants have been developed \[[@B33]\] and could be used in place of wild-type Tax for BLV promoter transactivation. Other mutations of Tax were shown to enhance BLV promoter activity in 293T cells by 10-fold over wild-type Tax \[[@B22]\]. However this mutant also transactivated the cellular proto-oncogene c-*fos*. Clearly, there is great potential to magnify the desirable traits of the BLV promoter/Tax system for mammalian expression vectors and minimize undesirable traits. Conclusions =========== To determine whether the BLV promoter could be a useful mammalian expression vector element, we compared its activity with the CMV immediate early promoter in dog osteosarcoma (D17), BLV-infected fetal lamb kidney (FLK), BLV-infected bovine B-lymphosarcoma (BL3.1), and primary bovine B-cells. Without concomitant Tax expression from a transgene or BLV infection, the BLV promoter activity was low compared to CMV promoter activity. In the presence of Tax or BLV expression, the BLV promoter activity became equally as active as the CMV promoter. The CMV promoter was not influenced by Tax or BLV. Tax overexpressed as a transgene in BLV infected cells resulted in BLV promoter expression greater than CMV promoter expression. The deacetylase inhibitor, trichostatin A was a potent upregulator of both BLV and CMV promoters. Our results indicate the BLV promoter has great potential use as an inducible promoter for mammalian expression vectors. Competing interests =================== None declared. Authors\' contributions ======================= JSH carried out cell culture work including transfection/transduction and luciferase assays, data preparation and analyses, and drafted the manuscript. KAE performed genetic engineering of vectors. LSK did preliminary work to establish study concepts. RDB and GAS participated in the design and coordination of the study. All authors read and approved the final manuscript. Acknowledgments =============== The authors thank Thomas Hope (University of Illinois-Chicago) for his gift of WPRE, Philip Griebel (VIDO, Saskatoon, Canada) for his gift of J558L Cells, and David Derse (National Cancer Institute) for his gift of pBLV913. This work was supported by NIH grant R44-CA88752.

PubMed Central

2024-06-05T03:55:47.401603

2004-8-24

PMC516021

Introduction ============ Advances in the characterization of cytokines and tumor antigens, coupled with our increasing ability to manipulate gene expression, have fostered a new era of tumor immunotherapy \[[@B1]\]. Interleukin-2 (IL-2) affects a variety of components of the cellular immune system, including B cells and macrophages by inducing the secretion of tumor necrosis factors (TNF) α and β and interferon-γ. Mainly, IL-2 is responsible for the proliferation of T cells. In animal and some human studies, systemic administration of IL-2 has antitumor effects, mediated by cytotoxic effector cells (such as lymphokine-activated killer -- LAK cells and cytotoxic T lymphocytes) \[[@B2]\]. Such systemic administration often induces high toxicity and is shown to be inferior to local continuous production of cytokine, for recruitment of T cells \[[@B3]\]. Cytokine-induced killer cells (CIK) are non-major histocompatibility complex-restricted cytotoxic lymphocytes generated by incubation of peripheral blood lymphocytes with anti-CD3 monoclonal antibody, interleukin (IL)-2, IL-1 and interferon gamma (IFN-γ). CIK represent cells with high antitumor cytotoxicity in vitro and in vivo \[[@B4]\]. CIK cells possess enhanced cytotoxic activity as compared to standard lymphokine activated killer (LAK) cells \[[@B5],[@B6]\]. CIK cells, express CD4 (45.4+/-3.2) % and CD8(47.7+/-11.0%) markers. It has been shown that NKT cells co-expressing CD3 and CD56 markers on their surface represent the major cytotoxic subset of CIK cells \[[@B7]\]. These NKT cells are derived from T cells \[[@B5]\]. Because of the increase in cytotoxicity and high proliferative response, CIK cells have a 73-fold increase in total lytic units per culture as compared to IL-2-stimulated LAK cells. Gene transfers of cytokine genes to CIK and tumor cells have been extensively studied. CIK cells transfected with cytokine genes have shown to induce antitumor effects \[[@B8]\]. Dendritic cells (DC) are specialized antigen-presenting cells located throughout the human body. They represent heterogeneous cell population, residing in most peripheral tissues where they represent 1--2% of the cell numbers. In the absence of ongoing inflammatory and immune responses, dendritic cells constitutively patrol through the blood, peripheral tissues, lymph and secondary lymphoid organs \[[@B9]\]. Morphologically, mature DC are large cells with elongated and stellated processes. They express high levels of MHC I and II, CD11 a, b, c, CD40, CD54, CD58, CD80, CD83, CD86. The most typical markers at present are MHC I, II and co-stimulatory markers such as CD80, CD86, \[[@B10]\] which present signals to CD4 and CD8 positive T cells. Once T cells are activated after interaction with a DC exhibiting the appropriate tumor-associated peptide antigen and class I molecule, they kill other cells that express these molecules such as tumor cells. Exocrine pancreatic carcinomas have a very poor prognosis and a resistance to conventional therapy. This is mainly induced by lack of immuno-competent cells. Therefore, it might be beneficial for patients with pancreatic cancer to induce an immune attack against the tumor by inserting a cytokine gene into the immunological effector cells. The aim of this study was to evaluate the antitumor immune responses of a cytokine immunotherapy using gene transfer to provide continuous and local cytokine production and therefore showing an improved cytotoxic effect against pancreatic cancer cells. Material and methods ==================== Generation of dendritic cells ----------------------------- DC were generated as described before \[[@B4],[@B7]\]. Blood was drawn according to our protocol accepted by the local ethics committee from healthy volunteers. Briefly, peripheral blood lymphocytes were isolated from buffy coats by Ficoll density gradient centrifugation (Lymphoprep, Nycomed, Oslo Norway). These cells were allowed to adhere in six-well-plates at a density of 5 × 10^6^cells/ml for one hour at 37°C in complete RPMI 1640 with 10% heat-inactivated fetal calf serum, 100 U/ml penicillin and 100 μg/ml streptomycin. The non-adherent cells were collected for generating CIK cells. The adherent cells were cultured in 2 ml RPMI 1640 with autologous, heat-inactivated serum, 750 IU GM-CSF and 500 IU IL-4 (Essex Pharma, Nürnberg, Germany), 100 U/ml penicillin and 100 μg/ml streptomycin per well for seven days for generating DC. The media along with the necessary cytokines were changed every third day. Generation of CIK ----------------- CIK cells were generated as described previously \[[@B11]\]. In brief, non-adherent Ficoll separated human peripheral blood mononuclear cells derived from healthy individuals were prepared and grown in RPMI 1640 medium (Gibco BRL, Berlin, Germany), containing 10% fetal calf serum (Gibco BRL), 25 mM Hepes, 100 U/ml penicillin and 100 μg/ml streptomycin. One thousand IU/ml human recombinant interferon γ (Boehringer Mannheim, Germany) was added on day 0. After 24 hrs of incubation, 50 ng/ml of an anti-CD3 (Orthoclone OKT 3, Cilag GmbH, Sulzbach, Germany), 100 U/ml interleukin-1β and 300 U/ml interleukin-2 (R and D Systems, Wiesbaden, Germany) were added. Cells were incubated at 37°C in a humidified atmosphere of 5% CO~2~and sub-cultured every third day in fresh complete medium with 300 U/ml IL-2 at 3 × 10^6^cells/ml. CIK cells were harvested on day +7 and were co-cultured for seven days with autologous DC at a stimulator (DC) to responder (CIK) ratio of 1:5. Cell lines ---------- The human pancreatic carcinoma cell line DAN-G was purchased from DSMZ (Deutsche Sammlung für Zellkultur, Braunschweig, Germany). The cells were maintained in RPMI 1640 supplemented with 10% fetal calf serum (FCS, PAA) 100 U/ml penicillin and 100 μg/ml streptomycin (Seromed, Jülich, Germany) and grown at 37°C in a humidified atmosphere of 5% CO~2~. Preparation of IL-2 plasmid --------------------------- cDNA of human IL-2 was cloned in the plasmid pMTV.05 (Invitrogen, Karlsruhe, Germany). The recombinant pMTv-hIL-2 (referred to as pIL-2) was transformed and the plasmid was eluted using a mini-prep column (Qiagen GmbH, Hilden Germany) according to the manufacturer\'s protocol. Pulsing of DC ------------- DCs were pulsed with tumor lysate of Dan-G cells on day +5 \[[@B12]\]. Gene transfer by nucleofection ------------------------------ CIK cells were subjected to a combination of electrical parameters and specific solution to deliver the DNA directly to the cell nucleus under mild conditions by using a commercially available nucleofection system on day 10 according to manufacturer\'s protocol (Nucleofector Amaxa Biosystems GmbH, Cologne, Germany). Five times 10^6^CIK cells were nucleofected in an electroporation cuvette along with pre-warmed nucleofector solution and 3 μg of pMTV-hIL-2 using the programme U-14. Once nucleofected, CIK cells were transferred into fresh pre-warmed media with the necessary cytokines and serum. IL-2 measurement ---------------- Cell culture supernatants from the nucleofected and non-nucleofected CIK cells were sampled at 24 hrs and 48 hrs, respectively. An enzyme linked immunosorbent assay for IL-2 with matched antibody pairs was performed according to the manufacturer\'s instructions (R and D Systems, Wiesbaden, Germany). Cytotoxicity assay ------------------ A DELFIA EuTDA^®^non-radioactive cytotoxicity assay was used as a fluorometric alternative to the ^51^Cr release assay (Perkin Elmer Wallac Life Sciences, Brussels, Belgium). The assay is based on loading target cells with a fluorescence enhancing ligand. After cytolysis the ligand is released and introduced to the DELFIA^®^Europium solution. The measured signal correlates directly with the amount of lysed cells \[[@B13]\]. Each experiment was performed in triplicates and the mean value was calculated. After incubation, 20-microl aliquots from all wells are transferred to a fresh 96-well plate. To each well of the plate, 180 microl of the Europium solution mix is added and incubated at room temperature for 15 min on a shaker. Fluorescence data are collected using a 96-well plate in a time-resolved fluorometer (PerkinElmer, Brussels, Belgium). Maximum release was obtained by incubating Dan-G cells with 1% lysis Buffer (Perkin Elmer Wallac Life Sciences, Brussels, Belgium). Target cells without effector cells are used as negative control (spontaneous release). Specific releases are calculated as percentage cytotoxicity = experimental release (counts) minus spontaneous release (counts) divided through maximum release (counts) minus spontaneous release (counts) of target cells. Students\'t\' test was applied. A *P*value \< 0.05 was considered significant. Results ======= In vitro generation of DC and CIK --------------------------------- DC were generated from CD14^+^monocytes using GM-CSF and IL-4. Adherent cells showed cytoplasmic processes typical for DC. After co-culture with CIK cells they formed typical cluster. Flow cytometry showed CD14 negative populations, expressing markers typical for DC (CD80^+^, CD83^+^, CD86^+^and HLA-DR expressing cells). The CIK cells were phenotyped with antibodies against CD3, CD8, CD16, CD40L, CD56, HLA-ABC and HLA-DR. Data was similar to other studies by our group (data not shown) \[[@B5],[@B11],[@B14]\]. Transfection efficiency ----------------------- Transfection efficiency was determined by eGFP expression analysis using a fluorescence activated cell sorter. Viable cells were determined by propidium iodide staining. Nucleofection efficiency for eGFP gene transfer into the stimulated CIK cells resulted in a transient expression of 43 +/- 3.8% of the cells after 24 hours. 17% of the cells were not viable after transfection. The amount of IL-2 was the maximum after 24 hrs. An irrelevant plasmid containing eGFP was nucleofected and compared to the plasmid containing the IL-2 insert in various samples (n = 8). Nucleofection resulted in the production of IL-2 with a mean of 478.5 pg /10^6^cells (range of 107.6--1079.3 pg /10^6^cells/24 h) compared to irrelevantly transfected (containing eGFP) CIK cells (31 pg/10^6^cells) (*P*= 0.05). CIK cells secreting IL-2 were co-cultured from days +7 to +14 with DC, 10 days of age. Cytotoxicity of effector cells was analyzed. Co-culture of effector cells with DC led to an increase in cytotoxic activity as measured in a Eu-release assay using Dan-G. Eu-release in co-cultured CIK cells transfected with pIL-2 and DC was 58.5 ± 3.2% at an effector:target ratio of 1:40 (Fig. [1](#F1){ref-type="fig"}) compared to non-transfected CIK cells co-cultured with DC (36.5 ± 5.3%, *P*= 0.03). In order to further enhance cytotoxic activity DC were pulsed with tumor lysate of Dan-G cells on day +5. However, lytic activity (50.3%) was not significantly enhanced (*P*= 0.33) when compared to non-transfected cells (Fig. [1](#F1){ref-type="fig"}). Lytic activity of DCs pulsed with tumor lysate and co-cultured with non-transfected CIK cells was 48.9% where as DC pulsed with tumor lysate and co-cultured with CIK cells transfected without the plasmid was 46.3% and CIK cells alone 40.3% (Fig. [1](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Cytotoxic activity of immunological effector cells that had been co-cultured with DC against Dan-G pancreatic carcinoma cells. Immunological effector cells from a donor were co-cultured from days +10 to +14 with autologous DC cultures seven days of age, as described in materials and methods. DC were pulsed at day +7 with 200 ng/ml of tumor lysate. Cytotoxic activity at various effector to target cell ratios was measured by Europium release assay. Dan-G cells were used as targets. Data represent results of three separate experiments and are shown as mean. CIKs = CIKS cells only DC+CIKS = naive DC co-cultured with CIK cells DC+CIKSpIL-2 = naive DC co-cultured with CIK cells nucleofected with pIL-2 DC-Tu+CIKS = DC pulsed with tumor lysate and co-cultured with CIK cells DC-Tu+CIKSpIL-2 = DC pulsed with tumor lysate and co-cultured with CIK cells nucleofected with pIL-2 DC-Tu+CIKS (nucleofected) = DC pulsed with tumor lysate and co-cultured with CIK cells nucleofected without plasmid :::  ::: Discussion ========== In this report, transfection of CIK cells with IL-2 demonstrated a prominent augmentation of antitumor immunity *in vitro*against pancreatic carcinoma cell lines via secreting significant amounts of IL-2. Ductal pancreatic adenocarcinoma is the fourth leading cause of cancer death in the Western world. Unfortunately, recent advances in diagnostics, staging, and therapy have not resulted in significant improvements. Thus, new approaches are necessary to improve the outcome of patients with exocrine pancreatic cancer. CIK cells are the most potent mediators of the lyses of autologous and allogeneic cancer cells *in vitro*in a non MHC restricted fashion \[[@B15]\], have a higher antitumor toxicity as compared to standard lymphokine activated cells \[[@B5],[@B6],[@B15]\] and may be suitable to remove tumor cells resistant to chemotherapy \[[@B8]\]. Therefore, they are ideal candidates for further enhancing cytotoxic activity. CIK cells have been shown to upregulate DC specific markers \[[@B16]\]. Transgene candidates to potentially activate systemic immune response include genes encoding for co-stimulatory molecules, lymphotactic chemokines, allogeneic MHC molecules, or cytokines like IL-2. Because of the serious toxicity of systemically administered IL-2 observed in clinical practice \[[@B17]\] it can be expected that local expression of IL-2 is less harmful to the patient than systemic administration to trigger the immune system. In this regard, adenoviral-mediated expression of IL-2 cytokine gene in several tumor models has been found to induce strong and specific antitumor responses \[[@B18]\] by stimulating immune cells including T and natural killer cells. But adenoviral transfection may raise safety questions in human gene therapy. Therefore, we were interested in evaluating the potential of IL-2 non-viral transfected CIK cells for their ability to stimulate and activate immunologic effector cells. The use of gene transfected lymphocytes were hampered by a poor efficiency of gene transfer in lymphocytes and a down regulation of cytokine expression \[[@B19]\]. In contrast, nucleofection is a fast and cost-effective method for transfection of large amounts of cells. Here, nucleofection resulted in a significant higher production of IL-2 compared to mock transfected CIK cells (*P*= 0.05). This results matches perfect to a previously reported result by our group \[[@B8]\]. To the best of our knowledge no further report about nucleofection of CIK cells were available. We then showed, that transfection of CIK cells with IL-2 enhances cytotoxic activity (Fig. [1](#F1){ref-type="fig"}) compared to non-transfected CIK cells co-cultured with DC. No significant cytotoxic activity was seen when DC were pulsed with tumor lysate of Dan-G cells were used with IL-2 transfected CIK cells. This may be due to inhibitory factors in the tumor lysate which may contribute to a decrease in lytic activity. This effect is due to increased amounts of CIK cells during co-cultivation of IL-2 secreting CIK cells with DC. IL-2 secretion by the CIK cells enhances the NK cell antitumor activity \[[@B20]\]. NK cells proliferate in the presence of IL-2 \[[@B21]\]. This led to a higher amount of effector cells resulting in a higher cytotoxic activity. This effect of inducing proliferation of tumoricidal lymphocytes is well known and the most important biologic effect of IL-2 on immune cells. The reproducible observation that virtually all malignant cells can be lysed by IL-2 stimulated lymphocytes in a manner directly related to the intensity of IL-2 administration encouraged the pursuit of aggressive, intensive clinical trials, especially in renal cell carcinoma and melanoma. Several authors have shown the efficacy of transfecting primary tumor cells and tumor cell lines with plasmid DNA/lipid complexes \[[@B22]\]. Local production of high concentrations of IL-2 and IFN-alpha at the tumor site was more effective in preventing tumor growth than systemic administration in patients with metastatic renal cell carcinoma \[[@B22]\]. There are several reports introducing IL-2 producing genes into pancreatic cancer, but there are no reports about IL-2 secreting lymphocytes functioning as immune enhancer cells. Therefore, our report is the first describing CIK cells to have enhanced in vitro cytolytic activity via non-viral interleukin-2 gene transfer against pancreatic cancer cell lines. Direct delivery of plasmid IL-2 gene to the established tumors in mice showed an increase in both early and long term survival \[[@B3]\]. Preclinical efficacy studies in a renal cell carcinoma, murine model also showed that direct intra-tumoral administration of an IL-2 plasmid DNA/DMRIE/DOPE complex resulted in the generation of tumor specific lymphocytes and complete tumor regression \[[@B23]\]. It is reasonable too, that these effector cells given in a pancreatic carcinoma model should enhance cytotoxic activity. These investigations are ongoing. Acknowledgments =============== This work was supported by H.W. & J. Hector-Stiftung, Weinheim, Germany.

PubMed Central

2024-06-05T03:55:47.403796

2004-8-25

PMC516022

Pathogenesis of brain abscess ============================= Brain abscesses develop in response to a parenchymal infection with pyogenic bacteria, beginning as a localized area of cerebritis and evolving into a suppurative lesion surrounded by a well-vascularized fibrotic capsule. The leading etiologic agents of brain abscess are the *streptococcal*strains and *S. aureus*, although a myriad of other organisms have also been reported \[[@B1],[@B2]\]. Brain abscess represents a significant medical problem, accounting for one in every 10,000 hospital admissions in the United States, and remains a serious situation despite recent advances made in detection and therapy \[[@B2]\]. In addition, the emergence of multi-drug resistant strains of bacteria has become a confounding factor. Following infection, the potential sequelae of brain abscess include the replacement of the abscessed area with a fibrotic scar, loss of brain tissue by surgical excision, or abscess rupture and death. Indeed, if not detected early, an abscess has the potential to rupture into the ventricular space, a serious complication with an 80% mortality rate \[[@B1]\]. The most common sources of brain abscess are direct or indirect cranial infection arising from the paranasal sinuses, middle ear, and teeth. Other routes include seeding of the brain from distant sites of infection in the body (i.e. endocarditis) or penetrating trauma to the head. Following brain abscess resolution patients may experience long-term complications including seizures, loss of mental acuity, and focal neurological defects that are lesion site-dependent. At the histological level, brain abscess is typified by a sequential series of pathological changes that have been elucidated using the experimental rodent models described in detail below \[[@B3]-[@B7]\]. Staging of brain abscess in humans has been based on findings obtained during CT or MRI scans. The early stage or early cerebritis occurs from days 1--3 and is typified by neutrophil accumulation, tissue necrosis, and edema. Microglial and astrocyte activation is also evident at this stage and persists throughout abscess development. The intermediate, or late cerebritis stage, occurs from days 4--9 and is associated with a predominant macrophage and lymphocyte infiltrate. The final or capsule stage occurs from days 10 onward and is associated with the formation of a well-vascularized abscess wall, in effect sequestering the lesion and protecting the surrounding normal brain parenchyma from additional damage. In addition to limiting the extent of infection, the immune response that is an essential part of abscess formation also destroys surrounding normal brain tissue. This is supported by findings in experimental models where lesion sites are greatly exaggerated compared to the localized nature of bacterial growth, reminiscent of an over-active immune response \[[@B5],[@B8],[@B9]\]. This phenomenon is also observed in human brain abscess, where lesions can encompass a large portion of brain tissue, often spreading well beyond the initial focus of infection. Therefore, controlling the intensity and/or duration of the anti-bacterial immune response in the brain may allow for effective elimination of bacteria while minimizing damage to surrounding brain tissue. The mechanisms elucidated to date in the immunopathogenesis of brain abscess are depicted in Figure [1](#F1){ref-type="fig"}. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Immunopathogenesis of brain abscess. Pyogenic bacteria such as *S. aureus*induce a localized suppurative lesion typified by direct damage to CNS parenchyma and subsequent tissue necrosis. Bacterial recognition by Toll-like receptor 2 (TLR2; **Y**) leads to the activation of resident astrocytes and the elaboration of numerous proinflammatory cytokines and chemokines. Microglia produce a similar array of proinflammatory mediators following bacterial stimulation; however, the receptor(s) responsible for *S. aureus*recognition and subsequent cell activation remain to be identified. Both microglia and astrocytes utilize TLR2 to recognize peptidoglycan (PGN) from the bacterial cell wall. Proinflammatory cytokine release leads to blood-brain barrier (BBB) compromise and the entry of macromolecules such as albumin and IgG into the CNS parenchyma. In addition, cytokines induce the expression of adhesion molecules (ICAM, intercellular adhesion molecule; VCAM, vascular cell adhesion molecule) which facilitate the extravasation of peripheral immune cells such as neutrophils, macrophages, and T cells into the evolving abscess. Newly recruited peripheral immune cells can be activated by both bacteria and cytokines released by activated glia, effectively perpetuating the anti-bacterial immune response that is thought to contribute, in part, to disease pathogenesis. :::  ::: *S. aureus*-induced experimental brain abscess model ==================================================== Although case reports of brain abscess in humans are relatively numerous, studies describing the nature of the ensuing CNS and peripheral immune responses are rare. Therefore, our laboratory has developed a mouse experimental brain abscess model to elucidate the importance of host immune factors in disease pathogenesis \[[@B5],[@B7]-[@B9]\]. Our mouse model was modified based on a previously published model in the rat \[[@B3]\] and utilizes *S. aureus*, one of the main etiologic agents of brain abscess in humans. The mouse brain abscess model accurately reflects the course of disease progression in humans, providing an excellent model system to study immunological pathways influencing abscess pathogenesis and the effects of therapeutic agents on disease outcome. We have successfully utilized this model to characterize inflammatory mediators induced in the brain immediately following *S. aureus*exposure \[[@B5]\] as well as identification of bacterial virulence factors critical for pathogenesis *in vivo*\[[@B8]\]. For example, we have demonstrated that *S. aureus*leads to the immediate and sustained expression of numerous proinflammatory cytokines and chemokines in the brain including tumor necrosis factor-alpha (TNF-α), interleukin-6 (IL-6), IL-1α,β, macrophage inflammatory protein-2 (MIP-2/CXCL2), monocyte chemoattractant protein-1 (MCP-1/CCL2), MIP-1α/CCL3, MIP-1β/CCL4, and regulated upon activation T cell expressed and secreted (RANTES/CCL5) \[[@B5],[@B7]-[@B9]\]. As mentioned earlier, lesion sites in both our experimental model and in human brain abscess are greatly exaggerated compared to the localized nature of bacterial growth, reminiscent of an over-active immune response. To account for the enlarged region of affected tissue involvement associated with brain abscesses compared to the relatively focal nature of the initial insult, we have proposed that proinflammatory mediator production following *S. aureus*infection persists, effectively augmenting damage to surrounding normal brain parenchyma \[[@B10]\]. Specifically, the continued release of proinflammatory mediators by activated glia and infiltrating peripheral immune cells may act through a positive feedback loop to potentiate the subsequent recruitment and activation of newly recruited inflammatory cells and glia. This would effectively perpetuate the anti-bacterial inflammatory response via a vicious pathological circle culminating in extensive collateral damage to normal brain tissue. Recent studies support persistent immune activation associated with experimental brain abscesses with elevated levels of IL-1β, TNF-α, and MIP-2 detected from 14 to 21 days following *S. aureus*exposure \[[@B9]\]. Concomitant with prolonged proinflammatory mediator expression, *S. aureus*infection was found to induce a chronic disruption of the blood-brain barrier, which correlated with the continued presence of peripheral immune cell infiltrates and glial activation \[[@B9]\]. Collectively, these findings suggest that intervention with anti-inflammatory compounds subsequent to sufficient bacterial neutralization may be an effective strategy to minimize damage to surrounding brain parenchyma during the course of brain abscess development, leading to improvements in cognition and neurological outcomes. Besides the potential detrimental roles cytokines may exert on surrounding normal brain parenchyma during the later stages of brain abscess, numerous proinflammatory cytokines such as IL-1β, TNF-α, and IL-6 may have beneficial effects on the establishment of host anti-bacterial immune responses. These cytokines exert numerous functions within CNS tissues including modulation of blood-brain barrier integrity, induction of adhesion molecule expression on cerebral microvascular endothelial cells, and subsequent activation of resident glia and infiltrating peripheral immune cells \[[@B11]-[@B17]\]. We recently examined the relative importance of IL-1, TNF-α and IL-6 in experimental brain abscess using cytokine knockout (KO) mice \[[@B7]\]. The IL-1 KO animals used for these studies were deficient in both IL-1α and IL-1β; therefore, potential caveats arising from redundancy in the activities of these two proteins were avoided. Despite the fact that these cytokines share many overlapping functional activities, IL-1 and TNF-α appear to play an important role in dictating the ensuing anti-bacterial response in brain abscess. This was evident by the finding that bacterial burdens were significantly higher in both IL-1 and TNF-α KOs compared to wild type mice which correlated with enhanced mortality rates in both KO strains \[[@B7]\]. In contrast, IL-6 was not found to be a major contributor to the host anti-bacterial immune response. These studies established important roles for IL-1 and TNF-α during the acute phase of experimental brain abscess development, indicating that these cytokines individually dictate essential functions for the establishment of an effective anti-bacterial response in the CNS parenchyma. Neutrophils are potent bactericidal effector cells and represent the major peripheral cell infiltrate associated with developing brain abscesses \[[@B5],[@B9]\]. Neutrophils exert their bactericidal activity through the production of reactive oxygen and nitrogen intermediates and hydrolytic enzymes that directly destroy bacteria. In addition, neutrophils serve as a source of proinflammatory cytokines, such as TNF-α that serve to amplify the host anti-bacterial immune response \[[@B18],[@B19]\]. However, the continuous release of these products by newly recruited and activated neutrophils can also contribute to tissue damage. Therefore, depending on the context of inflammation, neutrophils can have either beneficial or detrimental effects on the course of infectious diseases. We have recently revealed the functional importance of neutrophils in brain abscess development using antibody-mediated neutrophil depletion and CXCR2 KO mice where neutrophils lack the high-affinity receptor for the neutrophil chemoattractants MIP-2/CXCL2 and KC/CXCL2 \[[@B5]\]. Interestingly, in spite of elevated levels of the CXCR2 ligands MIP-2 and KC, neutrophil extravasation was impaired in CXCR2 KO mice, with cells remaining sequestered within small vessels in developing brain abscesses. Impaired neutrophil influx into evolving brain abscesses in both CXCR2 KO and neutrophil-depleted mice led to exacerbated disease typified by elevated bacterial burdens compared to wild type animals \[[@B5]\]. These studies demonstrate that CXCR2 ligands are the major chemotactic signals required for neutrophil influx into brain abscesses and that their activity cannot be substituted by alternative chemotactic factors such as complement split products (i.e. C3a, C5a), prostaglandins, leukotrienes, or other chemokines. Similar to our findings, the importance of neutrophils in *S. aureus*-induced acute cerebritis was demonstrated by Lo et al. where transient neutrophil depletion resulted in enhanced pathology \[[@B20]\]. In addition to MIP-2 and KC, numerous other chemokines are also detected within evolving brain abscesses including MIP-1α, MIP-1β, MCP-1, and TCA-3/CCL1 \[[@B5],[@B8]\]. The potential roles these chemokines play in the pathogenesis of brain abscess development remain to be defined. However, they could be envisioned to influence the accumulation of monocytes and lymphocytes into the brain and possibly the establishment of adaptive immune responses. Indeed, we and others have demonstrated the influx \[[@B21]\](Kielian, unpublished observations) and generation of *S. aureus*-specific lymphocytes \[[@B9]\] in experimental brain abscess. Staphylococci produce a wide array of virulence determinants that play a role in disease pathogenesis \[[@B22],[@B23]\]. These can be broadly subdivided into surface and extracellular secreted proteins. Surface proteins include structural components of the bacterial cell wall such as lipoteichoic acid and peptidoglycan. Secreted proteins are generally expressed during the exponential phase of bacterial growth and include such proteins as α-toxin, lipase, and enterotoxin. We recently reported that virulence factor production by *S. aureus*is essential for the establishment of brain abscess in the experimental mouse model \[[@B8]\]. Specifically, a requirement for ongoing bacterial replication and/or virulence factor production was supported by the finding that heat-inactivated bacteria were not sufficient to induce proinflammatory cytokine/chemokine expression or abscess formation in the brain. Using a series of *S. aureus*mutants with various defects in virulence factor expression, we identified α-toxin as a critical virulence factor determinant in the experimental brain abscess model. Replication of a *S. aureus*α-toxin mutant was significantly attenuated in the brain, which correlated with a reduction in proinflammatory mediator expression and the failure to establish a well-defined abscess \[[@B8]\]. We proposed that in wild type bacteria, α-toxin, which leads to pore formation in mammalian cell membranes and subsequent osmotic lysis, serves as an effective mechanism to eliminate CNS resident immunocompetent cells (i.e. microglia and astrocytes) as well as professional phagocytes that infiltrate brain abscesses and exert potent anti-bacterial activity (i.e. neutrophils and macrophages). This would effectively impair the efficacy of the ensuing anti-bacterial immune response, allowing bacterial burdens to expand unchecked during the acute phase of disease. In contrast, in the absence of α-toxin secretion, resident glia and infiltrating leukocytes would be capable of rapidly neutralizing bacteria, effectively facilitating the resolution of infection in a timely manner and thus preventing the establishment of a well-formed abscess. However, it is likely that additional virulence factors participate in *S. aureus*infection in the brain since the α-toxin mutant was not completely avirulent. Potential candidates include V8 protease, staphylococcal enterotoxin B, and protein A, the latter of which has been shown to bind to TNF receptor I in the host \[[@B24]\]. Recently, the *S. aureus*-induced experimental brain abscess model has been utilized by Stenzel et al. to demonstrate an important role for astrocytes in dictating the extent of brain abscess pathology \[[@B21]\]. Using glial fibrillary acidic protein (GFAP) KO mice, this group showed that brain abscess pathogenesis was exacerbated in KO animals where lesions were larger and typified by ill-defined borders, severe brain edema, and enhanced levels of vasculitis compared to wild type mice. In addition, GFAP KO mice exhibited a diffuse leukocyte infiltrate that extended into the uninfected contralateral hemisphere. Exacerbation of brain abscess severity in GFAP KO mice was attributed to the absence of a bordering function by astrocytes to contain the infection since strong GFAP immunoreactivity was observed along the abscess margins in wild type animals. It is intriguing that the absence of GFAP influences brain abscess evolution in such a dramatic manner, as astrocytes are still present and functional in these mice. It is possible that GFAP expression in activated astrocytes induces structural changes that influence the local cytoarchitecture leading to bacterial dissemination in brain abscess. Collectively, the studies to date performed in the mouse experimental brain abscess model have begun to elucidate critical mediators in the pathogenesis of disease and host cytokines that play a pivotal role in the generation of the CNS anti-bacterial immune response. However, there are numerous issues that remain to be resolved regarding the role of inflammatory mediators in the evolution of brain abscess. For example, the potential importance of other proinflammatory cytokines and chemokines detected in brain abscess remain to be defined. In addition, factor(s) that participate in the initiation of the anti-bacterial adaptive immune response remain to be elucidated. Evidence to support the establishment of an adaptive immune response is provided by our recent findings that *S. aureus*-specific lymphocytes are formed during the later stages of experimental brain abscess development \[[@B9]\]. It is not known whether the immune response generated during a previous brain abscess episode is capable of providing protection against a second CNS challenge. Another question relates to the potential dual role of various proinflammatory mediators during the course of brain abscess pathogenesis. As mentioned above, a dual role for IL-1 and TNF-α has been suggested by our findings that these cytokines are critical for establishing an effective host anti-bacterial immune response during the acute stage of brain abscess development. However, IL-1 and TNF-α expression persists within brain abscesses for at least 14 to 21 days following infection, suggesting an over-active immune response that is not down-regulated in a timely manner. We are currently using knockout mice to investigate the potential dual role these cytokines may exert during the evolution of brain abscess. Addressing these issues may facilitate the design of effective therapeutic regimens for brain abscess that would be capable of pathogen elimination without the accompanying destruction of surrounding brain parenchyma that normally occurs in disease. Responses of microglia to the brain abscess pathogen *S. aureus* ================================================================ Relevant to our experimental brain abscess model, recent studies from our laboratory have established that both intact *S. aureus*and its cell wall product peptidoglycan (PGN) serve as potent stimuli for proinflammatory mediator production in primary microglia \[[@B5],[@B10],[@B25]\]. Specifically, exposure to both stimuli led to a dose- and time-dependent induction of the proinflammatory cytokines IL-1β, TNF-α, IL-12 p40, and several chemokines including MIP-2, MCP-1, MIP-1α, and MIP-1β. The importance of microglia in the early host response to infection in brain abscess is suggested by the fact that proinflammatory mediator production is detected within 1 to 3 hours following the initial *S. aureus*infection, well before the significant accumulation of peripheral immune cell infiltrates \[[@B4]\]. Another study has also demonstrated that *S. aureus*induces IL-1β expression in neonatal rat microglia \[[@B26]\]. Microglia represent one of the main antigen presenting cells in the CNS \[[@B11],[@B27]\]. To achieve efficient activation of antigen-specific T cells, microglia must express sufficient levels of major histocompatability complex (MHC) class II (signal I) and co-stimulatory molecules such as CD40, CD80, and CD86 (signal II). Recognition of signal I without the concomitant engagement of signal II results in T cell non-responsiveness or anergy. Our group found that both heat-inactivated *S. aureus*and PGN are capable of inducing microglial MHC class II \[[@B10],[@B25]\], CD40, CD80, and CD86 receptor expression, similar to what has been described for microglia in response to the gram-negative bacterial product lipopolysaccharide (LPS) and interferon-γ (IFN-γ) \[[@B27]-[@B31]\]. The ability of *S. aureus*to augment the expression of receptors that are important for antigen presentation suggests that the ability of microglia to present bacterial peptides to antigen-specific T cells may be greatly enhanced following an initial exposure to *S. aureus*. The effects of *S. aureus*and PGN on microglial CD40, CD80, CD86, and MHC class II expression may either be a direct consequence of bacterial stimulation or indirect via the autocrine action of cytokines produced by activated microglia. Microglial activation is a hallmark of brain abscess \[[@B4],[@B5],[@B9]\]. They respond robustly to both *S. aureus*and PGN with significant proinflammatory mediator expression, and many of these same mediators are persistently elevated in brain abscess. Drawing on this relationship, we have proposed that chronic microglial activation may contribute, in part, to the excessive tissue damage characteristic of brain abscess. Therefore, attenuating chronic microglial activation subsequent to effective bacterial elimination in the brain may result in attenuation of the structural and functional damage associated with brain abscess. We have recently examined the efficacy of the cyclopentenone prostaglandin 15d-PGJ~2~to modulate microglial responses to *S. aureus*\[[@B10]\]. 15d-PGJ~2~was found to be a selective and potent inhibitor of *S. aureus*-dependent microglial activation through its ability to significantly attenuate the expression of numerous proinflammatory cytokines and chemokines of the CC family including IL-1β, TNF-α, IL-12 p40, MCP-1, and MIP-1β. In addition, 15d-PGJ~2~also selectively inhibited the *S. aureus*-dependent increase in microglial TLR2, CD14, MHC class II, and CD40 expression whereas it had no effect on the co-stimulatory molecules CD80 and CD86. The ability of 15d-PGJ~2~to modulate the expression of these receptors may serve as a means to regulate microglial and T cell activation during gram-positive bacterial infections in the CNS. Preventing microglial activation by 15d-PGJ~2~or related compounds may help to resolve inflammation earlier, resulting in reductions in brain abscess size and associated damage to surrounding normal brain parenchyma. Receptors utilized by microglia for bacterial recognition ========================================================= As detailed above, our laboratory has established that microglia are capable of recognizing *S. aureus*and respond with robust production of numerous proinflammatory mediators. However, to date, the receptor repertoire responsible for bacterial recognition remains to be defined. In macrophages, numerous receptors have been implicated in bacterial phagocytosis and subsequent activation leading to proinflammatory mediator release including Toll-like receptors (TLR), scavenger receptors, and mannose receptors. The fact that microglia and macrophages share many functional and phenotypical characteristics supports the contention that these receptors may play an important role in microglial responses to bacteria. Toll-like receptors are a family of surface receptors expressed on cells of the innate immune system that allow for the recognition of conserved structural motifs on a wide array of pathogens (referred to as pathogen-associated molecular patterns) \[[@B32],[@B33]\]. To date, eleven TLR have been identified, with TLR2 playing a pivotal role in recognizing structural components of various gram-positive bacteria, fungi, and protozoa \[[@B34]\]. Several groups have reported TLR2 expression in microglia, with receptor expression augmented following inflammatory activation \[[@B25],[@B35]-[@B38]\]. Relevant to brain abscess, we have demonstrated that both *S. aureus*and PGN lead to significant increases in TLR2 mRNA and protein expression, which may enhance microglial sensitivity to bacteria during the course of experimental brain abscess development \[[@B25]\]. Recent studies from our laboratory using primary microglia from TLR2 KO mice have revealed that TLR2 plays a pivotal role in recognition of PGN but not intact *S. aureus*(Kielian, manuscript in preparation). These findings indicate that an alternative receptor(s) is involved in mediating responses to intact bacteria. Candidates include the mannose receptor and members of the scavenger receptor family. Scavenger receptors encompass a broad range of molecules involved in receptor-mediated phagocytosis of select polyanionic acids such as lipoteichoic acid of *S. aureus*\[[@B39]\]. Although adult microglia do not express scavenger receptors in the normal CNS, their expression is induced following inflammation or injury \[[@B40]\]. In the context of brain abscess, a potential tripartite role for microglial scavenger receptors can be envisioned that would include regulating cell adhesion and retention within the inflammatory milieu, facilitating bacterial phagocytosis, and promoting the removal of apoptotic cell debris associated with the evolving abscess \[[@B41]\]. Preliminary data suggest that *S. aureus*and PGN differentially modulate the expression of several distinct scavenger receptors that may influence the nature and extent of phagocytosis (Kielian, unpublished observations). Scavenger receptors have been implicated in β-amyloid phagocytosis by microglia in the context of Alzheimer\'s disease, in part, by the finding that microglia associated with senile plaques express a high degree of scavenger receptor immunoreactivity \[[@B42],[@B43]\]. In addition, scavenger receptors have been implicated in β-amyloid uptake by microglia \[[@B44]-[@B47]\]. The functional importance of scavenger receptors in *S. aureus*phagocytosis by microglia remains to be established. Microglia have been shown to express functional mannose receptors that are responsible for the binding and phagocytosis of mannosylated and fucosylated ligands of bacteria \[[@B48],[@B49]\]. Interestingly, proinflammatory cytokines such as IFN-γ and LPS have been shown to downregulate mannose receptor expression on microglia \[[@B48],[@B49]\]. Using microarray analysis, we also recently demonstrated that mannose receptor levels were significantly attenuated in microglia following *S. aureus*exposure, suggesting that the regulation of mannose receptor expression is conserved among diverse stimuli \[[@B25]\]. Following the subsequent internalization of molecules via the mannose receptor by antigen presenting cells, an immune response can be generated in either a MHC class I, class II, or CD1-restricted manner \[[@B50]-[@B52]\]. In addition, some studies have indicated a functional coupling of the mannose receptor to microbiocidal activities, strongly suggesting a cytotoxic activity linked to mannose receptor-ligand interactions \[[@B53]\]. The functional importance of mannose receptors in the initial recognition and phagocytic events in microglia following *S. aureus*exposure remain to be defined. In addition to the receptors described above, there are additional candidates that may serve as receptors for *S. aureus*phagocytosis in microglia including complement receptor 3 (also known as CD11b/CD18) and CD14, the latter of which we have shown to be expressed on microglia and significantly upregulated following activation with either *S. aureus*or PGN \[[@B10],[@B25]\]. Responses of astrocytes to the brain abscess pathogen *S. aureus* ================================================================= Astrocytes play a pivotal role in the type and extent of CNS inflammatory responses. These cells likely play an important role in the initial recruitment and activation of peripheral immune cells into the CNS during neuroinflammation through the production of several cytokines and chemokines, such as IL-1, IL-6, IL-10, TNF-α, IFN-α/β, granulocyte-macrophage colony-stimulating factor (GM-CSF), macrophage-CSF (M-CSF), granulocyte-CSF (G-CSF), transforming growth factor-beta (TGF-β), RANTES, MCP-1, and IFN-γ-inducible protein-10 (IP-10/CXCL10) \[[@B12],[@B54]\]. Various studies have documented the ability of LPS to induce nitric oxide (NO), cytokine, and chemokine production in astrocytes \[[@B55],[@B56]\]. In contrast, the characterization of products produced by astrocytes following exposure to gram-positive bacteria had remained largely undefined until recently. Studies from our group have revealed that primary astrocytes are capable of recognizing both intact *S. aureus*and PGN and that they respond with vigorous proinflammatory cytokine and chemokine production \[[@B57]\]. Among the factors produced by *S. aureus*-activated astrocytes are NO, TNF-α, IL-1β, MIP-2, MCP-1, MIP-1α, and MIP-1β. These proinflammatory chemokines may serve as signals for neutrophil (MIP-2), monocyte and lymphocyte (MCP-1, MIP-1β) recruitment *in vivo*, whereas IL-1β and TNF-α likely alter blood-brain barrier permeability and induce the expression of critical adhesion molecules on CNS vascular endothelium required for immune cell extravasation into brain abscesses. Receptors utilized by astrocytes for bacterial recognition ========================================================== Astrocytes have recently been shown to express TLR2 \[[@B38],[@B58]\], and although these cells are capable of responding to the well-characterized TLR2 ligand PGN \[[@B58]\], the functional significance of this receptor was not directly demonstrated until recently. Using primary astrocytes from TLR2 KO and wild type mice, our laboratory was the first to report that TLR2 plays a pivotal role in the recognition of *S. aureus*and PGN and in subsequent cytokine and chemokine expression by astrocytes \[[@B57]\]. Interestingly, the production of these cytokines and chemokines was only partially attenuated in TLR2 KO astrocytes, suggesting that alternative receptors are also involved in bacterial recognition. There are numerous candidates for alternative receptors in astrocytes for gram-positive pathogens like *S. aureus*. For example, TLR2 has been shown to form functional heterodimers with TLR1 and/or TLR6 \[[@B59],[@B60]\], thereby increasing its range of antigen detection. It has recently been suggested that CD14 serves as a co-receptor for TLR2 \[[@B61]\] and enhances the recognition efficiency of many TLR2-specific ligands including PGN and lipoteichoic acid \[[@B62]-[@B64]\]. Recently, several studies have reported data that support the involvement of additional, as of yet uncharacterized pattern recognition receptors in bacterial recognition \[[@B61],[@B65]\]. Alternatively, activation through mannose and scavenger receptors that play an important role in the phagocytic uptake of bacteria and have been reported to be expressed by astrocytes \[[@B66]-[@B68]\] may be responsible for the residual proinflammatory mediator expression in TLR2 KO astrocytes. However, to date, the functional importance of these alternative receptors in mediating astrocyte activation in response to *S. aureus*and PGN is currently not known. Although astrocytes have been shown to possess phagocytic activity in response to β-amyloid \[[@B69]\], apoptotic cells \[[@B70]\], and yeast \[[@B71],[@B72]\], the phagocytic potential of astrocytes is still a subject of controversy. Data from our laboratory indicates that primary astrocytes are capable of phagocytosing *S. aureus*\[[@B57]\]. An active phagocytic process is supported by the finding that astrocytes rapidly internalize heat-killed *S. aureus*, indicating that bacterial uptake occurs via a phagocytic pathway and is not simply the result of productive infection by live organisms. Interestingly, TLR2 is not a major receptor for bacterial phagocytosis in astrocytes since both TLR2 KO and wild type astrocytes were equally capable of phagocytosing intact *S. aureus*organisms *in vitro*\[[@B57]\]. The receptor(s) responsible for mediating bacterial uptake in astrocytes are not known but could include the mannose and/or scavenger receptors described above. Studies to identify receptors responsible for *S. aureus*phagocytosis by astrocytes and the optimal conditions required for bacterial uptake are currently ongoing in our laboratory. Issues such as whether bacterial internalization is serum-dependent or requires other bacterial binding proteins must also be addressed. Conclusions and perspectives ============================ The incidence of brain abscess is expected to persist in the human population due to the ubiquitous nature of bacteria coupled with the recent emergence of antibiotic-resistant bacterial strains. Therefore, understanding the roles of both host anti-bacterial immune responses along with bacterial virulence factors may lead to the establishment of novel therapeutic treatments for brain abscess. The mouse *S. aureus*experimental brain abscess model provides an excellent tool for deciphering the importance of various mediators in disease pathogenesis. Especially appealing is the ability to examine the role of specific factors using transgenic and knockout mice because, in our experience, all of the mouse strains examined with this model have qualitatively similar inflammatory profiles following bacterial challenge. In addition, the consequences of *S. aureus*infection do not appear to be influenced by gender, as the responses of female and male mice are similar- another advantage when performing studies with knockout or transgenic mice where animal numbers are often limiting. The responses of microglia and astrocytes to *S. aureus*have been elucidated in terms of proinflammatory mediator expression and in general, have been found to be qualitatively similar to those observed following LPS exposure. Although studies with primary microglia and astrocytes from TLR2 KO mice reveal an important role for this receptor in mediating *S. aureus*-dependent activation, it is clear that additional receptors are also involved in glial responses to this bacterium. This functional redundancy is not surprising because these pathogens have the potential for devastating consequences in a tissue that has limited regenerative capacity such as the CNS. The implications of glial cell activation in the context of brain abscess are likely several-fold. First, parenchymal microglia and astrocytes may be involved in the initial recruitment of professional bactericidal phagocytes into the CNS through their elaboration of chemokines and proinflammatory cytokines. Second, microglia exhibit *S. aureus*bactericidal activity *in vitro*, suggesting that they may also participate in the initial containment of bacterial replication in the CNS. However, their bactericidal activity *in vitro*is not comparable to that of neutrophils or macrophages, suggesting that this activity may not be a major effector mechanism for microglia during acute infection. Third, activated microglia have the potential to influence the type and extent of anti-bacterial adaptive immune responses through their upregulation of MHC class II and co-stimulatory molecule expression. Finally, if glial activation persists in the context of ongoing inflammation, the continued release of proinflammatory mediators could damage surrounding normal brain parenchyma. Indeed, inappropriate glial activation has been implicated in several CNS diseases including multiple sclerosis and its animal model experimental autoimmune encephalomyelitis as well as Alzheimer\'s disease. The continued use of transgenic and knockout mice for *in vivo*studies will facilitate our understanding of immune mechanisms contributing to brain abscess pathogenesis. List of abbreviations ===================== BBB blood-brain barrier CCL CC chemokine ligand CD cluster of differentiation CSF cerebral spinal fluid CXCL CXC chemokine ligand CXCR CXC chemokine receptor GFAP glial fibrillary acidic protein GM-CSF granulocyte-macrophage colony-stimulating factor IFN interferon IL interleukin IP-10 interferon-inducible protein-10 KO knockout LPS lipopolysaccharide M-CSF macrophage colony-stimulating factor MCP monocyte chemoattractant protein MHC major histocompatability complex MIP macrophage inflammatory protein NO nitric oxide PGN peptidoglycan RANTES regulated upon activation T cell expressed and secreted TGF transforming growth factor TNF tumor necrosis factor Competing interests =================== None declared. Acknowledgements ================ I would like to thank Drs. Paul Drew and Nilufer Esen for critical review of the manuscript. This work was supported by grants from the National Institutes of Health NS40730 and MH65297.

PubMed Central

2024-06-05T03:55:47.405777

2004-8-17

PMC516023

Background ========== Human immunodeficiency virus 1 (HIV-1) Vpr is a small virion-associated protein that is incorporated into virions through a specific interaction with the p6 domain of the p55^gag^precursor protein \[[@B1],[@B2]\]. Increasing evidence suggests that Vpr plays important roles during HIV-1 replication and pathogenesis. First, virion-associated Vpr has been shown to act early in viral infection as a facilitator of HIV-1 preintegration complex (PIC) entry through the limiting nuclear pore. This activity of Vpr is thought to be responsible for Vpr\'s ability to enhance HIV-1 replication in nondividing cells, most notably in terminally differentiated macrophages \[[@B3]-[@B5]\]. Second, expression of Vpr induces a G2 cell cycle arrest, which is thought to indirectly enhance viral replication by increasing transcription from the HIV-1 long terminal repeat (LTR) \[[@B6],[@B7]\]. Even though the molecular mechanism of Vpr-mediated cell-cycle G2 arrest is still obscure, it has been known that Vpr expression leads to inactivation of the mitotic p34cdc2/cyclinB complex in human cells \[[@B8],[@B9]\] as well as in fission yeast *Schizosaccharomyces pombe (Sc. Pombe)*\[[@B10]-[@B14]\]. Involvement of protein phosphatase 2A (PP2A), Wee1, Cdc25C, and 14-3-3 proteins has also been implicated \[[@B8]-[@B12],[@B14]\] but the host cell proteins directly engaged by Vpr are not yet identified. Noteworthy, HIV-1 Vpr expression induces also a growth arrest in *Saccharomyces (S.) cerevisiae*\[[@B15]-[@B17]\]. Deletion mapping studies showed that the C-terminal 33 amino acids, including the H(S/F)RIG motif, contributed to this cytostatic effect \[[@B15],[@B18]\]. Although this region has also been implicated in Vpr-mediated cell-cycle dysregulation in mammalian and *S. Pombe*cells \[[@B19]-[@B25]\], the molecular mechanism of Vpr-growth arrest in budding yeast is thought to be distinct since growth arrest occurs independently of any evident block at the G2/M transition \[[@B16]\]. Accordingly, it has been reported that the G2/mitosis transition in budding yeast is regulated differently than in mammalian cells and fission yeast \[[@B26],[@B27]\]. Indeed, Vpr cytostatic effect observed in *S. cerevisiae*has been proposed to result from gross mitochondrial dysfunction \[[@B17]\] and/or cytoskeletal defects \[[@B16]\], rather than a cell cycle G2 arrest. In addition to nuclear import and cytostatic activities, HIV-1 Vpr exhibits cytotoxic properties. Elevated intracellular expression or addition of extracellular Vpr or derived peptides results in proapoptotic effects in human cells including neurons \[[@B6],[@B28],[@B29]\] as well as cytotoxicity in budding and fission yeasts \[[@B30],[@B31]\]. Jacotot *et al*. have provided evidence indicating that extracellular Vpr or peptides derived from Vpr C-terminus induce mitochondrial dysfunction in human cells by a mechanism involving a specific binding to the adenine nucleotide translocator (ANT), a component of the permeability transition pore complex (PTPC) in the mitochondrial membrane. The resulting mitochondrial membrane permeabilization (MMP) leads to a decreased membrane potential and the release of cytochrome c and apoptosis inducing factor (AIF) \[[@B32],[@B33]\]. This Vpr-mediated MMP is thought to initiate cell death through both caspase-dependent and independent mechanisms in human cells as well as cytotoxicity in budding yeast \[[@B32]-[@B37]\]. In addition, it has also been shown that extracellular Vpr is capable of forming cation-selective ion channels in planar lipid bilayers, which can depolarize intact cultured neurons, thus leading to cell death \[[@B28]\]. In a previous report, we have shown that expression of genetically-selected glutathione-S-transferase (GST)-fused di-tryptophan (di-W)-containing peptides inhibited Vpr-mediated growth arrest in *S. cerevisiae*presumably by interacting with Vpr \[[@B38]\]. Interestingly, these, di-W-containing peptides were also able to inhibit Vpr biological activities, including nuclear import, cell cycle G2 arrest and apoptosis, in mammalian cells or HIV-1 infected T cells \[[@B38]\]. Even though the inhibitory effect of these di-W-containing peptides correlated with their ability to interact with Vpr in budding yeast, the detailed mechanism underlying their mode of action remains to be defined. In addition, it is still unclear whether the growth arrest activity of Vpr in budding yeast is related to specific biological activities of Vpr in human cells. In this study, we have performed a mutagenic analysis of Vpr to identify Vpr domains important for di-W peptide binding and cytostatic activity in *S cerevisiae*. Results reveal that the inhibitory di-W-containing peptides target specifically a functional domain of Vpr directly involved in growth arrest in budding yeast. Furthermore, several previously well-characterized Vpr mutants unable to induce cell-cycle dysregulation and/or apoptosis in mammalian cells still exhibit strong growth arrest activity in budding yeast, indeed suggesting that Vpr carries out distinct functions in *S. cerevisiae*. Results ======= Analysis of Vpr sequence and/or structural determinants implicated in the interaction with di-W-containing peptides ------------------------------------------------------------------------------------------------------------------- We have previously used a genetic selection system in *S. cerevisiae*budding yeast and selected a panel of di-W-containing GST-peptides that specifically inhibit Vpr-mediated yeast growth arrest function presumably through their ability to bind HIV-1 Vpr \[[@B38]\]. In this study, we further investigated the molecular mechanism of this inhibition using a newly selected GST-fused di-W peptide WWSFKSV (GST-B4), which displayed an enhanced ability to bind Vpr and inhibit its growth arrest activity in budding yeast (Fig. [1A](#F1){ref-type="fig"} and [1B](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### GST-B4 peptide binds to HIV-1 Vpr in *S. cerevisiae*and rescues cell growth. (A) GST pull-down from yeast extracts. *S. cerevisiae*HP16 strain co-transformed with GST or GST-B4 plasmids and (R+) or (R-) Vpr expressor were metabolically-labeled with 150 μCi of ^35^S-Translabel in Vpr-inducible medium. Half volume of the cell extract was used for GST pull-down, while the remaining lysates were subjected to immunoprecipitation with polyclonal anti-Vpr antiserum. Total and GST-bound radiolabeled Vpr proteins were detected by autoradiography after SDS-PAGE. (B) GST-B4 suppression of Vpr-induced cell growth arrest. Yeast co-transformants were grown in non-inducible selective medium for two days. Similar number of yeast cells were then serially diluted, spotted onto either Vpr non-inducible (Trp^-^/Ura^-^, 2% raf) or Vpr-inducible plates (Trp^-^/Ura^-^, 2% gal) and incubated for 3 to 5 days to evaluate their growth rates. This data is representative of results obtained in two independent experiments :::  ::: Structural studies performed with synthetic forms of Vpr indicate that Vpr is characterized by a well-defined gamma turn (14--16)-alpha helix (α-helix I: 17--33)-turn (34--36), followed by an alpha helix(α-helix II: 40--48)-loop (49--54)-alpha helix (α-helix III: 55--83) domain and ends with a very flexible C-terminal arginine-rich sequence \[[@B39]\]. The α-helical determinants where shown to be required for Vpr virion incorporation, nuclear localization and oligomerization \[[@B39]-[@B44]\] and are believed to be involved in heterologous protein binding \[[@B45]\]. The arginine rich C-terminal region of Vpr has not been shown to have a predicted structure, however this region harbors protein phosphorylation sites and plays an important role in cell cycle G2 arrest and the nuclear localization of the protein in mammalian cells \[[@B6],[@B31],[@B46]\]. To further investigate the sequence and/or structural requirement of Vpr for GST-B4 binding, mutations were introduced in p424Gal1-Vpr expressor to target different regions of Vpr (Fig. [2A](#F2){ref-type="fig"}). The N-terminal Q3R mutant was shown to affect Vpr proapoptotic activity during HIV-1 replication \[[@B47]\]. Four amino acids Glu21, Leu23, Glu25 and Ala30 were separately changed to Lys or Phe (E21K, L23F, E25K and A30F) in order to disrupt the amphipaticity of the first α-helix \[[@B39]\] (Fig. [2A](#F2){ref-type="fig"}). The F34I was introduced in a γ-turn region which is just after the α-helix I \[[@B39]\]. The R62P and I63K mutations introduced in the third helix were aimed at interfering with the integrity of the α-helix and are known to abolish Vpr nuclear localization \[[@B41]\]. Four mutants in the C-terminal region, including, R77Q, S79A, R80A and R87, 88, were generated to replace positively-charged arginine residues or to remove the critical phosphorylation site (Ser 79) of the protein. Vpr mutants S79A and R80A were reported to be defective for cell cycle G2 arrest activity in mammalian cells, while the proapoptotic activity of R77Q was severely attenuated \[[@B6],[@B24],[@B48]\]. In addition, a frameshift mutation (R77fs) \[[@B40]\] and a truncation mutation (R86stop), which prematurely terminate the protein at amino acid 77 and 86 respectively were also constructed. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### HIV-1 Vpr mutants exhibit differential GST-B4 binding abilities. Each Vpr mutant used in this study with the exact location of the introduced mutation is described (A). (B) GST pull-down using a panel of Vpr mutants. Assays were performed as described in Fig. 1A. Protein extracts were prepared from radiolabeled cells expressing GST (lanes 1--2) or GST-B4 proteins (lanes 3--17) alone (R-), or in presence of wild-type Vpr (R+) or different mutant proteins, as indicated. Vpr bound to GST-B4 (upper panel) and the total amount of Vpr as determined using immunoprecipitation with anti-Vpr antiserum (lower panel) were separated by SDS-PAGE and detected after autoradiography. (C) The percentage of GST-B4-bound Vpr relative to the total amounts of Vpr for each mutant was quantified by autoradiography scanning and the level of wild type Vpr bound to GST-B4 was arbitrarily set as 100%. These data are representative of at least two independent experiments. :::  ::: To determine the impact of the Vpr mutations on GST-B4 peptide binding, HP-16 yeast co-transformed with mutated-Vpr expressors and either GST or GST-B4 vectors were radio-labeled in Vpr-inducible medium and subjected to GST pull-down assays (Fig. [2B](#F2){ref-type="fig"}), as described in Materials and Methods. Moreover, the amount of wild type Vpr or each mutant bound to GST-B4 peptide was evaluated by laser densitometric scanning of bands in autoradiograms and normalized to the total amounts of Vpr and GST proteins that were expressed in each transformants. The amounts of wild type Vpr bound to GST-B4 was arbitrarily set as 100% (Fig. [2C](#F2){ref-type="fig"}). Results of figure [2B](#F2){ref-type="fig"} reveal that all Vpr mutants were expressed at comparable levels, as determined by Vpr immunoprecipitation of induced-cell lysates with the exception of Vpr (R77fs), which indeed was previously reported to be less stable than wild type Vpr \[[@B40]\] (Fig. [2B](#F2){ref-type="fig"}, lower panel). While no Vpr interacted with GST (Fig. [2B](#F2){ref-type="fig"}, upper panel, lane 2), similar levels of wild type Vpr, E25K, F34I, I63K, R77fs, R80A, R87, 88 and R86stop mutants were pulled-down with GST-B4 (Fig. [2B](#F2){ref-type="fig"}, upper panel and [2C](#F2){ref-type="fig"}). Similar results were obtained for Vpr mutants Q3R, R77Q, S79A (data not shown). In contrast, E21K, L23F, A30F and R62P mutants, which are respectively located in α-helix I and α-helix III regions, were not co-pulled down with GST-B4 (Fig. [2B](#F2){ref-type="fig"} (upper panel, lanes 5, 7, 8 and 14) and [2C](#F2){ref-type="fig"}). Taken together, these results suggest that the integrity of the N-terminal α-helix I and the α-helix III of Vpr are crucial for GST-B4 binding, whereas the C-terminal domain is dispensable for the interaction. Vpr mutants defective for GST-B4 binding are unable to arrest yeast cell growth ------------------------------------------------------------------------------- We next tested the growth arrest activity of these Vpr mutants in HP-16 yeast. Growing yeast cells transformed with the empty vector (R-), wild-type (R+) or mutated Vpr expressors were serially diluted and spotted onto either a Vpr non-inducible plate (Trp^-^, 2% raf) or a Vpr-inducible plate (Trp^-^, 2% gal). Cell growth was evaluated following an incubation of 3--5 days at 30°C (Fig. [3](#F3){ref-type="fig"}). In Vpr non-inducible plate, all yeast transformants grew at similar rate (Fig. [3](#F3){ref-type="fig"}, left panel). Upon galactose induction, while the empty vector (R-)-transformed yeast grew efficiently (Fig. [3](#F3){ref-type="fig"}, lanes 1, 8 and 15), the wild-type Vpr (lanes 2, 9 and 16), the Q3R mutant and all proteins mutated in the C-terminal region, including R77Q, S79A, R80A exhibited a profound growth arrest activity (Fig. [3](#F3){ref-type="fig"}, right panel (lanes 10,13, 14 and 18). Similar results were obtained for R-87,88 and R86stop mutants (data not shown), indicating that the C-terminal arginine-rich region of Vpr is not involved in budding yeast growth arrest activity. Of note, R77fs showed an impaired growth arrest activity (Fig. [3](#F3){ref-type="fig"}, lane 17), which is most likely due to the shorter half-life of this truncated protein, as reported before \[[@B40]\]. In contrast, expression of helices I and III Vpr mutants, E21K, L23F, A30F and R62P, which displayed a strong attenuation of binding to GST-B4, did not lead to HP-16 budding yeast growth arrest (Fig. [3](#F3){ref-type="fig"}, right panel, lanes 3, 4, 6 and 11). On the contrary, helix I or III mutants E25K, F34I, and I63K, which were able to interact with GST-B4, still exhibited growth arrest activity, even though at reduced levels as compared to wild-type Vpr (Fig. [3](#F3){ref-type="fig"}, right panel, 5, 7 and 12). These results indicate that Vpr helices I and III represent an important functional domain involved in growth arrest in budding yeast. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### The growth arrest activity of different Vpr mutants. *S. cerevisiae*HP16 yeast was transformed with the p424Gal1 expressor alone (R-) or coding for wild-type (R+) or each mutant, as indicated at the left of photograph, and first grown in non-inducible medium for 2 days. Then, similar amounts of transformed yeast were serially 10× diluted and spotted onto either non-inducible (Trp^-^, 2% raf) or Vpr-inducible (Trp^-^, 2% gal) plates and incubated for 3 days to evaluate their growth rates. This data is representative of at least two independent experiments. The ability of each mutant to bind the B4 peptide is indicated on the right. (+) indicates efficient binding while (-) indicates absence of binding. :::  ::: GST-B4 inhibits the cytostatic activity of Vpr mutants and rescues cell growth ------------------------------------------------------------------------------ To further investigate the correlation between the inhibitory effect of GST-B4 and its Vpr-binding ability, GST or GST-B4 were co-expressed with two GST-B4-binding defective Vpr mutants E21K and L23F or with two GST-B4-binding competent mutants E25K and F34I in HP-16 yeast and the resulting cell growth was monitored in Vpr-inducible plates as described above. In agreement with the data of figure [3](#F3){ref-type="fig"}, in the presence of GST alone, mutants E25K and F34I induced significant yeast growth arrest, while such activity was severely impaired for B4-binding defective mutants E21K and L23F (Fig. [4](#F4){ref-type="fig"}, left panel). In contrast, GST-B4 co-expression strongly inhibited the growth arrest activity of the wild type Vpr, E25K and F34I mutants and indeed restored their cell growth at a level comparable to that of yeast cells expressing E21K or L23F (Fig. [4](#F4){ref-type="fig"}, right panel). A weak inhibitory activity of B4 was also observed with mutants E21K and L23F (lanes 3 and 4). It is possible that this may reflect a weak or instable interaction between B4 and Vpr mutants E21K and L23F, which could not be clearly detected in the binding experiments (Fig. [3](#F3){ref-type="fig"}). Overall, these results clearly indicate that GST-B4 specifically binds to structural determinants that are important for inducing cell growth arrest. Moreover, as described previously (38), the binding efficiency of B4 peptides correlates with the extent of the peptide inhibitory activity. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Comparison of GST-B4-mediated inhibition of the growth arrest activity of different Vpr mutants. *S. cerevisiae*HP16 yeast co-expressing GST (left panel) or GST-B4 (right panel) and a panel of representative Vpr mutants, as indicated, were serially 10× diluted and plated on Vpr-inducible and selective (Trp^-^/Ura^-^, 2% gal) plates as described in Fig. 3. The respective cell growth was evaluated after a 3-day incubation. This data is representative for two independent experiments. :::  ::: Discussion ========== We have previously shown that GST-fused di-W-containing peptides were able to interact with HIV-1 Vpr and as a result inhibit its multiple functions in budding yeast as well as in HIV-1 infected T cells \[[@B38]\]. In the present study we have further investigated the sequence and/or structural determinants required for Vpr/peptide interaction and determined their impact on Vpr cytostatic activity in budding yeast. Results clearly show that GST-fused B4 peptide interaction with Vpr involves the α-helical I and III structure of Vpr. Mutations affecting the integrity of these helical regions not only interfered with the interaction with GST-B4 peptide, but also failed to induce a cytostatic activity in budding yeast. Furthermore, Vpr mutants, including Q3R, R77Q, R80A and S79A, yet defective for cell-cycle arrest or apoptosis in mammalian cells, still induced a growth arrest in *S. cerevisiae*and displayed sensitivity to GST-B4 inhibition. Overall, these results indicate that GST-fused di-W-containing peptides directly target functional domains of HIV-1 Vpr responsible for inducing growth arrest in budding yeast and strongly suggest that the mechanism(s) underlying Vpr-induced cytostatic effect in budding yeast are distinct from those implicated in cell-cycle alteration and apoptosis in mammalian cells. Previous reports have indicated that the Vpr cytostatic activity in *S. cerevisiae*budding yeast was attributed to its last 63--96 amino acid (aa) and the critical domain was located in a conserved C-terminal HFRIGCRHSRIG sequence from aa 71 to 82 \[[@B15]\]. In contrast, our results showed that expression of a truncated Vpr encompassing aa 1 to 77 was sufficient to induce growth arrest (Fig. [3](#F3){ref-type="fig"}), suggesting that the sequence of HFRIG (aa 71 to 75), but not HSRIG (aa 78 to 82) and other C-terminal region of Vpr, may constitute one important determinant for this Vpr-induced phenotype. Consistently, a mutagenic analysis by Berglez *et al*., revealed that substitution mutations of aa His71 or Gly75 in this HFRIG sequence abolished Vpr cytostatic effect in budding yeast \[[@B18]\]. Interestingly, our analysis clearly reveal that the N-terminal α-helix I and the α-helix III are both contributing to Vpr cytostatic effect, which is in agreement with a previous finding by Gu *et al*showing that the Vpr F34I mutant was unable to induce a growth arrest phenotype in budding yeast \[[@B16]\]. On the basis of the Vpr NMR structure reported by Wecker *et al*\[[@B39]\], mutations E21K, L23F, E25K and A30F located within the α-helix I (from aa 17 to 33) were designed to disrupt either the negatively-charged cluster or the hydrophobic interface. With the exception of E25K mutant, all other mutations in this N-terminal α-helical region lead to the loss of Vpr cytostatic function (Fig. [3](#F3){ref-type="fig"}). In addition, disruption of the third α-helix by introduction of a proline at position 62 (R62P) suppressed Vpr-induced growth arrest, suggesting that integrity of α-helices I and III was required for Vpr cytostatic activity in budding yeast. It was also noted that E25K and I63K still induced a low level of growth arrest compared to other helical region mutants (Fig. [3](#F3){ref-type="fig"}). It could be possible that E25K is somewhat external to the spatially-aligned acidic cluster D17-E21-E24 \[[@B39]\], and may be therefore less critical. Similarly, the I63K mutation may have a minor impact on the tridimensional structure of helix III as compared to the introduction of a proline residue as with the R62P Vpr mutant. One striking observation of this study is that the four mutants (E21K, L23F, A30F and R62P) located in the α-helical I and III regions of Vpr, which were defective for the cytostatic activity in budding yeast (Fig. [3](#F3){ref-type="fig"}) also lost the ability to interact with the inhibitory GST-B4 peptide (Fig. [2](#F2){ref-type="fig"}). It indicates that GST-B4 directly targets a critical functional domain, possibly a structural cluster comprising both of α-helical I and III, that is responsible for cytostatic activity. Interestingly, the sequence of GST-B4 (GST-WWSKKSV) reveals that, in addition to a conserved di-W motif \[[@B38]\], it also harbors an overlapping WxxF motif, which has been previously isolated by phage-display as a Vpr-binding motif and is present in Vpr-interacting protein uracil DNA glycosylase (UDG) \[[@B49]\]. Coincidentally, a bipartite domain encompassing Vpr amino acids 15--27 and 63--77 was also shown to be involved in UDG binding \[[@B50]\]. Based on these observations, it appears that similar regions of Vpr are involved in binding to UDG and GST-B4 through targeting of a WxxF element. However, E25K and F34I mutants, which were shown to be defective for UDG binding in two-hybrid assays \[[@B21]\], were still able to interact with GST-B4 *in vivo*. Such a difference may specifically rely on the hydrophobic di-W motif, which is not present in UDG \[[@B49]\]. Up to date, how HIV-1 Vpr induces a growth arrest in budding yeast remains an open question. During HIV-1 replication, the expression of Vpr has been shown to induce a cell cycle G2 arrest resulting from the inactivation of the mitotic p34cdc2/cyclinB complex \[[@B51]\]. In contrast, Vpr-mediated growth arrest in budding yeast is thought to occur through a distinct mechanism(s), since it occurs independently of any evident block at the G2/M transition \[[@B16]\]. In this study, we tested a panel of well-characterized Vpr mutants for their ability to growth arrest HP-16 budding yeast. Interestingly, Vpr mutants (S79A and R80A) which were previously shown to be as stable as wild type Vpr but defective for cell-cycle G2 arrest in human cells \[[@B6],[@B19],[@B24]\] still induced strong growth arrest in budding yeast. Conversely, L23F, and R62P mutants, which are competent for cytostatic effect in mammalian cells, \[[@B20],[@B41]\] were unable to block yeast growth. Therefore, it can be concluded that Vpr structural determinants required for growth arrest in *S. cerevisiae*and human cells are clearly distinct, implying that different molecular mechanisms governs Vpr activities in these different cell species. Moreover, our study also demonstrates that Vpr cytostatic effect in budding yeast is not related to the cytotoxic activity of the viral protein. Vpr exhibits different cytotoxic properties that implicate distinct domains of the viral protein. First, wild-type Vpr and its first 40 N-terminal amino acids can form cation-selective ion channels in lipid bilayers \[[@B28],[@B52]\]. Depolarization of the plasma membrane resulting from inward sodium current eventually induces killing of polarized cells such as neurons. On the other hand, apoptosis in T cells is thought to be triggered by transduction of full-length Vpr or its C-terminal 52--96 moiety into cells and involves mitochondrial membrane permeabilization \[[@B33],[@B53],[@B54]\]. Resulting loss of mitochondrial transmembrane potential then induces the release of apoptogenic proteins, leading to caspase-dependent (37,55,48) or caspase-independent \[[@B55]\] cell killing. The fact that both 17--33 and 55--83 alpha-helices are required for growth arrest in *S. cerevisiae*strongly suggests that the cytostatic effect observed in budding yeast is mechanistically distinct from effects resulting from ion channels formation or mitochondria permeabilization. Consistently, Q3R, R80A, R77Q Vpr mutants, which were previously shown to be as stable as wild type Vpr, but yet defective for apoptosis induction in human cells \[[@B6],[@B47],[@B48]\] were still able to block yeast growth in a B4-sensitive way. Conclusions =========== Taken together, the results presented here provide evidence that Vpr triggers growth arrest in budding yeast by an undefined mechanism that is unrelated to Vpr-induced G2 arrest and apoptosis in mammalian cells. This Vpr-induced budding yeast growth arrest can be effectively inhibited by GST-fused di-W peptide through an interaction of di-W peptide with Vpr functional domain, which includes α helix I and III. These observations would support a model in which, Vpr interacts with a di-W-containing protein in *S. cerevisiae*to induce yeast growth arrest. The question that still remains unanswered at this point is whether this Vpr cytostatic activity in budding yeast can also play an important role during HIV-1 replication and viral pathogenesis and further investigations are currently underway to address this question. Materials and methods ===================== Yeast strain ------------ The *S. cerevisiae*yeast strain used in this study was the protease-deficient HP-16 strain (*MAT∝ ura3-52 his3Δ1 leu2 trp1Δ63 prb1-1122 pep4-3 prc1-407*) \[[@B56]\]. Plasmid transformation was performed using the lithium acetate method \[[@B57]\]. Plasmids, antisera and chemicals -------------------------------- The HIV-1 Vpr yeast expression plasmid (p424Gal1-Vpr) and the negative control plasmid p424Gal1-R^-^have been previously described \[[@B38]\]. To generate different p424Gal1-Vpr mutant expression plasmids, each of Vpr mutant cDNAs (Fig. [2A](#F2){ref-type="fig"}) was generated by a two-steps polymerase chain reaction (PCR)-based method \[[@B40]\] by using a 5\'-primer (5\'-CTGCTAGCGGATAGATGGGA-3\') harboring a *BamH*I site in front of the Vpr initiation codon, a 3\'-primer (5\'-GCATCGCTCGAGGATCTACTGGC-3\') containing a *Xho*I site after the stop codon of Vpr and the complementary oligonucleotide primers containing the desired mutations. Amplified Vpr cDNA harboring specific mutations were then cloned into the p424Gal1 vector at *BamH*I/*Xho*I sites. The Vpr mutants L23F, E25K, A30F, R62P, I63K, R77Q, R77fs and R80A were previously described \[[@B6],[@B41],[@B48]\]. The pPGK-GST plasmid was described previously \[[@B38]\] while the pPGK-GST-B4 expressor was isolated and purified from an *S. cerevisiae*HP-16 yeast colony that was resistant to HIV-1 Vpr-mediated growth arrest as previously described \[[@B38]\]. The rabbit anti-Vpr polyclonal serum was raised against bacterially expressed recombinant Vpr as described previously \[[@B58]\]. Galactose, raffinose and glucose were purchased from Sigma Inc. Evaluation of the growth arrest activity of Vpr mutants and the anti-Vpr activity of GST-peptide in budding yeast ----------------------------------------------------------------------------------------------------------------- The experimental procedures to evaluate protein expression, Vpr growth arrest activity and the anti-Vpr activity of GST-fused di-W peptide were described previously \[[@B38]\]. Briefly, HP-16 yeast cells transformed with p424Gal1-wild-type/mutant Vpr plasmids or co-transformed with Vpr expressors and pPGK-GST-B4 plasmid were first grown in a Vpr non-inducible selective medium (Trp^-^or Trp^-^/Ura^-^, 2% raffinose (raf^+^)) for 2 days. Then, suspensions of transformed HP-16 yeast cells (adjusted at similar cell densities) were serially diluted and spotted onto either a Vpr non-inducible plate (Trp^-^or Trp^-^/Ura^-^, 2% raf) or a Vpr-inducible plate (Trp^-^or Trp^-^/Ura^-^, 2% gal) to evaluate the growth of each co-transformed HP16 population. GST pull-down assay and anti-Vpr immunoprecipitation ---------------------------------------------------- HP16 co-transformants were radiolabeled with 150 μCi of ^35^S-Translabel (ICN Inc.) in Vpr-inducible medium and lysed in CHAPS buffer as previously described \[[@B38]\]. Cell extracts were then subjected to GST pull-down assay \[[@B4],[@B38]\]. Briefly, lysates were incubated with glutathione-sepharose 4B beads (Amerham Pharmacia Biotech Inc) for 2 hours at 4°C. Beads were washed 3 times and the radiolabeled protein complexes were eluted with an elution buffer (100mM reduced gluthathione, 120 mM NaCl, 100 mM Tris-HCl pH 8.5) by gentle shaking at 4°C for 1 hour. Eluted protein complexes were separated by SDS-PAGE and detected by autoradiography. For Vpr expression analysis, aliquots of labeled yeast lysates were immunoprecipitated with anti-Vpr antibodies as described previously \[[@B38],[@B40]\]. Authors\' contributions ======================= X-J Y designed the experiments, constructed most Vpr mutants and wrote the manuscript. NR carried out the binding assays and tested the effect of Vpr mutants on yeast cells growth. JL selected and characterized the B4 GST-di-W-containing peptide. GD participated in the design of the study and critically evaluated the manuscript. EAC participated in the design of the study and coordinated it. All authors read and approved the final manuscript. Acknowledgments =============== X-J. Yao is a recipient of a Médecine-Relève 2000-Messenger Foundation Award from the Faculté de Médecine, Université de Montréal. E.A. Cohen is the recipient of the Canada Research Chair in Human Retrovirology. This work was supported by grants from the Canadian Institute of Health Research (CIHR) and the CANVAC network of excellence to EAC.

PubMed Central

2024-06-05T03:55:47.408880

2004-8-16

PMC516024

Background ========== Profile-based methods of sequence analysis use multiple sequence alignments (MSA) to extract information about conserved features of a protein family, which are impossible to decipher from a single sequence. Such methods increase both the sensitivity of homology detection and the quality of produced alignments \[[@B1]-[@B10]\], mainly due to more accurate scoring of similarity between sequence positions. Here, we address the problem connected to but different from the problem of scoring positional matches. We focus on detecting confident dissimilarities between profile positions that are suggested to be equivalent. In particular, we sought conservative P-value estimates for the comparison of individual columns in MSA. Such estimates have at least three practical applications: (i) evaluation and optimization of methods predicting propensities for residue occurrence at protein positions; (ii) detection of potentially misaligned regions in automatically produced alignments and their further refinement; and (iii) detection of sites of functional or structural specificity in two related families. Statistical analysis at the level of individual MSA positions may be used to compare residue frequencies predicted from some model to the actually observed residue usage at the given position in sequence homologs. The model may represent, for example, a method for *in silico*sequence design that generates native-like sequences from a structural template. Detection of discrepancies between the model and the real data would assist the analysis of the model\'s performance and its further improvement. To our knowledge, such statistical assessment has not been proposed up to date. Several approaches have been proposed to detect potential regions of low alignment quality in sequence-sequence and sequence-profile alignments. These approaches range from identifying low-scoring regions in pairwise alignment \[[@B11]\] to more complicated schemes: comparing scores of the given alignment and the optimal alignment where this position is omitted \[[@B12]\], or analyzing the consistency of a given position among different alignments produced with various parameters of alignment construction \[[@B13],[@B14]\]. For multiple sequence alignments, positional residue conservation was proposed as a measure to detect potentially misaligned regions of high variability \[[@B15],[@B16]\]. Cline and co-authors \[[@B17]\] compared several methods for positional evaluation of sequence-profile alignments and recommended the approach based on the analysis of near-optimal alignments \[[@B13],[@B14]\]. However, detection of potentially misaligned regions in profile-profile alignments has not been addressed before. When the analyzed alignment is highly reliable, detecting positions of significant dissimilarity may reveal sites that determine functional or structural specificity of otherwise similar proteins. Several approaches have been proposed that use comparison of multiple sequence alignments in order to predict such sites \[[@B18]-[@B21]\]. However, these methods do not involve explicit estimation of statistical significance. Bejerano \[[@B22]\] has recently proposed a promising algorithmic approach to the exact P-value computation, which allows for a faster enumeration of possible outcomes. Despite a significant improvement in the computational efficiency, the algorithm still requires a considerable time to process realistic data in 20-dimensional space of residue frequencies. In this work, we consider approximate analytical estimates of P-value in two settings: (1) comparison of an alignment column to an emission vector of residue probabilities, and (2) comparison of two alignment columns. These estimates allow detecting cases where the null hypothesis (assumption of similarity) can be confidently rejected. We performed simulation experiments that show consistency of the estimates with the statistical model, and applied our method, PEAC ([P]{.underline}-value [E]{.underline}stimation for [A]{.underline}lignment [C]{.underline}olumns), to the analysis of real MSA. Results ======= Theory ------ As the statistical null model of a multiple alignment column, we assumed independent random draw of residues according to a vector of emission probabilities. We represented randomly generated columns by vectors of residue counts ***n***, with total count *N*equal to that of the real alignment column under evaluation. ### Statistical significance of similarity between a multiple alignment column and a vector of emission frequencies #### Null hypothesis **H**~**0**~^**(1)**^ given alignment column (vector of residue counts ***n*\***) is generated by given vector of emission probabilities ***f***. If this hypothesis is rejected, then the set of emission probabilities is inadequate for the description of the residue content in this alignment column. The assumed null model of random columns corresponds to a multinomial form of *ρ*(***n***\| ***f***), which is difficult for analytical consideration. To calculate the P-value, we use the multivariate Gaussian approximation of the multinomial distribution, based on the assumption of large statistical samples (large total residue counts *N*in the generated columns):  where ***x***= {*x*~*i*~} is a random *d*-dimensional vector of residue counts of size , ***f***is emission vector of residue frequencies,  is the mean vector of residue counts, *Σ*= \|\|cov(*x*~*i*~, *x*~*j*~)\|\| is the covariance matrix. This approximation of p.d.f. allows for the analytical expression for the P-value (Appendix 1 \[see [Additional file 1](#S1){ref-type="supplementary-material"} \]):  where  is a regularized gamma-function, *d*is dimensionality of vector ***f***. Thus, P-value is described by a *χ*^2^distribution with (*d*- 1) degrees of freedom. ##### Random simulation shows consistency of P-value estimates with null model In order to analyze whether the Gaussian approximation allows reasonable P-value estimates, we performed extensive random simulations and tested consistency of P-values based on this approximation (formula (2)) with P-values based on the multinomial model. In particular, we used a set of residue frequencies ***f***= {*f*~*i*~}~1~^20^to generate a large number *Ω*= 10^7^of random columns of a fixed size *N*, i.e. *Ω*sets of *N*residues drawn randomly according to probabilities *f*~*i*~. For each random column, residue counts ***n***= {*n*~*i*~}~1~^20^were derived and the multinomial probability of its generation was calculated as , where  is the multinomial coefficient. All *Ω*generated columns were sorted by *ρ*~mult~in the ascending order. For a given P-value *P*\*, the column number *Ω*P~0~was chosen from this sorted list. This column corresponded approximately to the multinomial P-value *P*\*. This P-value was compared to our estimate *P*~estim~(formula (2)) calculated for the chosen column in the Gaussian approximation of multinomial distribution. For each value *P*\* we performed 10 independent simulations and plotted average values of *P*~estim~against P\*, which showed their general consistency. Figure [1](#F1){ref-type="fig"} illustrates the results for three typical sets of emission frequencies ***f***derived from real alignment columns, and for three typical column sizes *N*. The accuracy of estimates becomes poorer for lower column sizes and more skewed frequency sets (Fig. [1A](#F1){ref-type="fig"}). However, even in such cases the accuracy of *P*~Gauss~within orders of magnitude is sufficient for the purpose of detecting the pronounced dissimilarities with *P*\<\< 0.05. Thus, the error introduced by Gaussian approximation still allows the use of P-value estimates under the initially assumed null model of random columns. ### Statistical significance of similarity between two columns of multiple alignments #### Null hypothesis ***H***~**0**~^**(2)**^ two observed columns ***m*\***and ***n*\***are generated by a single vector of emission probabilities. As the prior distribution of emission vectors, we use the maximum likelihood (ML) estimate based on ***m*\***and ***n*\***. Such prior should produce the conservative upper estimate of the P-value. Rejection of hypothesis *H*~0~^(2)^would mean that the two alignment columns are highly dissimilar. The P-value for this hypothesis is calculated in three steps: a). Given the two vectors of residue counts {***n***, ***m***}, we produce the ML estimate of the p.d.f. for emission vectors ***f***that can generate both columns simultaneously. We assumed a simple form of multivariate Gaussian distribution and calculated ML estimates of its mean  and variance values  (formulae B5). b). We use this p.d.f. *θ*(***f***) as the prior to calculate the posterior probability *ρ*(***n***, ***m*\|***θ*(***f***)) that a pair of random columns {***n***, ***m***} is produced by any single emission vector ***f***. Similarly to problem 1, we use multivariate Gaussian approximation of the multinomial distribution that assumes large total residue counts in the generated columns. The posterior probability density can be calculated as *ρ*(***m,n*** \| *θ*(***f***)) = *∫ρ* (***m,n*** \| ***f***) *θ* (***f***) *d****f***     (3) c). Using (3), we calculate P-value as the integral (Appendix 2 \[see [Additional file 2](#S2){ref-type="supplementary-material"}\]):  This value can serve as the upper estimate of the P-value, since the prior distribution *θ*(***f***) is a ML estimate based on the observed alignment columns. The partial integral *∫ρ* (***m,n*** \| ***f***) *d****m****d****n*** can be calculated analytically for any emission vector ***f***, but analytical calculation of full integral (4) is problematic. However, an approximate estimate of this value would suffice, since (i) expression (4) already contains approximations introduced by estimates of *θ*(***f***), *ρ*(***n***, ***m*\|***θ*(***f***)) and *ρ*(***n*\*, *m*\* \|***θ*(***f***)); and (ii) we are interested in a conservative estimate of the upper P-value limit. Hence, we calculate an approximate upper estimate of P-value (Appendix 2 \[see [Additional file 2](#S2){ref-type="supplementary-material"}\]):  where *erf*(*x*) is error function, and  ##### Random simulation shows consistency of upper P-value estimates with null model To assess the consistency of our estimates with the null model, we performed the following simulation experiments. A random emission vector of residue frequencies ***f***was used to produce a column of size *N*by random draw according to these frequencies. Having the vector residue counts ***n***in this column, we produced another vector of counts ***m***that made our estimated P-value *P*~estim~(***m***,***n***) equal to the specified value *P*~0~. To produce this vector, we considered sets of residue counts as points in multidimensional and randomly chose a straight line passing through the point ***n***. On this randomly directed line, we found the point ***m***as the solution of equation *P*~estim~(***m***,***n***) = *P*~0~, where *P*~estim~(***m***,***n***) is defined by formula (5). Thus, we generated a pair of columns that corresponded to the specified P-value according to the PEAC estimate. We compared this estimate to the actual P-value *P*\* calculated for the generation of ***m***and ***n***by the original vector ***f***. As shown by the plot of *P*\* against *P*~estim~(Fig. [2](#F2){ref-type="fig"}), a particular estimate of P-value may correspond to various actual values *P*\*. However, for low P-values, i.e. for the range of our interest, PEAC systematically produces *P*~estim~higher than actual values *P*\*, as expected from the upper P-value estimates. These conservative estimates ensure the absence of false positive results among detected cases of significant dissimilarity. We developed P-value estimates for the following null hypotheses (see Theory): (1) a given alignment column is generated by a given set of emission residue frequencies; and (2) two given alignment columns are generated by a single set of residue frequencies. We applied both types of estimates to the analysis of real multiple alignments, detecting cases of significant dissimilarity where the null hypotheses were confidently rejected. Application ----------- ### Comparison of an alignment column to a frequency vector Using our method, we assessed the consistency between predictions of residue frequencies based on structural considerations, and the frequencies in multiple alignments of sequence homologs. Specifically, we prepared a dataset of 1695 PDB structures and made predictions of residue propensities at each position, based on local structural environment. In parallel, the sequences corresponding to these structures were used as queries for PSI-BLAST searches, and profiles of detected confident sequence homologs were constructed (see Methods). The effective residue frequencies at profile positions were compared to the structure-based predictions, and P-values for each position were estimated using PEAC. The histogram of produced P-values for all positions is shown in Fig. [3A](#F3){ref-type="fig"}. These P-values ranged widely between 10^-320^and 1.0, with the median being approximately 0.01. To analyze the cases of most pronounced discrepancy between our structure-based predictions and residue frequencies observed among sequence homologs, we chose \~1000 protein positions (0.3% of the whole dataset) that had lowest P-values (*P*\< 10^-100^). These sites were located mainly in the secondary structure elements, most frequently at their ends, and corresponded to unusual local distortions of 3D conformations. We compared residue content in the corresponding subset of alignment columns to the whole dataset. As shown in Fig. [3B](#F3){ref-type="fig"}, alignment positions with low P-values demonstrated unusually high average frequencies of negatively charged residues, glutamate and aspartate. For a more detailed analysis, we considered the subset of 145 alignment positions with *P*\< 10^-100^that contained highly conserved D or E, and inspected corresponding D or E residues in tertiary structures. The vast majority of these residues were buried, as was indicated by the accessible surface area (ASA) of their carboxyl caps (data not shown). When we excluded glutamates and aspartates whose charge could be neutralized by contacts with positively charged arginine, lysine or histidine, the remainig portion of the set was still comprised of mostly buried residues (Fig. [4A](#F4){ref-type="fig"}). These buried residues with acidic side chains did not form salt bridges with basic side chains, which is the most typical way of neutralizing a charge in hydrophobic environment. Inspecting these positions manually, we found a less usual mode of charge neutralization, which involves contacts with other polar residues. A typical example of such conformation is a motif classified in the I-site database \[[@B23],[@B24]\] as aspartate beta bulge, located in the middle of a beta strand in bovine rhodanese (thiosulfate:cyanide sulfurtransferase, PDB ID 1 rhs, Fig. [4B](#F4){ref-type="fig"}). The contact between side chain oxygen of D32 and S34 distorts the regular beta-strand conformation. Our scheme of structure-based frequency prediction considered only most common classes of local conformations that involve nearest neighbor residues. This scheme could not account for less usual residue contacts and therefore failed to predict a high conservation of buried acidic residues at this position, which may have functional or structural importance \[[@B25]\] In summary, this application of our method assists detecting positions with discrepancies between the predicted and naturally occurring residue frequencies. A detailed analysis of these positions may highlight shortcomings of a predicting scheme and suggest possible directions for improvement. ### Comparison of two alignment columns Statistical comparison of two MSA positions may be used in two applications. (i) In automatically produced alignments of sequences or structures, consideration of profiles of confident homologs helps to detect inconsistencies. According to our observations, these inconsistencies are caused mainly by alignment errors. (ii) In the high-quality structure based alignments, where structural equivalence of residues is confident, the low P-values may indicate functional specificity of spatially aligned residues. #### Detection of errors in sequence alignments As an example of application (i), we evaluated our method by ability to predict erroneous residue matches produced by an automatic sequence aligner (ClustalW \[[@B26]\]), as compared to the high-quality reference alignments in a manually curated database, BaliBase \[[@B27]\]. For each BaliBase alignment, we (1) extracted individual sequences and generated their ClustalW alignment; (2) for the top and the bottom sequences of BaliBase alignment, produced MSAs of their homologs detected by PSI-BLAST; and (3) used the resulting alignment pair to estimate P-values for the sequence positions matched by ClustalW. We then sorted all ClustalW positional matches by ascending P-values and classified them as true or false predictions of ClustalW errors. For our purpose, the ClustalW matches different from those in BaliBase were considered true positive predictions; whereas correct matches were considered false positives. Having the ranked list of true and false positive predictions, we generated sensitivity curve (plot of the number of true positives vs. the number of false positives, Fig. [5](#F5){ref-type="fig"}). The curve shows the degree of discrimination between erroneous and correct positional matches. Among the top 1000 predictions, the method generated 151 false positives. Up to \~17,000 true positives, the rate of false positive predictions is slowly growing, then this rate considerably increases. This point approximately corresponds to the P-values of \~10^-2^. #### Detection of evolutionarily unrelated positions in structure-based alignments We applied our method to detect profile dissimilarity between protein positions that are aligned by an automatic structure based method. Specifically, we (1) collected pairs of protein domains that are structurally similar according to the DALI alignments \[[@B28]\] in the FSSP database \[[@B29],[@B30]\], (2) for each of these proteins, produced MSA of homologs detected by PSI-BLAST, and (3) used the resulting pairs of alignments to estimate P-values for the positions matched in FSSP. We used two sets of the FSSP domain pairs, with different sequence identities between the domains: 25 ± 1% (the upper limit of \"twilight zone\", which generally allows for homology detection and alignment construction based on sequences alone \[[@B31],[@B32]\])., and 15 ± 1% (a lower range of identity, where structural alignment is more difficult to reproduce by sequence comparison). Figures [5A,5B](#F5){ref-type="fig"} show the histograms of P-values produced for pairs of profile positions that correspond to structurally aligned residues. The distributions of P-values were different for the two ranges of sequence identities. For identities around 15% (Fig. [6A](#F6){ref-type="fig"}), the histogram had maximum at approximately 0.5 10^-2^and the median was approximately 0.1 10^-2^. For identities around 25% (Fig. [6B](#F6){ref-type="fig"}), the maximum was above 0.1 and the median was approximately 2.5 10^-2^, which shows much better consistency between structure-based and profile-based alignments. To analyze the most dissimilar profile positions, in each dataset we chose 0.3% of position pairs that had lowest P-values (639 pairs for identities 15 ± 1%, and 760 pairs for identities 25 ± 1%). Using Insight II suit for molecular modeling and simulation (Accelrys), we performed a detailed manual analysis of structural superposition for a portion of the corresponding structural alignments. We found that the majority of inspected positions were apparently misaligned. Approximately 80% of these residues were located within 5 positions from a gap introduced in the structural alignment. Vicinities of gaps generally correspond to less similar fragments of aligned structures, which are more difficult to superimpose and where alignment errors can occur more frequently. We considered residue contents of the MSA columns corresponding to these low-P-value position pairs, and compared these contents to the average residue frequencies in the whole MSA datasets. In the set corresponding to 15 ± 1% sequence identity, the most pronounced difference was a higher frequency of aspartate at the position pairs with low P-values (Fig. [6C](#F6){ref-type="fig"}). In the set for 25 ± 1% identity, the low-P-value positions had higher frequencies of methionine, leucine and isoleucine (Fig. [6D](#F6){ref-type="fig"}). We further concentrated on the aligned structural positions that showed unusual residue frequencies in the corresponding MSA columns. In the set corresponding to 15 ± 1% sequence identity, we considered positions with highly conserved aspartate, whereas in the set corresponding to 25 ± 1% sequence identity, we considered positions with high combined frequency of methionine, leucine and isoleucine. In an attempt to exclude apparently misaligned positions, we considered only those positions that were distanced more than 5 residues from gaps in the FSSP structural alignment. We selected and manually analyzed 16 of such positional matches. However, even among these selected matches most of the discrepancies were still caused by apparent alignment errors: 10 cases corresponded to structural misalignments (usually due to a shift in 1 position), and 3 cases were caused by biased residue frequencies at profile positions, due to errors in PSI-BLAST alignments of sequence homologs. The remaining 3 position pairs did not involve apparent errors of either DALI or PSI-BLAST. These pairs might represent real differences in residue preferences at structurally equivalent positions. Figure [7](#F7){ref-type="fig"} shows two examples of low P-values for protein regions that were superimposed by automatic structure aligners. The first example illustrates a typical case of apparent misalignment. The second example represents a case that is observed much rarer among automatic structural alignments: the structure superposition is correct but inconsistent with sequence-based similarity. Such inconsistency might represent a change in the structural role of evolutionary related positions. Human glyoxalase II (PDB ID 1qh5) \[[@B33]\] and bacterial metallo-beta-lactamase L1 (penicillinase, PDB ID 1sml) \[[@B34]\] belong to different families of the SCOP metallo-hydrolase/oxidoreductase superfamily. Although these proteins share only 16% sequence identity, their structures are highly similar (DALI Z-score 16.3). Both glyoxalase II and penicillinase bind two zinc atoms at similar locations. Fig. [7A](#F7){ref-type="fig"} shows a manual structural alignment of their fragments, beta hairpins that contain residues involved in Zn binding (D134 in 1qh5A and S185 in 1sml). These residues have similar orientation of their sidechains (shown in Fig. [7A](#F7){ref-type="fig"}, strand ***b***). In glyoxalase II, D134 binds zinc atoms directly \[[@B33]\], whereas in penicillinase, S185 is linked with zinc through a water molecule \[[@B34]\]. Figures [7B,7C](#F7){ref-type="fig"} and [7D](#F7){ref-type="fig"} show sequence alignments of these regions and corresponding positional P-values based on automated structure comparisons by DALI (Fig. [7B](#F7){ref-type="fig"}) and MAMMOTH \[[@B35]\] (Fig. [7C](#F7){ref-type="fig"}), and on the comparison of sequence profiles (Fig. [7D](#F7){ref-type="fig"}). Alignments of strands ***a***illustrate superposition errors as the typical source of low P-values for automatic structural alignments. DALI (Fig. [7B](#F7){ref-type="fig"}) constructed the correct alignment, which was the same as the manual structure alignment (Fig. [7A](#F7){ref-type="fig"}) and profile-based alignment (Fig. [7D](#F7){ref-type="fig"}). This alignment corresponded to high positional P-values. MAMMOTH (Fig. [7C](#F7){ref-type="fig"}) apparently misaligned strands ***a***by introducing a one-position register shift, which resulted in the low P-values for this region (Fig. [7C](#F7){ref-type="fig"}). Structural alignments of strands ***b***represent a rare example of an automatic alignment that corresponds to low positional P-values and yet is correct from the structural viewpoint. Both DALI and MAMMOTH produced the alignment consistent with the confident manual superposition. This structural superposition correctly aligns zinc-binding residues (D134 in 1qh5A and S185 in 1smlA). However, such alignment corresponds to low positional P-values, indicating a significant difference between structure-based and sequence-based position similarity. The optimal profile-based alignment (Fig. [7D](#F7){ref-type="fig"}) has a one-residue shift that dramatically increases positional P-values in this region, but is inconsistent with the topology of the beta strands and zinc-binding sites. Such a shift might represent a change in the structural roles of related protein positions in remote homologs. Indeed, zinc binding role in penicillinase 1smlA is transferred from residue D184, which is related to the zinc binding D134 of glyoxalase II (1qh5A \[[@B33]\]), to the neighboring S185 \[[@B34]\]. Thus, in the case of a high-quality structural alignment, low positional P-values may indicate evolutionary dissimilarity of spatially superimposed residues. Such cases, however, comprised a minor portion among automatic structure-based alignments and were overwhelmed by the cases of misalignment. #### Prediction of structurally and functionally specific protein positions As an example of possible predictions of functionally specific regions, we considered positions in multiple alignments of sequence homologs for two structurally similar but evolutionary divergent proteins: RNA 2\'-O ribose methyltransferase from *T. thermophilus*\[[@B36]\] (PDB ID 1ipaA) and hypothetical *E. coli*protein Ybea (PDB ID 1ns5A). These proteins possess the same *α*/*β*knot fold but belong to different SCOP families, SpoU-like RNA 2\'-O ribose methyltransferase and Ybea-like, respectively. Using manually curated structure-based alignment of the two proteins and MSAs of their homologs detected by PSI-BLAST, we considered structurally equivalent positions that were well aligned in space (C^*α*^distance less than 2 A, Fig. [8](#F8){ref-type="fig"}) but showed significantly different residue contents in the MSAs (*P*\< 0.01). We found 24 such positions, the majority being concentrated in the region of dimer interface, which includes the \'knotted\' C-terminal helix D (Fig. [8A](#F8){ref-type="fig"}). In RNA 2\'-O ribose methyltransferase, this region is suggested to be crucial for the molecular dimerization \[[@B36]\]. Positions detected in other regions mostly correspond to buried residues of hydrophobic core. The discrepancies in residue content at these positions may reflect different structural solutions for the sidechain packing within the core, as in the case of buried residues in helix C (W225 in 1ipaA *VS*C112 in 1ns5A, Fig. [8](#F8){ref-type="fig"}). Thus, the detected positional differences between SpoU-like and YbeA-like families highlight the functional importance of the \'knotted\' C-terminal helix and may suggest a family-specific mode of dimerization and dimer activity for the hypothetical protein YbeA. Discussion ========== Here, we applied the concept of statistical significance to comparison of single positions of multiple sequence alignments. We proposed rigorous problems of the P-value estimation for the comparison of an alignment column to an emission frequency vector; and for the comparison of two alignment columns. We suggested approximate analytical solutions to these problems and applied the resulting P-value estimates to the analysis of protein families. Comparison of an alignment column to an emission frequency vector ----------------------------------------------------------------- Using our method, we compared residue conservation among sequence homologs and residue propensities predicted from local structural environment. The cases of the highest discrepancy between observed and predicted residue frequencies were enriched with positions containing conserved buried residues D/E. Many of these acidic residues do not form a salt bridge with basic side chains, but use contacts with polar residues to neutralize the negative charge in hydrophobic environment. Surveys of such contacts formed by aspartate residues were previously performed by Singh and Thornton \[[@B37]\] and by Fiser *et al.*\[[@B25]\]. The observed residue conservation may indicate the importance of such motifs for protein structure or function. The structure-based statistic for the prediction of residue propensities used only common classes of structural environments and considered closest neighboring residues in polypeptide chain. Hence this statistic was unable to predict the found conservation of buried glutamate and aspartate. Detection of such contradictions between predicted residue propensities and actual residue frequencies in MSA has three main implications. First, analysis of these contradictions can assist evaluation and further optimization of the predicting schemes, including knowledge-based potentials \[[@B38]-[@B41]\] or environment-specific substitution tables \[[@B42],[@B43]\]. Second, the patterns of atypical relations between residue conservation and structural conformation may point to local motifs of potential structural or functional significance. Third, such atypical patterns, which are unlikely to coincide in two proteins by chance, may serve as signatures for homology detection. Comparison of two alignment columns ----------------------------------- We used our estimates to assess similarity between MSA positions. First, we evaluated our method by detection of erroneous residue matches produced by an automatic sequence aligner, ClustalW \[[@B26]\]. The evaluated automatic alignments were compared to the high-quality reference alignments in a manually curated database, BaliBase \[[@B27]\]. Second, we estimated P-values for MSA positions corresponding to structurally aligned residues in the FSSP database of automatic structure based alignments \[[@B29],[@B30]\]. We found that among detected cases of highest dissimilarity, the vast majority was caused by local structural misalignment. Correction of such alignment errors typically produced an increase of P-values (see results for strand ***a***in Fig. [7C](#F7){ref-type="fig"}*VS*Fig. [7B](#F7){ref-type="fig"}). These results suggest a potential value of the method for the detection of misaligned regions in automatic alignments. In our set of FSSP structural alignments, correctly aligned sites of low P-value were very rare. Such sites correspond to structurally equivalent positions that have different residue content in two related families. To illustrate the detection of such family-specific protein positions, we used a high-quality manually curated structural alignment of distantly related SpoU-like and YbeA-like families of the same *α*/*β*knot fold (Fig. [8](#F8){ref-type="fig"}). In addition to specific preferences for sidechain packing in the hydrophobic core, the statistically significant positional differences emphasized the importance of the \'knotted\' helix (Fig. [8A](#F8){ref-type="fig"}), which is essential for dimer formation \[[@B36]\]. These differences may suggest a family-specific mode of dimerization and dimer activity for the hypothetical protein YbeA. Conclusions =========== We proposed P-value estimates to assess statistical significance for (1) comparison of a single position in a multiple alignment to a set of emission residue frequencies; and (2) comparison of two alignment positions. Computational implementation of these estimates showed its potential value for several important tasks in sequence analysis: (i) evaluation and optimization of methods predicting propensities for residue occurrence at protein positions, such as protocols for *in silico*sequence design; (ii) detection of potentially misaligned regions in automatically produced alignments and their further refinement; and (iii) detection of sites that determine functional or structural specificity in two related families. Methods ======= Calculation of effective residue counts in multiple alignments -------------------------------------------------------------- Effective residue counts at alignment positions were calculated based on the PSIC \[[@B44]\] method. We calculated 21 counts *n*~*eff*~^*PSIC*^for each symbol in the alignment column (including gaps, which are considered the 21^st^symbol), and then applied the following transformation \[[@B16]\]:  Here, *n*~*eff*~corresponds to the number of randomly aligned sequences with the average number of residue types per position equal to *n*~*eff*~^*PSIC*^(for more details, see \[[@B16]\]). Profiles corresponding to fragments of protein structures --------------------------------------------------------- We applied our method to compare structure-based predictions of residue probabilities to the actual residue frequencies observed among sequence homologs. For such a comparison, we produced sequence profiles that correspond to fragments of known 3D structures. Briefly, we used a non-redundant set of structures from PDB (minimum 40 residues long, X-ray structure resolution no more than 2.5, NMR structures excluded, no pairs with sequence identity above 20%). SCOP \[[@B45],[@B46]\]. entries classified as membrane proteins or small proteins enriched with disulfide bonds or metal ions were excluded. The final dataset contained 1695 SCOP domains. Starting from the sequence of each domain, PSI-BLAST searches were performed to 5 iterations over the non-redundant NCBI database, with a conservative E-value cutoff of 10^-5^. In the resulting multiple alignments of detected homologs, we purged sequences whose identity to the query was less than 25%, so that only confident sequence homologs were used for profile construction. We split query sequence into fragments of fixed length *F*. For each fragment, we extracted the corresponding segment of the multiple alignment and removed the sequences with deletions (gaps) in this fragment. For a query of length *L*we produced *L*-*F*+1 sub-alignments and derived effective residue counts as described above. In this work, we used the library of profile fragments of length *F*= 6, which provided accurate results when applied to the prediction of local structural environment from a sequence profile \[[@B47]\]. Prediction of expected residue frequencies from local structural environment ---------------------------------------------------------------------------- The equilibrium frequency of an amino acid at a position in protein structure reflects the energetic fitness of the sidechain in the local structural environment \[[@B38],[@B39]\]. To estimate these frequencies, we employed the scheme similar to those used to derive statistical or knowledge-based potentials \[[@B38]-[@B41]\] or environment-specific substitution tables \[[@B42],[@B43]\]. In brief, we divided structural positions into discrete classes based on local structural environment, and analyzed residue contents for each class of positions in known protein structures. As the characteristics of local structural environment, we used the backbone conformations (*φ*and *ψ*dihedral angles) at the given position and the preceding position, and solvent accessibility of the sidechain at the given position. For a given position we used the partition of Ramachandran plot into 15 (*φ*, *ψ*) classes proposed by Shortle \[[@B39]\]., combined with 3 ranges of relative sidechain solvent accessibility as calculated by the NACCESS package \[[@B48]\] For the position preceding the given, we used a less detailed partition of Ramachandran plot into 6 classes. For each of the resulting 15 × 3 × 6 = 270 classes, we analyzed the set of PDB structures described in the previous section and derived the probabilities of residue types to occur in a class. These probabilities were used as frequency predictions at the structural positions that belong to the class. We assessed consistency of these predictions with residue frequencies in multiple alignments of sequence homologs. Pairs of profiles corresponding to pairs of similar structures -------------------------------------------------------------- As the second application, we estimated statistical significance of similarity between pairs of columns in multiple alignments. Namely, we used pairs of structurally similar proteins (according to the FSSP database \[[@B29],[@B30]\]), produced multiple alignments of their sequence homologs detected by PSI-BLAST, and assessed the consistency between structurally equivalent positions of these multiple alignments. We chose protein pairs with relatively low sequence identities, where detection of similarity between sequences is not straitforward. We focused on two identity ranges: 25 ± 1% (at the upper bound of twilight zone) and a lower range of 15 ± 1%. From each FSSP family, we extracted the parent sequence and all sequences of a significant structural similarity to the parent (Z-score greater than 5.0), with sequence identity to the parent within a given range. We found totally 494 and 1406 sequence pairs with identities 25 ± 1% and 15 ± 1%, respectively. These numbers were reduced by purging symmetric pairs and manual inspection of the remaining domains for the presence of repeats and low-complexity regions. For further analysis, we used 251 sequence pairs with identity 25 ± 1% and 340 pairs with identity 15 ± 1%, each pair representing a unique FSSP family. For each sequence, we ran 5 iterations of PSI-BLAST 2.2.1 against the NCBI nr database (E-value threshold for inclusion in the next iteration 0.005) and obtained multiple alignments of detected homologs. We then applied a procedure of the alignment processing similar to that implemented in PSI-BLAST \[[@B1]\] In particular, only one copy was retained of any rows that were \>97% identical to one another, and the columns with gaps in the first (query) sequence were purged. The resulting multiple alignments were used to calculate P-values for confident structure-based position matches (positions represented as capital letters in FSSP alignments). Calculation of solvent accessibility ------------------------------------ Solvent accessible surface area (ASA) for the residues of interest was determined using NACCESS package \[[@B48]\], which was applied to PDB structures, with heteroatoms excluded. To determine ASA for carboxyl groups of aspartate and glutamate, the sum of ASA for atoms of these groups was calculated. Residue contacts were determined using default settings of NACCESS. Authors\' contributions ======================= RS carried out the theoretical considerations, computational experiments, analysis of the results and drafted the manuscript. NG conceived of the study, and participated in its design and coordination. Both authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 \"P-value for multivariate Gaussian distribution\". ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 \"Upper estimate of P-value for similarity between two alignment columns\". ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We would like to thank Christopher Bystroff for the help with the I-sites database and HMMSTR server. This work was supported in part by the NIH grant GM67165 to NVG. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **P-value estimates for comparison between alignment columns and residue frequency vectors: random simulations show general consistency with null model.**(A-C) For different emission vectors based on real alignment columns, large sets of random columns of three different sizes *N*were generated, and P-value estimates by PEAC (P~estim~) were plotted against experimentally estimated P-values (P\*, see text for details). Sets of emission frequencies are shown as bar graphs in inserts. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Upper estimates of P-value for comparison between two alignment columns: random simulation tests for consistency with null model.**Using a random emission vector, a pair of columns was randomly generated that corresponded to a specified value of the PEAC estimate of (*P*~estim~). P-value for the generation of such a pair by the used emission vector (*P*\*) was plotted against *P*~estim~. See text for details. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Comparison of residue frequencies in multiple alignments of sequence homologs to the structure-based frequency predictions.**A. Histogram of P-values (log-log scale). B. Difference in residue frequencies between 0.3% alignment positions with lowest P-values and the whole population. Frequencies of glutamate and aspartate show the greatest elevation. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Structural environments for protein positions that show discrepancy between observed residue frequencies and their structure-based predictions.**A. Histogram of accessible surface areas (ASA) for D/E residues at the positions that show largest discrepancy (*P*\< 10^-100^), correspond to elevated frequencies of D/E among sequence homologs (effective frequencies more than 0.5), and do not contact with positively charged R, K or H residues. B. An example local structural distortion caused by contacts of buried acidic residues. A fragment of structure (PDB ID 1 rhs) with aspartate beta bulge is shown. Backbone is colored with blue, sidechains with red. Contact formed by aspartate D32 through the side chain oxygen is marked. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Sensitivity plot for the detection of errors in automatically produced multiple sequence alignments.**The set of positional matches in ClustalW alignments was ranked by ascending P-values for corresponding profile positions and classified as true or false predictions. True positive predictions are errors in ClustalW alignments compared to the BaliBase reference. False positive predictions are correst position matches consistent with BaliBase. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Detection of structurally aligned protein positions that correspond to dissimilar profile positions.**(A, B) Histograms of P-values for FSSP alignments between domains of different sequence identity (log-log scale): A. Sequence identity 15 ± 1%. B. Sequence identity 25 ± 1%. (C, D) Difference of residue frequencies at alignment positions with lowest P-values and the whole population: C. Sequence identity 15 ± 1%, *P*\< 10^-14^. D. Sequence identity 25 ± 1%, *P*\< 10^-18^. :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Two sources of discrepancies between structure-based and profile-based position similarities: misalignment and structural or functional specificity.**A. Manual alignment of fragments of human glyoxalase II (PDB ID 1qh5A) and bacterial penicillinase (PDB ID 1sml). Strands ***b***are shifted by one register, so that structurally similar residues correspond to different sequence positions. In these strands, sidechains are shown for the residues involved in Zn binding (D134 and S 185, respectively). (B-D) Alignments produced for these fragments by different methods, with corresponding positional P-values. Secondary structure elements (beta stands ***a***and ***b***) are shown with arrows. Structurally equivalent residues are connected with lines, residues involved in Zn binding are boxed. B. DALI correctly reproduces structural alignment of strands ***a***, resulting in relatively high positional P-values in this region. Alignment of strands ***b***is also structurally correct, but their shift in the two structures results in discrepancy between structural alignment and profile content, as reflected by low P-values. C. MAMMOTH misaligns strands ***a***by one register. Low P-values indicate significant profile dissimilarity between aligned positions in both regions ***a***and ***b***. D. Profile-based alignment is in good accord with profile content, as shown by higher P-values for regions ***a***and ***b***. Note that alignment in region ***b***differs from the structure-based alignment shown in B. :::  ::: ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **Analysis of positional differences in protein families can reveal potential sites of functional specificity.**Comparison of structurally equivalent positions in the multiple alignments of sequence homologs for two distant relatives, 2\'-O ribose methyltransferase from *T. thermophilus*(PDB ID 1ipaA) and hypothetical protein Ybea from *E. coli*(PDB ID 1ns5A). A. The ribbon diagrams of the structurally similar domain fragments (1ipaA, residues 116--263, and 1ns5A, residues 2--146) drawn by BobScript, a modification of the MolScript program \[50\]. Similar secondary structure elements are colored in blue (*α*-helices) and yellow (*β*-strands). The C-terminal helix D, which is involved in dimerization, is highlighted in red. Ball-and-stick models of sidechains are shown for the residues that (i) have C^*α*^distance less than 2A in the superimposed structures of the two domains, and (ii) correspond to significantly different positions in the multiple sequence alignments of the compared families (P \< 0.01). In the ball-and-stick models, C, N, O, and S atoms are shown in gray, blue, red, and yellow, respectively. B. Structure-based sequence alignment of the two domains. The *α*-helices and *β*-strands are displayed as arrows and cylinders, respectively. Highlighted: the protein positions that have significantly different residue content in the two corresponding multiple alignments of sequence homologs detected by PSI-BLAST. Cyan, spatially close protein positions, with C^*α*^distance in the superimposed structures less than 2A; gray, position pairs with higher C^*α*^distances. :::  :::

PubMed Central

2024-06-05T03:55:47.411696

2004-8-5

PMC516025

Background ========== Granulopoiesis, the differentiation of peripheral blood granulocytes, involves dramatic nuclear and cytoplasmic structural changes \[[@B1]\]. Committed bone marrow progenitor cells possess ovoid-shaped nuclei with prominent nucleoli and a paucity of heterochromatin. The mature terminally differentiated human neutrophil (polymorphonuclear granulocyte) exhibits a distinctly lobulated (segmented) nucleus with shrunken nucleoli and extensive peripheral heterochromatin. The mature neutrophil is released into the bloodstream, where it circulates as a round unpolarized cell. Responding to chemotactic agents produced by infection and tissue damage, the circulating neutrophil changes cell shape, converting to a rapidly migrating polarized cell. Mature granulocytes have a limited lifespan, succumbing to apoptosis within a few days following release into the bloodstream. Several established tissue culture cell lines have been investigated as model systems for understanding the events and mechanisms of granulopoiesis \[[@B2]\]. Previous studies from our laboratory have employed the HL-60 cell system to examine nuclear lobulation and cytoskeletal polarization \[[@B3]-[@B5]\]. HL-60 cells exhibit granulocytic differentiation in response to added retinoic acid (RA) \[[@B6]\], eventually undergoing apoptotic death \[[@B7]\]. Our studies on RA induced granulocytic differentiation of HL-60 cells implicated two major factors in the nuclear lobulation process: 1) very low cellular levels of lamins A/C and B1; 2) a significant increase in cellular levels of lamin B receptor (LBR). Ultrastructural studies suggested that during nuclear differentiation, nuclear envelope surface area was increased \[[@B3]\]. In addition to nuclear lobulation, these studies demonstrated the formation of extensive nuclear envelope outgrowths, denoted \"nuclear [e]{.underline}nvelope-[l]{.underline}imited [c]{.underline}hromatin [s]{.underline}heets\" or ELCS. The suspected role of LBR was confirmed in subsequent studies \[[@B8],[@B9]\], which demonstrated that a genetic deficiency of LBR correlates with hypolobulated granulocyte nuclei in the human Pelger-Huet anomaly \[[@B8]\] and the murine Ichthyosis mutation \[[@B9]\]. LBR is an integral membrane protein of the nuclear envelope inner membrane, with putative interactions to lamin B, chromatin and HP1α \[[@B10]\]. The mechanistic relationships between changes in nuclear envelope composition and granulocytic nuclear lobulation are currently unknown. The position of the interphase nucleus within the cell appears to be regulated by microtubules and associated dynein \[[@B11]\] and actin and associated spectrin-like proteins \[[@B12]-[@B15]\]. However, only fragmentary data exists examining the relationship, if any, between cytoskeletal elements and nuclear shape. Absence of intermediate filaments (vimentin) has been correlated with nuclear envelope folds or invaginations \[[@B16]\]. Employing HL-60 cells, evidence has been published that neither microtubules (MTs) nor the actin microfilament system are essential for the establishment of nuclear lobulation \[[@B17]\]. Our laboratory chose to investigate these conclusions concerning HL-60 cells in greater detail. Employing the same cell subline (HL-60/S4), we demonstrated that brief (i.e., 2 and 4 hour) treatments of undifferentiated or granulocytic (RA treated) cells with various cytoskeletal modifying chemicals had no obvious effects upon nuclear shape, although cell shape was strongly affected \[[@B4]\]. However, prolonged (i.e., 2 day) exposure of undifferentiated HL-60/S4 cells to nocodazole (NC, disrupts MTs) or taxol (TX, stabilizes MTs), but not cytochalasin D (CD, disrupts actin microfilaments), resulted in rapid apoptotic death. For this reason, the present study emphasizes the prolonged exposure of cytoskeleton-modifying chemicals on Bcl-2 overexpressing HL-60 cells \[[@B18],[@B19]\]. These cells are more refractory to undergoing apoptosis, but can still be induced with RA to exhibit granulocytic differentiation. Employing HL-60-*bcl*-2 cells \[[@B18]\], we demonstrate that the integrity of the MTs system, but not the actin microfilament system, is essential for the nuclear lobulation process during *in vitro*granulopoiesis. Prolonged exposure of HL-60-*bcl*-2 cells to CD or TX does lead to perturbations of nuclear shape, independently of RA induced nuclear differentiation. Results ======= HL-60/S4 cells undergo retinoic acid induced nuclear lobulation in the presence of cytochalasin D, but do not survive nocodazole or taxol treatment --------------------------------------------------------------------------------------------------------------------------------------------------- HL-60/S4 cells were cultivated for up to four days in medium with (or without) 1 μM RA and with (or without) 1 μM CD. Daily samples of these four cultivation conditions were cytospun on microscope slides and Wright-Giemsa stained for examination. The image data from day 3 are presented (Figure [1a,1b,1c,1d](#F1){ref-type="fig"}). Control cells (Figure [1a](#F1){ref-type="fig"}; no RA or CD) possessed single ovoid nuclei (\>99% of cells) throughout the entire period; CD treatment alone (Figure [1b](#F1){ref-type="fig"}) exhibited primarily multinucleated cells. RA treated controls (Figure [1c](#F1){ref-type="fig"}; no CD) revealed clear nuclear indentation and lobulation by day 3 \[[@B3]\]; RA plus CD treated cells revealed indented and lobulated multiple nuclei (Figure [1d](#F1){ref-type="fig"}). HL-60/S4 cells exposed to 1 μM CD, with or without RA, became mostly multinucleated by day 3 (approximately 4% mono-, 56% bi-, 20% tri-, and 18% tetranucleated). Evidence of apoptosis became apparent by day 4. In addition, immunostaining of RA plus CD treated cells (day 4) with anti-lamin B (data not shown) clearly indicated stained \"patches\" between nuclear lobes, interpreted as ELCS \[[@B3]\]. These results are in complete agreement with a previous publication \[[@B17]\], and monitored HL-60/S4 nuclear differentiation for a longer period. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Wright-Giemsa stained HL-60 cells following exposure to RA and CD.**HL-60/S4 cells (Panels a-d): a, undifferentiated; b, 1 μM CD; c, 1 μM RA; d, 1 μM RA plus 1 μM CD. Three days of drug exposure. HL-60-*bcl*-2 cells (Panels e-h): e, undifferentiated; f, 1 μM CD; g and h, 1 μM RA plus 1 μM CD. Seven days of drug exposure. Scale bar: 10 μm. :::  ::: The previous publication \[[@B17]\] also examined the role of MTs on nuclear lobulation in HL-60/S4 cells. They exposed RA treated cells to 0.1 μM (0.03 μg/ml) NC for up to two days, concluding that the integrity of MTs was not important to the formation granulocytic nuclear lobes. We repeated this experiment, including a separate comparison of HL-60/S4 cells exposed to 0.1 μM TX. The results clearly indicated \~100% apoptotic cell death by day 2 with either NC or TX, preventing any firm conclusion of whether MT integrity is important for nuclear lobulation. We concluded that exploration of this issue would be better accomplished using HL-60-*bcl*-2 cells, which are more refractory to undergoing rapid apoptosis. HL-60-*bcl*-2 cells exhibit nuclear lobulation in response to retinoic acid --------------------------------------------------------------------------- Bcl-2 overexpressing HL-60 cells remain viable for longer periods than the parent cell line, following exposure to RA \[[@B18],[@B19]\]. A significant fraction of RA treated cells survive for up to two weeks; some even for three weeks, exhibiting granulocytic characteristics including: nuclear lobulation, surface antigen expression and nitroblue tetrazolium reduction \[[@B18],[@B19]\]. Figure [2](#F2){ref-type="fig"} presents a morphological analysis of Wright-Giemsa stained HL-60-*bcl*-2 cells for up to three weeks following addition of 1 μM RA. For the sake of analysis, we distinguished four nuclear morphology categories: ovoid (Figure [2a](#F2){ref-type="fig"}), indented (Figure [2b](#F2){ref-type="fig"}), lobulated (Figures [2c](#F2){ref-type="fig"} and [2d](#F2){ref-type="fig"} left) and multilobed (Figures [2d](#F2){ref-type="fig"} right, [2e](#F2){ref-type="fig"} and [2f](#F2){ref-type="fig"}). A multilobed granulocytic nucleus is one that displays five or more lobes, a diagnostic hematological criterion observed in blood smears from humans with megaloblastic anemia \[[@B20]\]. Figure [3](#F3){ref-type="fig"} is a graphical representation of nuclear morphological changes during RA induced granulocytic differentiation. At one week following addition of RA, \~80% of the cells reveal nuclear differentiation, with \~95% viability (trypan blue exclusion). From 12 to 21 days post-addition of RA, small changes in the distribution of nuclear morphologies are apparent. Also during this period, cellular debris from dying cells becomes increasingly evident. These observations are in agreement with detailed viability measurements \[[@B18]\], which indicate a significant loss of viability after \~14 days post addition of RA. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Nuclear differentiation during RA induced granulopoiesis of HL-60-*bcl*-2 cells monitored by Wright-Giemsa staining.**Examples of the four nuclear morphological states: a, ovoid; b, indented; c, lobulated; d, lobulated (left) and multilobed (right); e and f, multilobed. Scale bar: 10 μm. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Time course of nuclear differentiation in RA treated HL-60-*bcl*-2 cells over a three-week period.**The percentage of cells in each of the four nuclear categories from day 0 to day 21 (following addition of RA): ovoid; indented; lobulated; multilobed. :::  ::: An immunoblotting analysis from total cell extracts of RA differentiating HL-60-*bcl*-2 cells is presented in Figure [4](#F4){ref-type="fig"}, focusing upon nuclear envelope components. For a detailed comparison with total cell extracts from differentiating HL-60/S4 cells, see \[[@B5]\]. Based upon densitometric analyses from three separate immunoblotting experiments, LBR showed an increased amount for days 7 to 14 (\~2 to 3-fold, compared to the level at day 0). This increase was seen with both the expected \~70 kD protein band and a \~53 kD band, which may represent a proteolytic product \[[@B4]\]. By contrast, other nuclear proteins (lamin B2, LAP2 α, LAP2 β and emerin) exhibited little-or-no increase in levels, but revealed decreases (especially after day 12). LAP2 β and emerin consistently revealed double bands, which might represent protein modification (e.g., phosphorylation). The low levels of lamins A/C and B1 revealed a relative constancy throughout the period of RA-induced granulocytic differentiation, consistent with earlier observations on HL-60/S4 cells \[[@B5]\]. The loss of proteins on (or after) day 12 may reflect a combination of programmed gene expression and programmed protein degradation. Therefore, experiments on the effects of various cytoskeleton-modifying drugs upon HL-60-*bcl*-2 cells were confined to a 10-day period following the addition of RA. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Immunoblot of total cell extracts of RA treated HL-60-*bcl*-2 cells for up to three-weeks of differentiation.**Nuclear antigens: LBR, lamin B receptor; LMNB2, lamin B2; LMNB1, lamin B1; LMNA/C, lamin A/C; LAP2β; LAP2α; Emerin. The lanes contained comparable amounts of total proteins as judged by Ponceau S staining of the PVDF membrane. Days following addition of RA are indicated. :::  ::: Confocal immunofluorescence studies of HL-60-*bcl*-2 reveal the existence of micronuclei ---------------------------------------------------------------------------------------- Immunofluorescent staining was performed on undifferentiated and RA treated HL-60-*bcl*-2 cells to better document nuclear shape and antigen localization. In the course of the experiments on undifferentiated cells, we discovered that \~10% of the cells exhibited one or more micronuclei, in close proximity to the main nucleus. Furthermore, co-localization experiments revealed that the vast majority of micronuclei contained lamin B, but reduced levels of LBR, in comparison to the adjacent main nucleus. Figure [5](#F5){ref-type="fig"} presents a gallery of anti-LBR and anti-lamin B stained cells with a single main nucleus and adjacent micronuclei. As revealed by TO-PRO-3 staining, the micronuclei also contained DNA. During granulocytic differentiation the round micronuclei continue to be visualized, even as the main nuclei are undergoing indentation and lobulation (Figure [6](#F6){ref-type="fig"}). Again the majority of micronuclei maintain a relative deficiency of LBR, compared to lamin B. The nuclear envelope and patches of ELCS (bright yellow regions) appear to possess significant local concentrations of both LBR and lamin B. Also shown in Figure [6](#F6){ref-type="fig"} (right column, arrow) is a cell containing apoptotic bodies, exhibiting persistence of LBR staining, with reduced lamin B reactivity, in agreement with observations that lamin B is degraded prior to LBR \[[@B21]\]. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Gallery of confocal immunofluorescent images of stained nuclei from undifferentiated HL-60-*bcl*-2 cells.**Cells were selected that demonstrate micronuclei adjacent to a main nucleus: LMNB, lamin B; LBR, lamin B receptor; DNA, TO-PRO-3 stain. Notice that micronuclei frequently exhibit a relative deficiency of LBR compared to the main nuclei, but show comparable amounts of LMNB. Scale bar: 10 μm. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Gallery of confocal images of LMNB and LBR stained nuclei during RA induced granulocytic differentiation.**Note the increased lobulation and the persistence of micronuclei (frequently exhibiting a relative deficiency of LBR). The arrowhead points to an apoptotic cell with apoptotic bodies. Scale bar: 10 μm. :::  ::: In an effort to better define the nature of the nuclei in differentiating HL-60-*bcl*-2 cells, we explored their immunostaining properties (Figure [7](#F7){ref-type="fig"}). Our results indicate that main nuclei and many micronuclei contain centromeres (reactivity with CREST antisera), heterochromatin (reactivity with anti-dimethyl H3K9) and nucleolar staining. Figure [7A](#F7){ref-type="fig"} also indicates that the Golgi apparatus (anti-p58) retains its discrete juxtanuclear form in undifferentiated and granulocytic cell types. In addition, the ER (anti-calreticulin) is present in the spaces between nuclear lobes of granulocytic HL-60-*bcl*-2 cells. A confocal stereo image of immunostained HL-60-*bcl*-2 cells is also presented in Figure [7B](#F7){ref-type="fig"}, demonstrating a micronucleus deficient in LBR and containing multiple centromeres. Micronuclei have not been reported to occur in granulocytes or during granulopoiesis, underscoring that HL-60 cells are clearly abnormal in certain aspects of nuclear and cellular physiology (see \[[@B3]\] for other differences, compared to normal granulocytes). ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Confocal immunofluorescent images of HL-60-*bcl*-2 cells.**A. Gallery of undifferentiated and RA treated (7 days) cells. The columns (left to right) are: anti-centromere (CREST); anti-dimethyl H3K9; anti-nucleolus; anti-Golgi p58; anti-calreticulin. LMNB is always shown in green; the different antigens in red. Scale bar: 10 μm. B. Stereo image of undifferentiated HL-60-*bcl*-2 cells stained with anti-centromere (red), anti-LMNB (blue) and anti-LBR (green). Note the presence of centromeres within a micronucleus, which exhibits a deficiency of LBR. Scale bar: 10 μm. :::  ::: Retinoic acid induced nuclear differentiation of HL-60-*bcl*-2 cells occurs in the presence of cytochalasin D ------------------------------------------------------------------------------------------------------------- HL-60*-bcl-*2 were exposed to 1 μM CD with (or without) 1 μM RA for up to 7 days, cytospun onto microscope slides, methanol-fixed, air-dried and stained with Wright-Giemsa. The results were generally similar to those described above for HL-60/S4 cells. Several examples of Wright-Giemsa stained HL-60*-bcl-*2 cells treated with CD or with RA plus CD are shown in Figure [1e,1f,1g,1h](#F1){ref-type="fig"}. Viability of both the CD-only and the RA plus CD cells were \~60% by day 7. By day 2 following addition of CD with (or without RA), \~70% of the cells were binucleated. A slow progression to tri-, tetra- and \>tetranucleated cells was observed, leading to a population with \~20% mono-, \~50% bi- and \~30% multinucleated cells by day 7. Lobulation and ELCS formation was evident in the cells exposed to RA for 7 days; but nuclei within the CD-only cells also revealed distorted and folded nuclear shapes. This was best observed when cytospun HCHO-fixed cells were immunostained and viewed by confocal microscopy (Figure [8](#F8){ref-type="fig"}). Therefore, it seems reasonable to conclude that CD does not prevent nuclear lobulation and ELCS formation induced by RA. Furthermore, prolonged exposure to CD alone exerts direct (or indirect) effects upon nuclear shape in undifferentiated HL-60*-bcl-*2 cells. ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **Gallery of confocal images of CD treated undifferentiated and RA differentiated HL-60-*bcl*-2 cells.**CD and RA treatment were for 7 days. The columns (left to right) are: anti-nucleolus; anti-Golgi p58; anti-α-tubulin. LBR is always shown in green; the other antigens in red. :::  ::: Nocodazole prevents nuclear lobulation in retinoic acid treated HL-60-*bcl*-2 cells ----------------------------------------------------------------------------------- HL-60*-bcl-*2 cells were exposed to 0, 0.1 or 1.0 μM NC, with 1 μM RA for up to 10 days. Samples were harvested for Wright-Giemsa staining at 1, 2, 3, 4, 7 and 10 days; for immunostaining at day 8. Figure [9](#F9){ref-type="fig"} summarizes the distribution of nuclear profiles in the cytospun, methanol-fixed, air-dried and Wright-Giemsa stained preparations. The progressive disappearance of ovoid and increase in indented and lobulated nuclear forms is clear when NC is not present. On the other hand, the presence of 0.1 or 1.0 μM NC blocked the appearance of any significant numbers of indented or lobulated nuclear forms. Ovoid-shaped nuclei were present in at least 80% of the NC-treated cells. Immunostained preparations of 0.1 μM NC treated cells (HCHO-fixed and not air-dried, to preserve 3-D structure) are presented in Figures [10](#F10){ref-type="fig"} and [11](#F11){ref-type="fig"}. The ovoid nuclei give a slightly wrinkled appearance, with considerable surface infolding; but clearly not lobulations. Short tufts of MTs could be visualized in juxtanuclear positions, consistent with a general depolymerization of MTs. The nuclear envelope exhibited clear staining with anti-lamin B and anti-LBR. Nuclear interiors were stained with anti-centromere, anti-nucleolar and anti-dimethylated H3K9 antisera. Anti-Golgi antibodies revealed some dispersion of the p58 antigen. Dispersal of the Golgi resulting from MT disruption is well documented \[[@B22]\]. In one set of experiments (data not shown), NC was added to HL-60-*bcl*-2 cells at day 2 or day 4 after the addition of RA and nuclear morphology was observed on day 8. NC had a clear inhibitory effect upon RA induced nuclear lobulation, even when added 4 days after RA. Thus the inhibitory effect of NC is not an early event in granulocytic differentiation. As noted earlier, we have observed essentially 100% cell death of HL-60/S4 cells exposed to 0.1 μM NC by day 2. Viability of HL-60*-bcl-*2 cells was much better; but protection was not complete. By day 8 of 0.1 μM NC, viability had dropped to \~23%. This could also be observed as increased levels of cellular debris in the Wright-Giemsa stained slides. The cells shown in Figures [10](#F10){ref-type="fig"} and [11](#F11){ref-type="fig"} were judged to have been viable prior to fixation and staining, based upon the intactness of their nuclear envelopes. Prolonged exposure of cells to NC has clear detrimental effects. But within this limitation, it can be concluded that exposure to NC prevents RA induced nuclear lobulation in HL-60*-bcl-*2 cells. ::: {#F9 .fig} Figure 9 ::: {.caption} ###### **Nuclear differentiation in RA treated HL-60-*bcl*-2 cells exposed to varying concentrations of NC.**Panels: 0, 0.1 and 1.0 μM NC. The percentage of cells in each of the four nuclear categories from day 0 to day 10 (following addition of RA): ovoid; indented; lobulated; multilobed. :::  ::: ::: {#F10 .fig} Figure 10 ::: {.caption} ###### **Stereo confocal images of HL-60-*bcl*-2 cells exposed to 1.0 μM RA and 0.1 μM NC.**Treatment was for 8 days. The rows are: anti-centromere (CREST); anti-nucleolus; anti-α-tubulin. LMNB is always shown in green; the different antigens in red. Scale bar: 10 μm. :::  ::: ::: {#F11 .fig} Figure 11 ::: {.caption} ###### **Confocal images of HL-60-*bcl*-2 cells exposed to 1.0 μM RA and 0.1 μM NC.**Treatment was for 8 days. Left column: top, anti-LBR; middle, anti-dimethyl H3K9; bottom, anti-LMNB. Right column: top, anti-Golgi p58; middle, anti-LMNB; bottom, merge. LMNB is always shown in green. Scale bar: 10 μm. :::  ::: Taxol treatment of HL-60-*bcl*-2 cells results in nuclear lobulation and micronuclei, independently of exposure to retinoic acid -------------------------------------------------------------------------------------------------------------------------------- As described earlier, addition of 0.1 μM TX to HL-60/S4 cells resulted in \~100% cell death by day 2, independently of the presence (or absence) of RA. Apparently the cytotoxicity of TX is potentiated in cells with elevated levels of c-*myc*\[[@B23]\]. HL-60 cells are well known to possess c-*myc*amplification \[[@B24],[@B25]\]. Preliminary studies with HL-60-*bcl*-2 cells indicated that as early as day 2, addition of 0.1 or 1.0 μM TX resulted in the appearance of indented and lobulated nuclei in the absence of RA. Consequently, Wright-Giemsa staining analysis of TX treated cells (in the absence of RA) was performed for up to 10 days (Figure [12](#F12){ref-type="fig"}; identical results were obtained in the presence of RA). There was a rapid decline of cells with ovoid nuclei and a corresponding increase of cells with indented and lobulated nuclear forms. There was also a progressive increase in cell death and debris in the stained preparations. Indeed by day 4, HL-60-*bcl*-2 cells exposed to 0.1 μM TX displayed only \~20% viability. Even so, there were a sufficient number of viable cells to perform immunofluorescent staining. Besides the evident nuclear lobulation, considerable numbers of cells revealed formation of micronuclei. This, combined with the extensive bundling of MTs, resulted in dramatic images of highly perturbed cells (Figure [13](#F13){ref-type="fig"}). There was no obvious spatial relationship between the positions of the MT bundles and the nuclear lobulations and micronuclei. This may signify that the action of TX on cells occurs earlier (e.g., disruption of the mitotic spindle), with later rearrangements during interphase. TX induced micronuclei revealed heterochromatin (anti-dimethylH3K9) and nucleolar materials (Figure [14](#F14){ref-type="fig"}). Clearly, the dramatic effects of TX upon nuclear shape do not depend upon the presence of RA. It remains to be demonstrated whether these effects occur by perturbation of interphase and/or mitotic MTs. ::: {#F12 .fig} Figure 12 ::: {.caption} ###### **Nuclear shape changes in undifferentiated HL-60-*bcl*-2 cells exposed to varying concentrations of TX.**Panels: 0, 0.1 and 1.0 μM TX. The percentage of cells in each of the four nuclear categories from day 0 to day 10: ovoid; indented; lobulated; multilobed. :::  ::: ::: {#F13 .fig} Figure 13 ::: {.caption} ###### **Stereo confocal images of undifferentiated HL-60-*bcl*-2 cells exposed to 1.0 μM TX.**Treatment was for 4 days. Rows: top, control cells, unexposed to TX; middle and bottom, 4 days of TX. LMNB is shown in green; α-tubulin in red. Scale bar: 10 μm. :::  ::: ::: {#F14 .fig} Figure 14 ::: {.caption} ###### **Stereo confocal images of undifferentiated HL-60-*bcl*-2 cells exposed to 1.0 μM TX.**Treatment was for 4 days. Rows: top, anti-nucleolus; middle, anti-dimethyl H3K9; bottom, anti-Golgi p58. LMNB is shown in green; other antigens in red. Scale bar: 10 μm. :::  ::: The centrosomal region retains its proximity to the nucleus in HL-60-*bcl*-2 cells under all conditions ------------------------------------------------------------------------------------------------------- The position of the centrosomal region was visualized with monoclonal antibodies against γ-tubulin (Figure [15](#F15){ref-type="fig"}). In all the conditions tested (undifferentiated; RA treated; RA and CD treated; TX treated), the centrosomal region appeared to be near the nuclear envelope; often surrounded by nuclear lobes in the RA treated cells. The proximity of the centrosomal region to the nucleus in RA treated HL-60-*bcl*-2 cells contrasts with earlier observations on RA treated \"polarized\" HL-60/S4 cells \[[@B4]\]. Light microscope observations of living cells revealed fewer polarized HL-60-*bcl*-2 cells, than previously observed for HL-60/S4 cells (data not shown). It is possible that the HL-60/S4 cells are more readily \"activated\", like normal neutrophils, than are HL-60-*bcl*-2 cells. They also differentiate to granulocytic form faster (\~3 to 4 days), compared to HL-60-*bcl*-2 cells (\~7 days). ::: {#F15 .fig} Figure 15 ::: {.caption} ###### **Confocal images of HL-60-*bcl*-2 cells immunostained for γ-tubulin.**Cell treatments: 0, undifferentiated; RA, 7 days; RA + NC, 7 days with 1 μM RA and 0.1 μM NC; TX, 3 days with 1 μM TX. γ-tubulin is shown in red; LBR in green. Scale bar: 10 μm. :::  ::: Discussion ========== HL-60-*bcl*-2 cells and granulocytic nuclear differentiation ------------------------------------------------------------ Prolonged exposure of HL-60 cells to various cytoskeleton-modifying chemicals (i.e., nocodazole and taxol) are very harmful to cell viability, inducing rapid apoptosis. The present study was conducted primarily on a Bcl-2 overexpressing subline of HL-60 cells, which is more refractory to apoptosis and exhibits the chemically induced differentiation properties of the parent cell line \[[@B18]\]. Lethal effects of NC or TX are delayed in HL-60-*bcl*-2 cells, compared to HL-60/S4 cells, allowing a window of time for determining the effects of these MT modifying chemicals on nuclear shape and nuclear differentiation. There are two additional differences between HL-60-*bcl*-2 and HL-60 cells observed in this study. The first difference: \~10% of the undifferentiated HL-60-*bcl*-2 cells exhibit micronuclei, compared to \~0.8% of HL-60 cells \[[@B26]\]. Micronuclei are generally regarded as the products of abnormal mitoses, where the enclosed chromosomes or chromosome fragments fail to congress at the mitotic plate, but are still surrounded by a post-mitotic nuclear envelope \[[@B27]-[@B30]\]. In normal human cells growing in culture (e.g., lymphocytes), they occur in less than 0.5% of the cells \[[@B31]\]. They can be induced in cells by treatment with a variety of DNA breakage conditions (e.g., irradiation) or spindle disrupting agents (e.g., colchicine). Micronuclei in HL-60 cells are described as representing amplified acentric euchromatic genes (such as c*-myc*) and appear to form dynamically during S phase \[[@B32]\]. We observed that micronuclei in HL-60-*bcl*-2 cells exhibited immunostaining of centromeres, heterochromatin and nucleolar antigens (Figure [7](#F7){ref-type="fig"}), which suggests differences from the earlier interpretation of the nature of micronuclei in HL-60 cells. Especially puzzling was our observation that the majority of micronuclei possessed comparable amounts of lamin B, but reduced amounts of LBR, in comparison to the companion main nucleus. A recent study employing MCF-7 cells observed micronuclei containing lamins A/C and B1 and a relative deficiency of LBR \[[@B33]\]. The same study demonstrated that when significant numbers of micronuclei were induced by prolonged exposure to the spindle disrupting chemical curcumin, the resulting nuclear envelopes contained LBR. Current views on the sequence of protein additions to post-mitotically reformed nuclei agree that LBR enters the nascent nuclear envelope well before lamin B \[[@B34],[@B35]\]. It is possible that untreated HL-60-*bcl*-2 and MCF-7 micronuclear envelopes form later than the main nucleus, after the cytoplasmic pool of LBR is exhausted, whereas curcumin \"induced\" micronuclei form at about the same time as main nuclei. The second difference: RA treated HL-60-*bcl*-2 cells exhibit a small population (up to 20%) with multilobed nuclei, which we almost never observed with HL-60/S4 cells. There may be some clinical significance to this latter observation. A small percent of granulocyte nuclei with 5 or more nuclear lobes in human blood smears is considered diagnostic for megaloblastic anemia (vitamin B12 or folic acid deficiency) \[[@B20]\]. The present data with RA treated HL-60-*bcl*-2 cells suggests that delayed or dysfunctional apoptosis might play a role in these human diseases. It is of interest that neutrophil nuclear multilobulation (hypersegmentation) has been described in two circumstances that delay apoptosis: 1) Glucocorticoid administration to patients induces hypersegmentation \[[@B36]\], and *in vitro*glucocorticoid treatment of neutrophils prolongs their survival \[[@B37]-[@B39]\]. 2) Granulocyte colony-stimulating factor (G-CSF) administration to rats induces hypersegmentation in mature neutrophils \[[@B40]\], and *in vitro*treatment of neutrophils with G-CSF prolongs their survival \[[@B38],[@B41]\]. Major cytoskeletal influences on nuclear shape in HL-60-*bcl*-2 --------------------------------------------------------------- The most important present observation is the suppression of RA induced nuclear lobulation in HL-60-*bcl*-2 cells by simultaneous exposure to NC (Figure [9](#F9){ref-type="fig"}). This observation implies that MTs must be intact during nuclear differentiation. Furthermore, we have observed (data not shown) that HL-60-*bcl*-2 cells can be made 0.1 μM NC on day 2 or 4 after addition of RA, still exhibiting inhibition of nuclear lobulation. This observation suggests that the requirement for intact MTs is not an early event in the nuclear differentiation process. When RA and NC treated HL-60-*bcl*-2 cells were examined by confocal immunofluorescent staining with anti-lamin B, the ovoid nuclei revealed extensive \"wrinkling\" of the nuclear envelope (Figures [10](#F10){ref-type="fig"} and [11](#F11){ref-type="fig"}). We suggest that this \"wrinkling\" reflects expansion (growth) of the nuclear envelope in the absence of nuclear lobulation. The present study demonstrates a lack of requirement of an intact actin microfilament system for RA induced granulocytic nuclear differentiation (Figure [1](#F1){ref-type="fig"}). This conclusion is based upon the observation that incubation of HL-60/S4 and HL-60-*bcl*-2 cells with 1 μM CD in the presence of RA does not inhibit nuclear lobulation or the formation of ELCS. Our results agree with an earlier study \[[@B17]\]. In addition, we extended the incubation (with CD and RA) to times where nuclear differentiation is more definitive. Our data also demonstrated that prolonged incubation with CD alone produces significant nuclear envelope folding, which is best appreciated by confocal immunostaining of the nuclei (Figure [8](#F8){ref-type="fig"}). The mechanism for the nuclear shape changes in undifferentiated cell nuclei exposed to prolonged incubation with CD is presently unknown. There is increasing evidence that actin microfilament interactions with the nuclear envelope exist, mediated via actin binding spectrin-like proteins, variously named Syne 1 and 2/ nesprin-1 and 2/ ANC1/ NUANCE \[[@B12]-[@B15]\]. However, a counter argument for the relevance of this actin interacting system to nuclear differentiation in the HL-60 cell system can be made. Nesprin-1 binds directly to lamin A and emerin \[[@B14]\]; but undifferentiated and granulocytic HL-60 cells possess negligible amounts of lamin A, with emerin primarily localized in the cytoplasm \[[@B5]\]. Furthermore, published \[[@B13]\] and unpublished data (A. Karakesisoglu, A. Olins and D. Olins) indicate that HL-60 cells, undifferentiated or RA treated, possess only trace amounts of NUANCE. The finding that TX treatment of cells has dramatic consequences to interphase nuclear structure has been reported before. Exposure of human carcinoma cells (Ishikawa and HeLa) to 0.01--0.1 μM TX for up 20 hours, followed by incubation in drug-free media for up to 72 hours, led to nuclear envelope \"unraveling\" and clustering of nuclear pores \[[@B42]\]. The authors observed lobulated and micronuclei, much as observed here. In the case of HL-60-*bcl*-2, the nuclear structural changes superficially mimic the effects of RA treatment, but occur much faster (Figure [12](#F12){ref-type="fig"}). The bundling of MTs in HL-60 cells in response to exposure to 1.0 μM TX for 24 hours has also been reported \[[@B43]\], but no mention was made of the nuclear envelope changes. Furthermore, the authors noted that at 0.1 μM TX (or greater) the HL-60 cells showed clear apoptosis by 24 hours. In a later paper by the same group \[[@B44]\], increased expression of Bcl-2 or Bcl-x~L~yielded HL-60 cells with considerably greater resistance to TX induced apoptosis. It seems to us that the dramatic cell and nuclear changes that we observe have no obvious relationship to normal granulocytic nuclear lobulation, but underscore an effect of MT integrity upon nuclear shape. The formation of micronuclei, in particular, suggests that TX may exert its effects on nuclear shape by interfering with normal mitotic chromosome distribution. A model for granulocytic nuclear lobulation ------------------------------------------- Although far from a complete understanding of all the molecular forces involved in shaping the granulocytic nucleus, a working model that incorporates existing data and emerging concepts provides a useful perspective for future experimentation (Figure [16](#F16){ref-type="fig"}). The process of granulocytic nuclear lobulation can be viewed as a dynamic balance of stabilizing and distorting forces. The working model contains the following assumptions: 1) the flexible nuclear envelope (due to the paucity of lamins A/C and B1) is \"tacked down\" to the underlying heterochromatin (enhanced by the elevated LBR); 2) the nuclear envelope undergoes invaginations in the region of the centrosome due to dynein movement along MTs; 3) new membrane materials are added to the nuclear envelope via lateral diffusion from the ER, resulting in net membrane growth \[[@B3]\]; 4) constraints on nuclear shape by actin and spectrin-like proteins are weak, due to the paucity of lamins A/C and NUANCE and the cytoplasmic localization of emerin; 5) constraints on nuclear shape by vimentin-envelope interactions are minimal in the differentiating HL-60 cell system. ::: {#F16 .fig} Figure 16 ::: {.caption} ###### **Model for granulocytic nuclear lobulation.**A. Postulated changes in nuclear envelope flexibility arising from changes in nuclear envelope composition. HL-60 cell states: undifferentiated; granulocyte, RA treated; monocyte, TPA treated. Abbreviations: inm, inner nuclear membrane; LMNB2, lamin B2; LMNA/C, lamins A/C; LMNB1, lamin B1; HP1, heterochromatin protein 1; LBR, lamin B receptor. Due to a paucity of lamins A/C and B1, the nuclear envelope is believed to be more flexible in the undifferentiated and granulocytic cell states. B. Balance of forces postulated to be affecting granulocytic nuclear shape. Microtubules (green) and affiliated dynein motors (red circles) are assumed to produce nuclear envelope invaginations (bent arrows). Actin with affiliated spectrin-like proteins and vimentin are assumed to be pulling outwards on the nuclear envelope (thin arrows). Current evidence does not favor a major contribution by actin; any significant contribution by vimentin is presently unknown. The large straight arrows indicate continued influx of nuclear envelope components from the endoplasmic reticulum, allowing sustained membrane growth. Abbreviations: MTs, microtubules; IFs, intermediate filaments; ER, endoplasmic reticulum; Eu, euchromatin; H, heterochromatin. The nuclear compartment is colored blue; the cytoplasm, pink. :::  ::: We suggest that nuclear envelope deformability is an important factor and depends upon the amount of underlying lamins: the less lamin protein, the more pliable the nuclear envelope. Granulocytic forms of HL-60 exhibit deficiency of lamins A/C and B1, whereas both types of lamins are present in monocyte/macrophage forms \[[@B5]\]. The absence of lamins A/C in normal granulocytes and their presence in macrophages has been previously noted \[[@B45]\]. The present data on granulocytic differentiation in HL-60-*bcl*-2 demonstrates that nuclear lobulation correlates with low levels of lamins A/C and B1 coupled with a rise in LBR levels (Figure [4](#F4){ref-type="fig"}). The pivotal role of LBR in determining granulocytic nuclear lobulation was demonstrated with the human Pelger-Huet anomaly and murine Ichthyosis mutations \[[@B8],[@B9]\]. These studies demonstrated that LBR functions in a dose-dependent manner: homozygous mutants present a more severe phenotype and lower amounts of LBR, than in the heterozygous state. The influence of LBR on granulocytic nuclear lobulation is consistent with its known properties \[[@B46]\]. LBR is embedded within the nuclear envelope inner membrane via 8 transmembrane domains (the C-terminal \~400 aa), and associated with lamin B, chromatin and HP1α in the N-terminus (\~200 aa). In the absence of sufficient LBR, nuclear lobulation is prevented and the normally peripheral heterochromatin is redistributed into a more centrally condensed form \[[@B8],[@B9]\]. A role for MT integrity is implicit in our present observation that exposure of HL-60-*bcl*-2 cells to NC prevents nuclear lobulation during exposure of the cells to RA. Direct interaction between MTs and the interphase nuclear envelope in mammalian cells have not documented. However, a recent model for mitotic nuclear envelope breakdown \[[@B47],[@B48]\] can be adapted to the situation of nuclear lobulation. The nuclear envelope breakdown model proposes that cytoplasmic dynein attaches MTs to the nuclear envelope, pulling the envelope towards the centrosomal region. The excess envelope near to the centrosome produces nuclear invaginations; the tension on the non-growing nuclear envelope generates tears and the mixing of nuclear and cytoplasmic materials. If we assume that the nuclear envelope is still growing in the case of RA differentiating HL-60 cells, invaginations and lobulations might be expected to accumulate within the intact nuclear envelope. The present study demonstrates proximity of the centrosome to nuclear lobulation (Figure [15](#F15){ref-type="fig"}). But as yet, there is no direct evidence for cytoplasmic dynein playing a role in granulocytic nuclear differentiation. Our present data suggests that an intact actin microfilament system does not play a major role in granulocytic nuclear lobulation (Figure [1](#F1){ref-type="fig"}). The best described mechanism of actin interacting with the nuclear envelope involves spectrin-like proteins, which may bridge cytoplasmic actin to nuclear envelope proteins, such as lamin A and emerin \[[@B12]-[@B15]\]. But undifferentiated and granulocytic HL-60 cells possess very little lamin A/C and emerin is primarily cytoplasmic \[[@B5]\], suggesting that this bridging system may not be functional in these cell forms. In SW-13 cells the absence of intermediate filaments (vimentin) has been correlated with nuclear envelope folds or invaginations \[[@B16]\]. We have observed a decrease in vimentin during differentiation of granulocytic HL-60/S4 cells \[[@B3]-[@B5]\], suggesting that reduced vimentin concentrations may contribute to granulocytic nuclear lobulation. The proposed model for granulocytic nuclear lobulation yields several testable predictions: 1) Expression of lamins A/C and B1 in HL-60 cells should strengthen the nuclear envelope, minimizing nuclear lobulation following exposure of the cells to RA; 2) \"Knock-down\" or expression of a \"dominant negative\" form of LBR in HL-60 cells would be expected to suppress granulocytic lobulation; 3) Overexpression of dynamitin in HL-60 cells should inhibit dynein activity \[[@B49]\], preventing nuclear lobulation following RA induced differentiation. A number of these experiments are already in progress. Conclusions =========== Employing Bcl-2 overexpressing HL-60 cells, which are more refractory to induced apoptosis than the parent cell line, we demonstrated that disruption of the MTs by nocodazole prevented nuclear lobulation in response to RA treatment. These results implicate the necessity of an intact MT system for granulocytic nuclear shape. Cytochalasin D, on the other hand, did not suppress RA induced nuclear lobulation. Combined with the decreasing levels of intermediate filaments (vimentin) during differentiating granulocytic forms of HL-60 cells, the role of the MT system appears to be quite central to the nuclear shape change. Recent models on the role of a MT bound motor (dynein) in facilitating mitotic nuclear envelope breakdown suggest that similar tension forces on the differentiating granulocyte nucleus, combined with continued influx of nuclear envelope components, could explain nuclear invaginations and lobulation. Methods ======= Cells and chemicals ------------------- Two HL-60 cell sublines were employed in this study: HL-60/S4 \[[@B50]\], which achieves maximum nuclear lobulation in 4 days following addition of RA; HL-60-*bcl*-2 \[[@B18]\], which achieves maximum lobulation in \~7 days, comparable to the parent cell line. Cultivation conditions were exactly as described previously \[[@B3],[@B4]\]. The following chemicals were purchased from Sigma-Aldrich (St. Louis, MO): *all-trans*retinoic acid (RA), nocodazole (NC), taxol (TX), cytochalasin D (CD). Stock solutions of these chemicals were stored at -20°C as described earlier \[[@B3],[@B4]\]. Antibodies ---------- Guinea pig antisera were the generous gifts of two Ph. D. students in the laboratory of H. Herrmann (German Cancer Research Center, Heidelberg): anti-LBR and anti-emerin, from C. Dreger; anti-lamin A and anti-lamin B1, from J. Schumacher. Goat anti-lamin B was obtained from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Rabbit anti-4 × dimethyl H3K9 (lysine 9 on histone 3) was generously provided by T. Jenuwein (Biocenter, Vienna). Human auto-antisera anti-centromere (CREST) and anti-nucleolus (antigens unknown) were purchased from Antibodies Inc. (Davis, CA). Rabbit anti-calreticulin was from Calbiochem (San Diego, CA). Mouse monoclonal antibodies against α-tubulin, γ-tubulin and Golgi p58 were all purchased from Sigma-Aldrich. Mouse monoclonal anti-vimentin (3B4) was a gift of H. Herrmann and has been described before \[[@B51]\]. FITC-, Cy3-, Cy5- and HRP-conjugated donkey secondary antibodies were all purchased from Jackson ImmunoResearch Laboratory, Inc. (West Grove, PA). TO-PRO-3 and SlowFade were obtained from Molecular Probes, Inc. (Eugene, OR). Fixation and staining --------------------- For Wright-Giemsa staining, cells were cytospun onto ethanol-cleaned microscope slides, fixed in room temperature methanol for 15 min, air-dried and stained as described earlier \[[@B3]\]. For analysis of the percentage of cells in the various nuclear morphology categories (Figures [3](#F3){ref-type="fig"}, [9](#F9){ref-type="fig"} and [12](#F12){ref-type="fig"}), approximately 150 cells were observed and classified in each experiment. In the majority of immunostaining experiments the procedure followed that described previously \[[@B4]\], with some modifications, as follows: 1) Microscope slides were soaked overnight in 1/1 ethanol/ether and freshly coated with poly-L-lysine (MW \~150--300,000; Sigma-Aldrich), just before centrifugation of the cells. 2) No coverslip was used during antibody incubations, to minimize loss of cells. 3) Prior to the application of primary antibodies, slides were incubated with 5% normal donkey serum (Jackson ImmunoResearch Laboratory) in PBS for 15--30 min at 37°C in a moist chamber. 4) In the cases of monoclonal anti-γ-tubulin and anti-vimentin, where antigenicity appeared to be destroyed by HCHO fixation, slides were fixed in methanol (-20°C, 10 min), acetone (-20°C, 1 min) followed by three washes in PBS (5 min each). The cells were not excessively flattened by either procedure, maintaining sufficient 3-D structure to justify stereo viewing. Confocal images were collected on a Zeiss 510 Meta. Stereo images were obtained as ± 15° projections through the stack of confocal images. Immunoblot analysis ------------------- SDS total cell extracts, obtained from undifferentiated and RA treated HL-60-*bcl*-2 cells over a three-week period (simultaneously with preparation of slides for Wright-Giemsa staining), were analyzed by immunoblotting as previously described \[[@B5]\]. No attempt was made to separate viable cells from debris prior to SDS extraction. Comparable amounts of total cell protein were loaded into each lane of the SDS-PAGE, as judged by Ponceau S staining of the PVDF membrane after protein transfer. Several different ECL exposures were collected on X-ray film and subsequently scanned with a Bio Rad Chemi Doc for densitometric analyses. Abbreviations ============= RA, retinoic acid; MT, microtubules; NC, nocodazole; TX, taxol; CD, cytochalasin D; ELCS, nuclear envelope-limited chromatin sheets Authors\' contributions ======================= ALO performed the microscopy and prepared the figures. DEO performed the tissue culture, immunostaining and immunoblotting. Both authors were involved in the conception of the study and have read and approved the final manuscript. Acknowledgements ================ This work was supported by funds from the Department of Biology, Bowdoin College. We wish to express our gratitude to members of the Bowdoin faculty who have encouraged our research, especially W. Steinhart, B. Kohorn and D. Page. We also wish to express our appreciation to our colleagues (H. Herrmann and P. Lichter) at the German Cancer Research Center for stimulation and enlightened criticism.

PubMed Central

2024-06-05T03:55:47.415634

2004-8-19

PMC516026

Background ========== The Campanulaceae *sensu stricto*are a nearly cosmopolitan angiosperm family consisting of latex-bearing, primarily perennial herbs or occasional subshrubs that typically have alternate leaves, sympetalous corollas, inferior ovaries, and capsular fruits. Allied to the Campanulaceae are the Lobeliaceae, Cyphiaceae, Cyphocarpaceae, Nemacladaceae, Pentaphragmataceae, and Sphenocleaceae; at times, all of these taxa have been included in the Campanulaceae at varying taxonomic rank by different authors (Table [1](#T1){ref-type="table"}). Taxonomic treatments lack consensus (Table [1](#T1){ref-type="table"}) and phylogenetic work has only recently been attempted. Campanulaceae in the strict sense are recognized as 600 \[[@B1]\] to 950 \[[@B2]\] species distributed among 35 \[[@B1]\] to 55 \[[@B2]\] genera. Generic circumscription and intrafamilial classification vary widely according to author. Within the family as few as two \[[@B3]\] and as many as 18 \[[@B4]\] tribes have been recognized (Table [1](#T1){ref-type="table"}). Fedorov\'s more recent work \[[@B5]\] recognized eight tribes (Table [1](#T1){ref-type="table"}), but only included taxa present in the former Soviet Union. Although Kolakovsky\'s treatment of Old World Campanulaceae \[[@B4]\] is the most recently published attempt to produce a more complete intrafamilial classification of the Campanulaceae (Table [1](#T1){ref-type="table"}), the scope of the work is limited compared to that of either A. de Candolle \[[@B3]\] or Fedorov \[[@B5]\]. In all treatments, the Campanuleae and Wahlenbergieae (at whatever rank) are typically the largest, most inclusive taxa, with segregate tribes consisting of only one to a few genera. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Classification systems of Campanulaceae. All major intrafamilial subdivisions are included (level of subdivisions indicated by number of dashes) but only those genera sampled in this study are included. If sampled genera are not listed, the genus was not recognized by the author but, rather, was subsumed into one of the listed genera. ::: A deCandolle AP deCandolle Schönland Federov Takhtajan Kovanda Kolakovsky --------------------- --------------------- ----------------------- ----------------------- --------------------- --------------------- ------------------------- **Campanulaceae** **Campanulaceae** **Campanulaceae** **Campanulaceae** **Campanulaceae** **Campanulaceae** **Campanulaceae** **-Subtribus I** **-Wahlenbergieae** -**Lobelioideae** **-Sphenocleoideae** -**Cyanantheae** -**Campanulinae** -**Prismatocarpoideae**  *Jasione*  *Jasione* **-Cyphioideae** **-Campanuloideae**  *Cyananthus*  *Campanula*  *Prismatocarpus*  *Codonopsis*  *Platycodon* **-Campanuloideae** \--**Campanuleae** **-Wahlenbergieae**  *Symphyandra*  *Roella*  *Platycodon*  *Codonopsis* **\--Pentaphragmeae**  *Campanula*  *Wahlenbergia*  *Legousia* -**Canarinoideae**  *Wahlenbergia*  *Wahlenbergia* **\--Sphenocleae**  *Symphyandra*  *Edraianthus* -**Wahlenberginae** -**Wahlenbergoideae**  *Prismatocarpus*  *Prismatocarpus* **\--Campanuleae** **\--Peracarpeae**  *Jasione*  *Wahlenbergia* \--**Wahlenbergieae**  *Roella*  *Roella* **Campanulinae** **\--Ostrowskieae**  *Codonopsis*  *Codonopsis*  *Jasione* **-Subtribus II**  *Edraianthus*  *Symphyandra* **\--Michauxieae**  *Merciera*  *Cyananthus*  *Wahlenbergia*  *Petromarula* **-Campanuleae**  *Trachelium* **\--Phyteumateae**  *Roella*  *Roella*  *Codonopsis*  *Campanula*  *Petromarula*  *Campanula*  *Asyneuma*  *Prismatocarpus*  *Edraianthus*  *Platycodon*  *Trachelium*  *Campanula* **Wahlenberginae**  *Legousia* **-Platycodoneae**  *Jasione*  *Cyananthus*  *Symphyandra*  *Trachelium*  *Cyananthus* \--**Wahlenbergieae**  *Platycodon* **-Platycodinae** **\--Azorineae**  *Musschia*  *Symphyandra*  *Jasione*  *Codonopsis*  *Musschia*  *Platycodon* \--**Musschieae** **-incertae sedis**  *Musschia*  *Prismatocarpus* **\--Edraiantheae** -**Campanuleae**  *Musschia*  *Merciera* **-Merciereae**  *Merciera*  *Edraianthus*  *Campanula* \--**Echinocodoneae**  *Merciera*  *Edraianthus* \--**Jasioneae**  *Legousia* \--**Annaea**  *Wahlenbergia*  *Jasione*  *Triodanis* **\--Muehlbergelleae**  *Codonopsis* -**Michauxieae** **\--Theodorovieae**  *Roella* **-Phyteumaeae** **\--Gadellieae** **Platycodinae**  *Asyneuma* **\--Ostrowskieae**  *Platycodon*  *Trachelium* -**Campanuloideae**  *Musschia*  *Petromarula* \--**Campanuleae** **-Peracarpeae**  *Campanula*  *Symphyandra*  *Trachelium* **\--Phyteumateae** **\--Peracarpeae** **\--Sergieae** **\--Michauxieae** \--**Neocodoneae**  *Asyneuma*  *Legousia* \--**Edraiantheae**  *Edraianthus* **\--Sachokieleae** **\--Mzymteleae** ::: The most comprehensive treatment of the Campanulaceae remains the monograph of A. de Candolle \[[@B3]\], who recognized two groups corresponding to the Wahlenbergieae and Campanuleae (Table [1](#T1){ref-type="table"}). Simple basal leaves and simple, alternate or occasional whorled, cauline leaves that are often different in shape than the basal leaves, characterize the Campanuleae in de Candolle\'s sense. Flowers are solitary or borne in cymes or racemes, and have five corolla lobes that are mostly fused proximally. The inferior ovary usually has 3--5 carpels and develops into a capsule that mostly dehisces by lateral pores (rarely a berry). The Wahlenbergieae are mostly perennials characterized by simple, alternate, cauline leaves. Flowers are solitary or borne in cymes or heads, and petals may be free, proximally fused, or distally fused. The ovary is inferior, semi-inferior, or superior, and consists of two, three, or five carpels. The fruit is generally a capsule dehiscing by apical pores or valves (rarely a berry). Both groups have five stamens with filaments that are often proximally dilated and anthers with introrse dehiscence; nectaries are generally present, and many ovules are attached to axile placentae. The entire family is characterized by secondary pollen presentation in which protandry is combined with a close association of anthers around the style and introrse pollen discharge onto the style for presentation to pollinators. This syndrome is similar to that found in Lobeliaceae and Asteraceae, but invaginating stylar hairs are unique to the Campanulaceae. Capsule characters vary considerably and provide the basis for most intrafamilial classification schemes. Campanuleae typically include taxa with capsules dehiscing by lateral pores, whereas Wahlenbergieae usually include taxa with capsules dehiscing by apical valves. Ovary characters, such as carpel number and position, have also been important in traditional classifications. For example, the monotypic tribe Platycodoneae \[[@B6]\] or subtribe Platycodinae (Table [1](#T1){ref-type="table"}) is sometimes segregated. It is defined by carpels that are equal in number to and alternate with the calyx lobes, whereas in Campanuleae and Wahlenbergieae the carpels are often fewer than the calyx lobes, or if the same in number then opposite them \[[@B1],[@B7],[@B8]\]. Little correlation appears to exist among diagnostic features; therefore there is considerable taxonomic disagreement among classifications. In certain instances it is difficult to discern the rationale behind tribal placement of individual genera. The high level of disagreement among both inter- and intrafamilial classifications of the Campanulaceae indicates that phylogenetic assessment of the family is needed. Cosner, in her thesis \[[@B9]\], included an early version of a portion of the work described here, and Eddie, in his thesis \[[@B10]\] developed phylogenetic hypotheses based on ITS sequence data and morphology. An expanded version of the ITS work has been published \[[@B11]\] but leaves some major lineages unsampled and the relationships among some major groups are unresolved or poorly supported. Further phylogenetic work is clearly warranted. The chloroplast genome has proven to be a useful tool for phylogenetic reconstruction. Chloroplast DNA (cpDNA) of land plants is highly conserved in nucleotide sequence as well as gene content and order; its relatively slow rate of evolution makes it an excellent molecule for phylogenetic and evolutionary studies \[[@B12]\]. Chloroplast genomes of photosynthetic angiosperms average about 160 kilobase pairs (kb) in size; the circular chromosome is divided by two copies of a large (in angiosperms usually about 25 kb) inverted repeat (IR) into large and small single copy regions (LSC and SSC, respectively) \[[@B13],[@B14]\]. Restriction site mapping, gene sequencing, and analysis of gene order rearrangements have been used to study cpDNA variation for phylogenetic investigations \[[@B12]\]. Here we use the distribution of gene order changes in the chloroplast genomes of the Campanulaceae to estimate phylogenetic relationships in the family. Generally, major gene order changes are rare. Therefore, when they occur, such mutations are extremely useful as phylogenetic markers because they are readily polarized and typically lack homoplasy \[[@B15]-[@B17]\]. Four categories of cpDNA gene order rearrangements have been proposed: 1) inversions, 2) insertions or deletions, 3) IR expansion or contraction or loss, and 4) transpositions; all of which may have occurred during chloroplast genome evolution in the Campanulaceae \[[@B18]\]. When rearrangements have been discovered elsewhere, they are generally few and easily characterized. The distributions of such characters make effective markers of monophyletic groups. For example, both the loss of one copy of the IR and inversions are extremely useful characters in legume phylogeny \[[@B19],[@B20]\], defining large clades within the family. Other examples of phylogenetically informative inversions are found within Asteraceae \[[@B21]\], Ranunculaceae \[[@B22],[@B23]\], ferns \[[@B24],[@B25]\], and vascular plants \[[@B26]\]. Many other examples could be cited. The earlier work of Cosner \[[@B9],[@B18]\] and Knox \[[@B27],[@B28]\] characterized some chloroplast genomes of the Campanulales and identified a number of rearrangements relative to the consensus gene order of angiosperms found in tobacco. Members of the Lobeliaceae exhibit multiple rearrangements but are less rearranged than the Campanulaceae. Three rearrangements may be shared between the two families -- a loss of the *accD*gene, the expansion of the inverted repeat into the small single copy region, and, perhaps, an inversion of the region corresponding to tobacco probes 40--44. Then, within the Campanulaceae, more than 40 inversions, more than eight putative transpositions, two additional gene losses, additional IR expansion or contraction events and 18 large insertions greater than 5 kb in size may have contributed to observed differences among the chloroplast genomes sampled \[[@B9]\]. Due to this unprecedented number of gene order mutations, it is not possible to unambiguously determine the evolutionary order of most events or in some cases to even define the events themselves. This complex situation poses special problems for using these rearrangements to estimate phylogenetic relationships. In this paper we develop alternative character codings for the data and compare the results of parsimony analyses of the different data sets. In addition, we compare the ability of the gene order data to support robust phylogenetic hypotheses to that of sequence data from *rbcL*and ITS. Finally, the phylogenetic implications of the cpDNA rearrangement data for the Campanulaceae are discussed. Results ======= Our data indicate that the eighteen mapped Campanulaceae chloroplast genomes (Table [2](#T2){ref-type="table"}) are drastically rearranged relative to those of other land plants (Fig. [1](#F1){ref-type="fig"}). The tobacco cpDNA gene order represents the consensus gene order for angiosperms \[[@B13],[@B15]\]. Therefore rearrangements in Campanulaceae chloroplast genomes were identified relative to tobacco. Because characterizing specific mutational events was not always possible three different coding methods (Matrix 1, 2 and 3) were developed. Matrix 1 coded all gene order changes as endpoints (derived adjacencies, relative to tobacco, were identified and scored for presence/absence). Matrix 2 and 3 involved recoding some endpoint characters to recognize 31 specific mutations. Matrix 2 and 3 were analyzed with and without weighting. See Methods for additional details on character encoding and analyses. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Species of Campanulaceae mapped for chloroplast DNA structural rearrangments. ::: **Species** **Source** **Voucher**^a^ -------------------------------------------- ---------------------- ---------------- *Adenophora confusa*Nannf. R.C. Haberle 179 TEX *Asyneuma virgatum*(Labill.) Bourm. Berlin-Dahlem^b^0104 *Campanula elatines*L. T. Ayers 88--287 BH *Codonopsis viridis*Wall. T. Ayers 88--229 BH *Cyananthus lobatus*Wall. ex Benth. M. Cosner 179 OS *Edraianthus graminifolius*(L.) A.DC. T. Ayers 88--195 BH *Jasione heldreichii*Boiss. & Orph. T. Ayers 88--208 BH *Legousia falcata*(Ten.) Fritsch ex Janch. Berlin-Dahlem^b^0143 *Merciera tenuifolia*(L.f.) A. DC. K. Steiner 2445 OS *Musschia aurea*Dumort T. Ayers 88--274 BH *Petromarula pinnata*(L.) A. DC. T. Ayers s.n.^c^ BH *Platycodon grandiflorus*(Jacq.) A. DC. T. Ayers 88--216 BH *Prismatocarpus diffusus*(L.f.) A. DC. K. Steiner 2448 OS *Roella ciliata*L. T. Ayers s.n.^c^ BH *Symphyandra hofmannii*Pant. T. Ayers 88--225 BH *Trachelium caeruleum*L. M. Cosner 173 OS *Triodanis perfoliata*(L.) Nieuwl. M. Cosner 178 OS *Wahlenbergia gloriosa*Lothian T. Ayers 88--217 OS a, abbreviations for herbaria: BH = Bailey Hortorium (Cornell University, Ithaca, NY); OS = Ohio State University Herbarium (Columbus); TEX = University of Texas Herbarium (Austin) b, Botanischer Garten and Botanisches Museum, Berlin-Dahlem c, s.n. = *sin numero*(no number assigned by collector) ::: ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Linearized cpDNA maps for 18 species (Table 2) of Campanulaceae (plus tobacco) showing order in which the consecutively numbered tobacco probes hybridized. Lines under maps indicate location and extent of IR. Asterisks indicate the position of the putative *23S rDNA*duplicative transposition; parenthetical asterisk (\*) indicates partial deletion/divergence of the *23S rDNA*transposition. Size and location of large insertions designated by \"i\" followed by size in kb (insertions less than 5 kb not shown). :::  ::: Seventy-nine variable characters were included in the endpoints only matrix (Matrix 1). Forty-two of the derived character states were unique to a single taxon and 37 were phylogenetically-informative. Six trees of 97 steps were obtained with consistency indices of 0.81 with autapomorphies included and 0.67 with autapomorphies excluded (CI = 0.81/0.67). Ten nodes were common to the six shortest trees (Fig. [2](#F2){ref-type="fig"}). Eight of those ten nodes have bootstrap values (BS) greater than 50, but BS exceeded 90 for only three nodes. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### One of six shortest trees obtained in the maximum parsimony analysis of Matrix 1. Tree length is 97 steps; consistency index is 0.81 (with autapomorphies, 0.66 without). The number of character changes is given above the branches and bootstrap values (where greater than 50) are given below. Arrows indicate nodes that collapse in the strict consensus of all six shortest trees. :::  ::: To construct Matrix 2 and Matrix 3, we interpreted endpoints as events where possible. Under our interpretation, several types of rearrangements contributed to cpDNA evolution in the family, including multiple inversions (scored primarily as endpoints), five IR expansion or contraction events, eight transpositions, two deletions, and 14 large insertions greater than 5 kb in size (Table [3](#T3){ref-type="table"}). Although transposition probably does occur, at least occasionally, in the chloroplast genome \[[@B29]\], it is not a common mechanism of rearrangement. Still, in some instances transposition could explain rearranged gene orders with fewer steps than multiple inversions and so we hypothesized transposition events in some cases. Matrix 2 and 3 each were composed of 84 variable characters of which thirty-one and thirty-four, respectively, were parsimony informative. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### List of chloroplast DNA rearrangement characters. Numbers refer to tobacco cpDNA hybridization probes. Endpoints are given as novel probe adjacencies. Other rearrangement types are indicated as: T = transposition (T\' = secondary transposition \[14\] of most of 53--56); I = inversion; i = insertion; D = deletion (or divergence); IRc and IRe = IR contraction or expansion, respectively (followed by single copy region affected). Characters marked with asterisks (\*) are those rescored in Matrix 3 relative to Matrix 2. ::: ----------------- ------------------- ----------------- ----------------- \*1. 11/60 22\. T (53,54) 43\. 40/56 64\. i (15) \*2. 56/53 23\. T (53--56) 44\. 39/37 65\. IR^e^(LSC) \*3. 49/37 24\. 98/28 45\. 37/44 66\. T (5--8) 4\. 40/35 25\. 53/98 46\. 56/61 67\. T (6--9) \*5. 28/11 26\. T\' (53--56) 47\. 76/96 68\. i (9) 6\. 15/76 27\. 49/39 48\. 77/106 69\. i (18) 7\. 60/27 \*28. 37/40 49\. i (5) 70\. l (60--61) \*8. 26/44 29\. 56/39 50\. 39/25 71\. T (28) 9\. 41/47 30\. 53/40 51\. 9/37 72\. T (16--17) 10\. 48/36 31\. 37/28 52\. 48/56 73\. i 7 11\. 35/25 32\. 40/26 53\. 49/61 74\. IR^e^(SSC) 12\. 16/90 33\. 27/44 54\. 5/29 75\. i (8) 13\. T (93) 34\. 8/40 55\. 50/28 76\. i (9) 14\. 84/76 35\. 39/26 56\. 26/50 77\. i (9) 15\. 84/90 36\. 27/11 57\. 40/16 78\. i (15) 16\. 10/49 37\. 56/36 58\. 25/10 79\. i (16) 17\. 53/28 38\. 44/10 59\. 15/56 80\. D (93) 18\. 37/60 39\. 9/53 60\. IRc (LSC) 81\. i (5) 19\. 56/11 40\. 49/36 61\. IR^e^(SSC) 82\. i (6) 20\. I (16--17) 41\. 8/36 62\. IR^e^(LSC) 83\. i (19) 21\. 56/27 42\. 9/40 63\. D (45--46) 84\. i (5) ----------------- ------------------- ----------------- ----------------- ::: The unweighted analysis of Matrix 2 produced 241 equally parsimonious trees of 93 steps (CI = 0.90/0.79). The strict consensus of the 241 trees includes six resolved nodes (Fig. [3a](#F3){ref-type="fig"}) all six of which were supported by BS values of at least 50 and five nodes were supported at 90% or above. The weighted analysis of Matrix 2 (Fig. [3b](#F3){ref-type="fig"}) resulted in 12 equally parsimonious trees of 125 steps (CI = 0.93/0.82). The strict consensus of the twelve trees retains ten resolved nodes. Seven of the ten nodes have BS values over 50 and for five nodes BS ≥ 90. Both analyses of Matrix 3 generated the same two equally parsimonious trees (Fig. [4](#F4){ref-type="fig"}). The lengths of the two trees were 87 steps (CI = 0.97/0.92) or 118 steps (CI = 0.97/0.93) depending on whether unweighted (Fig. [4a](#F4){ref-type="fig"}) or weighted (Fig. [4b](#F4){ref-type="fig"}) analyses were conducted. Only three endpoint characters are homoplasious in the Matrix 3 analyses (Fig. [5](#F5){ref-type="fig"}). The strict consensus of these two trees retains nine resolved nodes, all nine of which are supported with BS ≥ 50. Six (or five in the weighted analysis) nodes received strong support (BS ≥ 90). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Trees obtained in unweighted (a) and weighted analyses (b) of Matrix 2. Fig. 3a shows one of 241 shortest trees of 93 steps; consistency index is 0.90 (including autapomorphies, 0.78 without). Fig. 3b shows one of 12 shortest trees of 125 steps; consistency index is 0.93 (including autapomorphies, 0.80 without). The number of character changes is given above the branches and bootstrap values are given below. Arrows indicate nodes that collapse in the strict consensus of all the shortest trees. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### The two shortest trees obtained in both the unweighted and weighted analyses of Matrix 3. The trees are 87 steps long without weights and 125 steps when weights are applied. The consistency index of the trees in the unweighted analysis is 0.97 (including autapomorphies, 0.92 without); in the analysis with weights applied the CI is 0.97 (including autapomorphies, 0.93 without). Values given on the upper tree (Fig. 4a) pertain to the unweighted analysis, values on the lower tree (Fig. 4b) to the weighted analysis. The number of character changes is given above the branches and bootstrap values are given below. Arrows indicate nodes that collapse in the strict consensus of the shortest trees. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Strict consensus tree of the two equally parsimonious trees from the analysis of Matrix 3 (Fig. 4) showing character changes. Number and type of each change are indicated by e = endpoint, IV = inversion, IS = insertion \>5 kb, T = transposition, and D = deletion/divergence. Three of the 84 characters (18, 31 and 37) exhibit homoplasy; each character is an endpoint that changes twice within the tree. Homoplastic changes are shown by the character number below the branch upon which the change occurs. To the right of the tree, brackets and letter/number designations indicate major clades discussed in the text. Symbols at the ends of the terminal branches indicate the geographic distribution of the taxa: □ = eastern Asia; ● = Europe (includes North Africa); ◆ = Americas (primarily North America); \* = Southern Hemisphere (mainly Southern Africa). :::  ::: All results (Figs. [2](#F2){ref-type="fig"},[3](#F3){ref-type="fig"},[4](#F4){ref-type="fig"},[5](#F5){ref-type="fig"}) indicate that *Codonopsis*, *Platycodon*, and *Cyananthus*are basal within the family. Analyses on Matrix 2 and 3 support a *Codonopsis*+ *Cyananthus*sister group relationship and a monophyletic basal clade whereas the Matrix 1 analysis supports a *Codonopsis*+ *Platycodon*sister group and a paraphyletic basal grade. Neither outcome is very well supported; the alternative scenarios each require only a single additional step in the other data set. Within the fifteen derived taxa some of the relationships are not resolved or resolved but weakly supported. However, some groupings are well supported in all analyses. The South African taxa, *Merciera*, *Prismatocarpus*and *Roella*, form a clade (BS = bootstrap value = 98 - 100). *Wahlenbergia*is the sister to these three taxa in all analyses with varying levels of support (BS = 82, 60, 69, 99, 99, in the five analyses based on gene order changes). Other groupings include a *Symphyandra*+ *Edraianthus*clade (BS = 86-91) and *Legousia*+ *Asyneuma*+ *Petromarula*+ *Triodanis*(BS = 94-100). The five analyses had somewhat different characteristics (Table [4](#T4){ref-type="table"}). For example, Matrix 1 and Matrix 3 analyses generated fewer equally-parsimonious trees than Matrix 2. The Matrix 1 analyses resolved the most nodes. Matrix 3 analyses exhibited the lowest amounts of homoplasy and supported the highest number of nodes BS ≥ 90. Comparing all results, no nodes with high bootstrap values (BS ≥ 90) were conflicted by other nodes of equally high value. However, there were three instances of incongruence involving nodes of lesser support -- Matrix 1 and 3 analyses supported *Campanula*+ *Adenophora*, whereas Matrix 2 supported *Adenophora*+ *Jasione*; Matrix 1 supported *Codonopsis*+ *Platycodon*(BS = 50), whereas Matrix 2 and 3 supported *Codonopsis*+ *Cyananthus*(BS = 94-99); and Matrix 1 supported (weakly, BS = 57) the placement of *Cyananthus*at the base of the derived clade, whereas Matrix 2 and 3 analyses supported the monophyly of the basal group (BS = 56-68). One clade, *Legousia*+ *Asyneuma*(BS = 87), was recovered only by the Matrix 1 analysis within a clade not further resolved by the other gene order analyses. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Comparison of characteristics from the different analyses, best values for each characteristic shown in bold. ::: Matrix 1 Matrix 2, no weights Matrix 2 weighted Matrix 3, no weights Matrix 3 weighted *rbcL* ITS\* ---------------------------------------------------------- ----------- ---------------------- ------------------- ---------------------- ------------------- ----------- ----------- Number of MP trees 6 241 12 2 2 9 **1** Number of resolved nodes in consensus of MP trees 10 6 10 9 9 **14** 13 Nodes retained in consensus of all trees to 1% longer 6 5 5 6 **7** 5 2 CI (with / without autapomorphies) 0.81/0.67 0.90/0.79 0.93/0.82 0.97/0.93 **0.97/0.94** 0.77/0.66 0.69/0.60 Number of nodes BS ≥ 50 8 7 7 9 9 **13** 10 Number of nodes BS ≥ 90 3 5 4 **6** 5 5 **6** Average bootstrap value of resolved nodes 72 **91** 72 85 86 78 74 Number characters (PI) 37 31 34 116 **195** Total homoplastic characters (number/percent of PI) 15/40.5% 8/25.8% **3/8.8%** 63/54.3% 141/71.9% Homoplastic characters with one excess change 12/32.4% 7/22.6% **3/8.8%** 48/41.4% 82/41.8% Homoplastic characters with two excess changes 3/8.1% 1/3.2% **0** 14/12.1% 41/20.9% Homoplastic characters with three or more excess changes **0** **0** **0** 1/0.1% 18/9.2% \*the ITS analysis includes 3 fewer taxa than the others ::: We included sequence data here mainly to allow for a comparison with gene order data in terms of phylogenetic utility. The *rbcL*data from the same eighteen taxa (Table [5](#T5){ref-type="table"}) provided 116 parsimony-informative characters that, when analyzed, yielded nine shortest equally-parsimonious trees of 338 steps (C = 0.77/0.66). The strict consensus of the nine trees retained fourteen nodes (fig. [6a](#F6){ref-type="fig"}), thirteen of which had BS ≥ 50 and four of which were supported BS ≥ 90. The ITS data of Eddie et al \[[@B11]\] from taxa equivalent to fifteen of the eighteen mapped taxa (Table [5](#T5){ref-type="table"}) provided 196 parsimony-informative characters from which a single most parsimonious tree of 716 steps (fig. [6b](#F6){ref-type="fig"}) was generated (CI = 0.69/0.60). The tree contains thirteen resolved nodes of which ten had BS ≥ 50 and four had BS ≥ 90. The two sequence data sets had lower CI values than any of the gene order analyses and a higher percentage of homoplastic characters (Table [4](#T4){ref-type="table"}). The ITS data had especially high levels of homoplasy; the ITS data had a higher percentage of characters that change three or more times in excess than the Matrix 3 analyses had for total homoplastic characters (Table [4](#T4){ref-type="table"}). In the Matrix 3 analysis only three characters (endpoints) are required to change more than once over the most parsimonious tree; each has one excess change. With the inclusion of the sequence-based analyses, there were additional instances of incongruence between weakly supported nodes: 1) The placement of *Musschia*and *Jasione*varies between the *rbcL*and ITS results (the placement of these taxa is largely unresolved by the gene order data); 2) In both sequence-based trees, *Campanula*and *Adenophora*are separate lineages (rather than sister taxa) basal to the *Legousia*-*Asyneuma*-*Triodanis*-*Petromarula*clade, whereas in the gene order analyses they are allied to *Symphyandra*-*Edraianthus*, and *Trachelium*; 3) Matrix 1 supports a *Legousia*-*Asyneuma*clade, whereas a *Legousia*-*Triodanis*clade occurs in the sequence-based trees; and 4) Matrix 1 and ITS support a *Codonopsis*-*Platycodon*grouping within the basal clade, whereas *rbcL*and Matrix 2 and 3 analyses support *Codonopsis*-*Cyananathus.*Among these instances of disagreement between weakly supported nodes, there is no general pattern of disagreement between the sequence data and the gene order analyses. And among strongly supported nodes, again, there is complete agreement, among all analyses-sequence and gene order. ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Taxa for which *rbcL*and ITS data were analyzed. ::: **Gene Order/rbcL Species** **GenBank Accession*rbcL*** **ITS \"equivalent\" taxon** **GenBank Accession ITS \[11\]** ----------------------------- ----------------------------- ------------------------------ ---------------------------------- *Lobelia cardinalis* AY655144 *Lobelia tenera* AF054938 *Adenophora confusa* AY655145 *Adenophora divaricata1* AY322005 & AY331418 *Asyneuma virgatum* AY655146 *Asyneuma japonica* AF183437 & AF18343 *Campanula elatines* AY655147 *Campanula lusitanica* AY322025 & AY331438 *Codonopsis viridis* AY655148 *Codonopsis lanceolata* AY322048 & AY331461 *Cyananthus lobatus* L18795 \[40\] *Cyananthus lobatus* AY322050 & AY331463 *Edraianthus* AY655150 *Edraianthus* AY322052 & AY331465 *graminifolius* *graminifolius* *Jasione heldreichii* AY655151 *Jasione crispa* AY322059 & AY331472 *Legousia falcata* AY655151 *Legousia speculum-* AY322065 & AY331478 *veneris* *Merciera tenuifolia* AY655153 No equivalent NA *Musschia aurea* AY655154 *Musschia aurea* AY322067 & AY331481 *Petromarula pinnata* AY655155 *Petromarula pinnata* AY322069 & AY331482 *Platycodon grandiflorus* AY655156 *Platycodon grandiflorus* AY322074 & AY331487 *Prismatocarpus diffusus* AY655157 No equivalent NA *Roella ciliata* AY655158 *Roella ciliata* AY322074 & AY331487 *Symphyandra hofmanni* AY655159 *Symphyandra hofmanni* AY322076 & AY331489 *Trachelium caeruleum* L18793 40 *Trachelium caeruleum* AY322078 & AY331491 *Triodanis perfoliata* AY655160 *Triodanis leptocarpa* AY322079 & AY331492 *Wahlenbergia gloriosa* AY655161 No equivalent NA ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### Trees obtained from sequence data. The values below nodes are bootstrap percentages; the values above the nodes indicate the number of changes that occur on that branch. Fig. 6a. The one shortest tree (716 steps) based on ITS sequence data. The CI is 0.69 with autapomorphies and 0.60 without. Fig. 6b. One of the nine shortest trees (338 steps) based on *rbcL*sequence data, CI = 0.77/0.66. The node that collapses in the strict consensus of the nine trees is marked with an arrow. :::  ::: Discussion ========== Phylogenetic analysis of cpDNA rearrangements --------------------------------------------- The relatively large number of gene order mutations that have occurred in the Campanulaceae chloroplast genomes causes difficulties when interpreting their phylogenetic significance. The phylogenetic analysis of such a complex set of cpDNA rearrangements within a group of plants is without precedent. The first problem was simply defining individual mutational events. Although the ideal way to analyze rearrangement data is to determine presence or absence of specific events, in the Campanulaceae, this was not possible in many cases given our present knowledge. Where multiple overlapping rearrangements have occurred between genomes, the two specific endpoints that define a particular inversion may not be determinable. Because of the inherent complexity of the data, we felt a new method of character analysis of the rearrangement data was warranted. Our approach involved coding endpoints, along with more easily defined rearrangements, as characters for different cladistic analyses. Endpoints were defined as two non-contiguous tobacco regions that are now adjacent in the genomes of one or more species. Using endpoints as characters is advantageous. It allows for the incorporation into the analysis of data that could not be used if only unambiguously interpreted events were included. However, using endpoints as characters has several drawbacks, including the inadvertent weighting of certain events over others. Inversions necessarily produce two endpoints, and transpositions three, whereas gene losses and IR boundary changes produce a single endpoint. Therefore inversions would be included twice and transpositions three times if scored as \"independent\" endpoints rather than events. Plus, if both endpoints of an inversion are still intact in a genome, the inversion is scored twice, if only a single endpoint remains the inversion is counted only once, and if both endpoints have been lost (through further mutation) the inversion will not be included at all. This may represent a problem in the Campanulaceae analyses because there appears to be a mixture of event types and at least some endpoint reuse \[[@B18]\]. Our inclusion of transposition as a possible mechanism for gene order mutation in the Campanulaceae chloroplast genomes is problematic. Definitive evidence supporting the occurrence of transposition in the plastid genome is lacking. Transposition has been invoked to explain chloroplast DNA rearrangements, for example in \"subclover\" \[[@B30]\] and wheat \[[@B31],[@B32]\]. In these cases, transposition has been supported using parsimony arguments (one transposition explaining a change with fewer steps than three inversions) or using the existence of inverted- and direct-repeat sequence motifs near the boundaries of rearrangements \[[@B33]\]. In Campanulaceae, some lines of evidence in addition to parsimony suggest the possibility of transposition as a mechanism. First, the abundance of rearrangement events within the family suggests some mechanism that facilitates gene order mutation; transposition is one such process. Second, the segment of the genome defined by tobacco probes 53--56 is now located, in most of the derived taxa, within the inverted repeat. The region from which it has been removed appears otherwise undisturbed. In *Asyneuma*, the 53--56 region has been secondarily removed from the IR and returned to near its original location leaving behind small portions of 53 and 56 in the IR, detectable using southern hybridization. In *Wahlenbergia*, *Merciera*, *Prismatocarpus*and *Roella*, the 53, 54 portion of the 53--56 block has moved from the IR back to the LSC. One explanation for the high level of rearrangement apparently associated with this segment is that the region contains a transposable element. Third, a possible duplicative transposition is suggested (Fig. [1](#F1){ref-type="fig"}) in *Trachelium*\[[@B18]\] and five other taxa \[[@B9]\]. In addition to a full-length (presumably functional) copy of the 23S rRNA gene, a partial copy is located within *ycf1*. Transposition is one manner in which segments of DNA can be both copied and moved within a genome. None of our data are definitive. The observed rearrangements could have taken place as the result of multiple inversions. Therefore, it is important to note that if transposition is not active in the Campanulaceae genome, our phylogenetic results will not be greatly affected. Events coded in Matrix 2 and 3 as single transpositions would be underweighted inversions if incorrectly interpreted. The fact that the analysis of Matrix 1 yields results compatible with those of the matrices that include transpositions suggests that, if our interpretation is erroneous, it does not affect the phylogenetic conclusions. Our three methods of character scoring did yield largely compatible results in our analyses. Relationships that were strongly supported in one analysis were found in all analyses. Events make more desirable characters but they will only improve analyses if the postulated events are the correct ones. Comparing analyses that include event interpretations with endpoint only analyses is one way to determine the phylogenetic effects of the hypotheses of events used. Endpoint only analyses also allow studies that minimize a priori assumptions about the evolutionary events. It is possible that more complex evolutionary scenarios occurred, in which some inversions evolved in parallel, or in which similar gene orders resulted from a different set of inversions. The parsimony analyses may underestimate the number of inversions shared between primitive and advanced genera, because evidence of shared inversions may have been lost. Although we have attempted to produce the simplest evolutionary schemes, it is very possible that longer, more complicated scenarios actually occurred, especially given that the Campanulaceae seem predisposed to cpDNA rearrangements. However, given the congruence of the results among our various analyses, we feel our phylogeny is a reasonable estimate of relationships within the family. Elsewhere, we have analyzed a reduced subset of characters and taxa for the Campanulaceae cpDNA data set using endpoint scoring and constructing trees using breakpoint distances among other methods (e.g., \[[@B34]\]). Other computational biologists have also used this reduced data matrix to test different methods of phylogeny reconstruction based on gene order data (e.g., \[[@B35],[@B36]\]). These various studies produced trees that are largely congruent with those generated in this paper suggesting that the Campanulaceae cpDNA gene order data are providing a consistent estimate of phylogenetic relationships given any logical method of scoring and analysis. Although the rearrangements in Campanulaceae are complex, the phylogenetic utility of the gene order data is evident. In most previous examples of phylogenetic use of rearrangements the small number of events allowed for the circumscription of only very broad groups \[[@B15]\]. Because there are so many rearrangements in the Campanulaceae, smaller groups can be identified. This has resulted in the most highly resolved phylogeny as yet developed based entirely on cpDNA rearrangements. Not only do these data support a well-resolved phylogeny but they provide robust support of several nodes. Matrix 3 supports as many nodes at BS ≥ 90 as ITS and more than *rbcL*. In matrix 3, only three endpoint characters (8.8% of parsimony-informative characters) are homoplastic, each changing one extra time over the tree. In contrast, within ITS, 18 characters (9.2% of parsimony-informative characters) change three or more extra times over the tree, and 71.9% of characters are homoplastic. Presumably because of this high level of homoplasy, only two nodes are retained in the consensus of all ITS trees from the shortest to 1% longer, whereas seven nodes are retained in matrix 3 trees \"to 1% longer\" -- the highest number of any of the analyses. Matrix 3, the matrix in which characters are most interpreted as mutational events, is especially strong in its performance, exceeding both sequence data sets in average bootstrap value per resolved node and CI (in addition to those characteristics just discussed). This suggests that the closer we can get to scoring the actual mutations the stronger gene order data will perform. Although the endpoint only matrix provides useful insights on relationships, we would argue that the extent to which these gene order characters cannot recover the phylogeny is directly related to our ability to define individual mutational events. Phylogenetic implications of the rearrangement data --------------------------------------------------- Most traditional classifications of the Campanulaceae are based mainly on capsule dehiscence and ovary position and arrangement. As Kovanda \[[@B1]\] and Thulin \[[@B8]\] recognized, classification of the Campanulaceae based on capsule characters alone brings together otherwise radically different taxa. Neither the Campanuleae nor Wahlenbergieae (at whatever taxonomic rank) are monophyletic based on cpDNA rearrangements (Fig. [5](#F5){ref-type="fig"}). Likewise, no traditional classification (Table [1](#T1){ref-type="table"}) suggests that *Codonopsis*, *Platycodon*, and *Cyananthus*are basal in the family as supported by both gene order and sequence data. Takhtajan\'s system \[[@B2]\] is something of an exception among traditional classifications; however, he suggested only *Cyananthus*(in its own tribe Cyanantheae) as the most primitive member of the family, placing *Platycodon*and *Codonopsis*in other tribes (Table [1](#T1){ref-type="table"}). In contrast, studies of pollen ultrastructure have indicated that *Platycodon*, *Codonopsis*, and *Cyananthus*are basal members of Campanulaceae \[[@B37],[@B38]\]. These taxa have colpate to colporate apertures, whereas the remaining family members (as surveyed here) have porate grains \[[@B37]-[@B41]\]. The evolutionary scheme based on pollen morphology presented by Dunbar \[[@B38]\] suggests that *Cyananthus*(colpate) and *Codonopsis*(colpate) are more closely related to each other than either is to *Platycodon*(colporate), which is also supported by the gene order tree (Clade B, Fig. [5](#F5){ref-type="fig"}). Thulin \[[@B8]\] believed that pollen morphology should constitute a key part of any modern reassessment of relationships in the Campanulaceae. He suggested that all taxa with elongated apertures should be removed from Campanuleae and Wahlenbergieae, and those with porate grains removed from Schönland\'s Platycodinae. Following the removal of colpate and colporate taxa, Campanuleae *sensu*Schönland are comprised of Northern Hemisphere genera, whereas Wahlenbergieae contain Southern Hemisphere taxa, with the exceptions of *Edraianthus*and *Jasione*(although *Jasione*occurs in North Africa as well as Europe). The gene order data indicate that the affinities of *Jasione*and *Edraianthus*lie with Northern Hemisphere species rather than with Wahlenbergieae. The gene order data also are compatible with other available nucleotide data in addition to those reported here \[\[[@B10],[@B11],[@B42]\], L. Raubeson, A. Oestriech and R. Jansen, unpublished data\], a morphology-based cladistic study \[[@B10]\] and are also largely congruent with a serological study of the Campanulaceae \[[@B43]\]. Although the gene order and serological studies differed somewhat in the taxa sampled, both included a group containing *Trachelium*and *Campanula*. They also agreed in the grouping of *Asyneuma*and *Petromarula*. The only discrepancy was in the placement of *Legousia*; the serological study placed this genus basal to all others surveyed \[[@B43]\]. The groups delimited by cpDNA rearrangements also exhibit geographical integrity. *Wahlenbergia*is primarily a Southern Hemisphere Old World genus \[[@B44]\]; *W. gloriosa*, mapped for this study, is Australian \[[@B44]\]. *Roella*, *Merciera*, and *Prismatocarpus*are all endemic to South Africa \[[@B45]-[@B47]\]. The nine genera in the *Trachelium*and *Legousia*clades are primarily European to Eurasian, although *Triodanis*is endemic to North America and *Campanula*has a few North American representatives \[[@B5],[@B48]-[@B50]\]. *Musschia*is endemic to the island of Madeira \[[@B51]\]. There has been considerable debate regarding the relationships among the four centers of taxonomic diversity of the Campanulaceae: Asia, Europe (especially the Mediterranean), South Africa, and western North America. Bentham \[[@B52]\] hypothesized a northern origin for Campanulaceae but he did not specify a particular region. Takhtajan \[[@B2]\] suggested a basal position of the Asian genus *Cyananthus*. Studies of pollen ultrastructure indicated that the Asian genera *Codonopsis*, *Cyananthus*, and *Platycodon*are basal members of the Campanulaceae \[[@B37],[@B38]\]. Recent studies of the Campanulales \[[@B42],[@B53],[@B54]\] indicate that the order consists of several families, including the Campanulaceae, Cyphiaceae, Cyphocarpaceae, Lobeliaceae, Nemacladaceae, and Stylidaceae. Several of these families are restricted to the Southern Hemisphere (all but Nemacladaceae from North America and Campanulaceae which is cosmopolitan), implying that the Southern Hemisphere may be the ancestral area for the Campanulaceae \[[@B54]\]. Phylogenies based on *rbcL*sequence data position the Campanulaceae sister to the North American family Nemacladaceae \[[@B42],[@B54]\]. Our cpDNA phylogeny based on genome rearrangements (Fig. [5](#F5){ref-type="fig"}) provides strong support for the basal position of the three examined Asian platycodonoid genera, suggesting that the early radiation of the family may have occurred in Asia rather than Africa. The genera from the Southern Hemisphere (*Merciera*, *Prismatocarpus*, *Roella*, and *Wahlenbergia*) are in a much more derived position in the cpDNA tree. In addition, the gene order data suggest affinities of several controversial genera (Fig. [5](#F5){ref-type="fig"}). Schönland \[[@B48]\] united *Musschia*and *Platycodon*as Platycodinae, clearly incompatible with both our results and pollen evidence. *Musschia*is placed in the derived clade (A), although its exact placement varies among all the analyses, including *rbcL*and ITS. De Candolle \[[@B3]\] was unsure of *Merciera*\'s taxonomic position because its four basal ovules and single-seeded (by abortion) unilocular capsule \[[@B43]\] are unique in the Campanulaceae \[[@B55]\]. This genus was later recognized as a separate tribe, Merciereae \[[@B56]\], but is allied with other southern African genera in the cpDNA analysis (Fig. [5](#F5){ref-type="fig"}). Takhtajan \[[@B2]\] placed *Merciera*with *Wahlenbergia*, *Roella*and *Prismatocarpus*in his Wahlenbergieae but also included other genera forming a polyphyletic group according to our results. *Adenophora*and *Symphyandra*have been segregated from *Campanula*based on the presence of a conspicuous tubular nectariferous disc and connate anthers, respectively. *Adenophora*and *Campanula*are sister taxa in the gene order analyses (except those based on Matrix 2) and *Adenophora\'*s chloroplast genome is derived relative to *Campanula\'*s (Fig. [5](#F5){ref-type="fig"}). Further sampling within *Adenophora*and the large genus *Campanula*will be necessary to determine if this is a general result. *Symphyandra*is more closely related to *Edraianthus*than *Campanula*but all are within the A3 Clade (Fig. [5](#F5){ref-type="fig"}). *Edraianthus*has traditionally been considered close to *Wahlenbergia*\[[@B3]\] but this is not supported by any of the results reported here or by morphological studies of Hilliard and Burtt \[[@B57]\]. Much controversy surrounds the taxonomy of the genera *Triodanis*and *Legousia*. In some treatments, both genera were included under the illegitimate name *Specularia*(e.g. \[[@B3],[@B48]\]). McVaugh \[[@B58],[@B59]\] and Fernald \[[@B60]\] disagreed regarding the circumscription of the genera; Fernald felt that *Triodanis*as a genus is very weak and should be merged with *Legousia*. McVaugh \[[@B58]\] argued that the two genera should either remain separate or both be subsumed into *Campanula*. In his system, both species studied here (*T. perfoliata*and *L. falcata*) belong to *Triodanis*. As expected, *Triodanis*and *Legousia*belong to the same cpDNA clade (A2), united by an unusual mutation that transferred a large segment of the large single copy (LSC) region to the SSC region \[[@B9]\]. However, *Legousia*has a putative transposition not found in *Triodanis*, whereas *Triodanis*has a unique large insertion \[[@B9]\]. Conclusions =========== Despite the difficulties in interpreting such a complex set of rearrangements, the systematic utility of chloroplast DNA in the Campanulaceae is evident. Our results support the division of the family into two groups previously unrecognized in any taxonomic treatment. In addition, numerous groupings within the larger, derived clade are strongly supported. The data indicate that traditional classifications based on fruit and ovary characters are unnatural, and suggest affinities of several difficult genera. Additional sampling within large genera, such as *Campanula*and *Wahlenbergia*, will be necessary to fully elucidate relationships among chloroplast genomes. It is likely that intrafamilial relationships can be further resolved by including other genera in rearrangement analyses. Although homoplasy is not absent in our data, it is low and considerably lower than some sequence data such as ITS. Although any reasonable scoring method for the gene order data generates results that are largely compatible among the different analyses, the more that the gene order data can be interpreted as actual mutational events (and the presence or absence of those events used as characters) the stronger will be the results. Even in cases such as this, with high levels of gene order complexity, cpDNA gene order mutations make excellent phylogenetic markers. Methods ======= Total DNA was isolated from one species in each of 18 genera in the Campanulaceae (Table [2](#T2){ref-type="table"}) according to the CTAB method of Doyle and Doyle \[[@B61]\]. DNAs were digested with the restriction endonucleases *Bam*HI, *Bgl*II, *Eco*RI, *Eco*RV, *Hind*III, and *Sst*I, and double digests were carried out using *Hind*III and the remaining five enzymes. Hybridization probes consisted of 106 small tobacco cpDNA probes (average size 1.2 kb) provided by J. Palmer \[[@B62]\]. Twenty-one cloned *Hind*III cpDNA fragments from *Trachelium caeruleum*of the Campanulaceae were also used as hybridization probes \[[@B18]\]. Complete single and double digest restriction site maps were constructed for 16 of the 18 taxa, and nearly complete maps were constructed for the remaining two taxa, *Jasione*and *Roella*\[[@B9]\]. It was not possible to map the small single copy (SSC) region of *Roella*because hybridization signals became increasingly weak in later rounds of hybridization. In *Jasione*, rearrangements involving the IR/SSC junction and SSC region prohibited the complete resolution of the map. The restriction site maps were then interpreted as linear \"number\" maps representing the hybridization patterns of 106 consecutively numbered tobacco cpDNA probes for the 18 taxa (Fig. [1](#F1){ref-type="fig"}). Rearrangements were recognized as any change in the order of gene segments relative to the order observed in tobacco. The recognition of such disruptions is straightforward; the interpretation of the disruptions as actual mutational events can be quite complicated. As a hypothetical example, the ancestral order in a region may be 1-2-3-4-5-6; while the order 1-2-5-3-4-6 may be observed in a rearranged genome. In the rearranged genome 2-5, 5-3, and 4-6 are adjacencies that are derived relative to the ancestral order. But what set of events is responsible for the change? A simple transposition of 5 to the position between 2 and 3 can account for the difference in a single event. Alternatively, two inversions with one shared endpoint may be responsible or two inversions with unique endpoints followed by a transposition can explain the differences. Additional explanations would also be compatible with these data. On what basis do we choose among multiple scenarios? As an actual example, the chloroplast genome of *Platycodon*could have evolved from a tobacco-like ancestor by two different models each involving seven inversions (Fig. [7](#F7){ref-type="fig"}); not one inversion is common to the two scenarios. Thus in our initial approach to data analysis (generating Matrix 1) we did not define events, but utilized endpoints only. In the hypothetical example 2-5, 5-3, and 4-6 are \"endpoints\" -derived adjacencies absent in the ancestral gene order. Taxa with genomes that exhibit the derived adjacencies are coded as 1 for those characters and those with the ancestral condition as 0. ::: {#F7 .fig} Figure 7 ::: {.caption} ###### Two alternative models of seven inversions each to explain the evolution of *Platycodon*cpDNA structure from a tobacco-like ancestor. Numbers in parentheses show order of hybridized tobacco cpDNA probes. Inversion endpoints are shown by the arrows. Locations of regions represented by probes 6-7, 8, and 9 are believed to be the result of transposition \[14\]; these events are required in addition to the inversions to completely explain the new gene order. :::  ::: We constructed two additional matrices that did include events since some endpoints (or combination of endpoints) seemed readily interpretable. For example, if a region of the genome was simply reversed in order (i.e., 1-4-3-2-5 relative to 1-2-3-4-5) we assumed that an inversion had taken place to result in the different arrangements of gene segments. Likewise if genomes differed in content of the IR, we assumed that single duplication or loss events were responsible. Making such inferences, we constructed Matrix 2 that is composed of 31 events and 53 endpoints. We then went further and constructed hypotheses of rearrangement events to account for the differences among the genomes of the three major clades delimited among the fifteen derived taxa \[[@B9]\]. If these scenarios indicated that an inversion likely was shared between two or more genera, the taxa were coded as having an endpoint even if the endpoint has been lost due to disruption by subsequent events. We were conservative in our application of this approach and only six endpoint scorings were modified in Matrix 3 compared to Matrix 2. To summarize, we produced three data matrices that represented increasing levels of interpretation of endpoints as actual events. Cladistic analyses were performed on each of the three data sets using equal weighting of all included characters. The second and third matrices were also analyzed giving weights of two to all non-endpoint characters. This weighting represents an attempt to compensate for the unintentional weight given to inversions in which both endpoints are present. This, of course, results in down-weighting inversions in which only one endpoint remains and fails to include inversions whose endpoints are both absent. Finally, to allow for a direct comparison of performance between gene order data and sequence data over the same taxa, we conducted maximum parsimony analyses of ITS and *rbcL*data. The ITS sequences were generated and aligned by Eddie et al \[[@B11]\]. We determined taxa equal or equivalent to our taxa and performed analyses on just those taxa from the Eddie matrix (Table [5](#T5){ref-type="table"}); only fifteen of our eighteen taxa were represented. We generated *rbcL*data to add to the two taxa already available \[[@B42]\] so that we had *rbcL*sequence data from all of the eighteen mapped taxa. Exactly the same DNAs were used. In generating the *rbcL*sequences, we PCR-amplified about 1370 bp of the gene in 50 μl reactions containing: 1 μl unquantified total genomic DNA, 0.2 mM each dNTP, 2.5 mM MgCl2, 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 0.4 μM each primer, and 1 unit Taq polymerase. Cycling conditions were as follows: 1 95°C denaturation step for 3 minutes 30 seconds, 30 cycles of 1 minute at 95°C, 1 minute at 55°C, and 1 minute 30 seconds at 72°C, and finally a 7 minute 72°C step. The PCR primers plus two internal primers were used for sequencing; the forward amplification primer and two internal primers were designed by G. Zurawski (his Z-1, Z-427 and Z-895). The Zurawski primer commonly used as the reverse amplification primer did not work in many Campanulaceae; we designed an alternative: 5\'-GTATCCATTGCGCAAACTC-3\'. For sequencing, two successful PCR reactions were combined and then cleaned (and concentrated) using the Qiagen QIAquick PCR Purification Kit (catalog number 28104). Depending on the concentration of the recovered product, 0.5--2 μl of this template was cycle sequenced and resolved on an ABI Prism 377 Automatic DNA Sequencer. Electropherograms were inspected, and then sequences were edited and assembled using Sequencher, vers 3.1 (Gene Codes Corp.) The sequences have been deposited in GenBank (accession numbers in Table [5](#T5){ref-type="table"}). Alignment was performed by Sequencher and adjusted manually. Alignment of the *rbcL*sequences was very straightforward. For all parsimony analyses, searches were conducted using the branch and bound algorithm in PAUP\* 4.0b10 (PPC) \[[@B63]\]. Tobacco was used as the outgroup for the gene order data since it has the ancestral chloroplast genome gene order for the angiosperms \[[@B15],[@B26]\] and *Lobelia*was used as the outgroup for the sequence data. See the [additional file - data file 1](#S1){ref-type="supplementary-material"} -- for the Nexus file used in the PAUP analyses. This file includes the three gene order matrices and the *rbcL*alignment. The ITS alignment of Eddie et al \[[@B11]\] is available online \[[@B64]\]. The strength of the support, in each data set, for monophyletic groups was evaluated by calculating bootstrap values \[[@B65]\] using 10,000 heuristic (TBR, multrees option) replicates. In addition, for each matrix, analyses were performed to generate all trees from the shortest to one percent longer. We used a percentage rather than an equal number of steps in an attempt to make an equivalent comparison among the different sized data sets. A consensus of these trees was determined and the number of nodes retained \"to 1% longer\" was calculated. Authors\' contribution ====================== MEC performed the DNA isolations and Southern hybridizations, mapped the genomes, performed character codings and analyses for Matrix 2 and Matrix 3, and wrote up her work as a chapter of her Ph.D. thesis. LAR confirmed genome maps and character codings, updated MEC\'s analyses, added MATRIX 1 and its analysis, generated the *rbcL*sequence data, analysed ITS and *rbcL*data, and modified the thesis chapter for publication. RKJ assisted in all aspects of the work. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 One additional file is provided -- data file 1. This file is in NEXUS format and includes data for Matrix 1, Matrix 2, and Matrix 3 encodings and the *rbcL*data, concatenated for each taxon. PAUP statements in the file identify each individual data matrix and the characters in Matrix 2 and 3 that are weighted in some analyses. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ This work was supported by National Science Foundation grants DEB 9101239 to MEC, DEB 9982092 to RKJ, RUI/DEB 0075700 to LAR and DEB0120709 to RKJ and LAR. The CWU Faculty Research and Development Committee also supported LAR\'s work on this project through a Faculty Research Leave. We thank Tina Ayers for material, Jeff Palmer for tobacco probes, and Bill Eddie, Tom Lammers and two anonymous reviewers (especially J. Palmer) for helpful suggestions on improving the manuscript. Darlene Boykiw and Gwen Gage assisted in the preparation of tables and figures, respectively. This paper represents a portion of a dissertation submitted to the Graduate School of The Ohio State University by MEC under the supervision of Dan Crawford.

PubMed Central

2024-06-05T03:55:47.419093

2004-8-23

PMC516027

Background ========== Endogenous reference alse referred to as house keeping genes defines in biology the theoretical assumption that certain genes are ubiquitously expressed in nucleated cells possibly because their stable expression is essential for cell survival and welfare in all physio-pathological circumstances. In practical terms, endogenous reference genes provide a useful constant reference to normalize the expression of test genes in different tissues and in different conditions. This is obviously important when estimates of gene expression are provided in relative terms rather than absolute units of measurement. Thus, endogenous reference genes are used as common denominator in biological fractions where the expression of a test gene is described as the relative ratio over an arbitrarily selected internal control presumed to be stably expressed in all circumstances relevant to the experiment \[[@B1]-[@B3]\]. Most frequently, glyceraldehydes-3-phosphate dehydrogenase (GAPDH) \[[@B4],[@B5]\], albumin (for hepatocytes) \[[@B6]\], β-, γ-actins \[[@B7],[@B8]\], cyclophilin \[[@B9],[@B10]\], α-, β-tubulins \[[@B7],[@B11]\], hypoxantine phosphoribosyltransferase (HRPT) \[[@B12],[@B13]\], L32 \[[@B14],[@B15]\] and 18S, 28S ribosomal RNA (rRNA) \[[@B16]-[@B18]\] have been used as endogenous reference genes. Depending upon the experimental design, endogenous reference genes have been used individually or in combination for Northern blot analysis, reverse transcription polymerase chain reaction (RT-PCR) and quantitative real-time PCR (qRT-PCR) analysis \[[@B19],[@B20]\]. With the development cDNA microarray technology endogenous reference genes have been used for array data normalization. However, accumulation of extensive data bases suggests that the expression of frequently used endogenous reference genes can vary substantially according to materials and conditions studied \[[@B1],[@B2],[@B6],[@B14],[@B17],[@B18],[@B20]-[@B27]\]. Powerful insights in patterns of gene expression could be attained recently through cDNA or oligonucleotide-based global transcript analysis tools that apply a constant reference system to determine ratios of gene expression across large data sets \[[@B28],[@B29]\]. The constant reference is provided for each gene in question by consistently co-hybridizing individual test samples with a differentially labeled reference sample maintained identical throughout all the hybridization experiments. Gene expression data are then expressed as the ratio of expression between test and reference samples for each gene. By keeping the reference sample identical the resulting ratio represents a precise estimate of the relative expression of each gene across the various conditions tested bypassing the need to normalize with endogenous reference genes. This holds true if the hybridization kinetics between test and reference sample are accurately reproducible. We will refer to this concept as \"reference concordance\" and in the results we will discuss how reference concordance was used to validate the reproducibility of cDNA array data from which putative candidate endogenous reference genes were identified. In the present study, we tested a set of 419 consecutive experiments performed on a 17,000 gene cDNA array platform to which RNA from neoplastic or normal tissues were consistently co-hybridized with a differentially-labeled reference RNA derived from peripheral blood mononuclear cells (PBMC) pooled from six normal donors. The following steps were pursued: 1) [Reproducibility assessment of the data set through determination of reference concordance]{.underline}. This was achieved by repeating 14 reference experiments using the melanoma cell line A375 as test sample (Cy5) co-hybridized with pooled PBMC as reference (Cy3). To test for inter-array and printing variation, slide number one and every other 25 slides in sequential order of printing (100 slides per printing set) were used for the repetitive A375 / pooled PBMC hybridizations. In addition, to assess labeling bias, reciprocal labeling was alternatively applied as previously described \[[@B30]\]. In this fashion a pool of genes expressed with high level of reference concordance was selected. 2) [Identification of putative endogenous reference genes]{.underline} was performed on 384 array experiments of unequivocal quality by selecting genes that had demonstrated high reference concordance (\>90% of the genes in the arrays) and ranking them from the lowest to the highest variance of ~Log2~test / reference ratios across all array experiments. 3) [Validation of the candidate endogenous reference genes as predictor of relative gene expression in large data series]{.underline}. For this purpose, we tested the relative estimates of expression of the melanocytic lineage-restricted melanoma differentiation antigen gp100/Pmel17 (gp100) \[[@B31]\] in melanocytic and non-melanocytic tissues. Estimates of expression of gp100 were compared after normalization with different endogenous reference genes. For this analysis, we used a new tool that spots cDNA libraries from different tissues on an array platform to allow high-throughput evaluation of individual gene expression across broad tissue collections. We termed this tool: \"transcriptome array\". 4) [Validation of endogenous reference gene-based normalization of gene expression by qRT-PCR and protein expression]{.underline}. Results ======= Data processing --------------- The GenePix Pro 4.0 software was used for array image analysis and calibration. All statistical analyses were performed with the SPLUS package <http://www.insightful.com>. Thirteen arrays with missing data in more than 30% of the spots and 8 arrays with irregular distribution according to M-A plots (M and A representing respectively log-ratios and average intensities) \[[@B32]\] were excluded. Spots in which \> 50% of the pixels reached saturation in either channel, flagged spots and spots with intensity ≤ 200 in one channel and ≤ 500 in the other channel were filtered out. If the intensity in one channel was \< 200 but that of the other channel was \> 500, we arbitrarily assigned an intensity of 200 to the channel with the lowest intensity. The ~log2~Cy5/Cy3 ratios were normalized by approximating median values to zero. Spots with ratios \> 100 or \< 1/100 were truncated at 100 and 1/100 respectively. Data normalization was done by median centering ~Log2~ratio of all genes for each array. For data normalization we used certain arbitrary cutoff criteria to remove spots with weak signal. Weak signal approximating background fluorescence is not reliable and the corresponding log ratios are often distorted resulting in disproportionately high or low values that would bias the statistical results. Because the interpretation of spots with low signal is difficult to make we have adopted the policy of excluding them to focus the analysis on genes whose expression pattern is more reliably tracked by the array tool. In regard to spots with a ratio of \>100 or \<-1/100, most of them are due to extremely weak signal in one channel which generates a disproportionately extreme value. Although there is no published \"gold standard\" for the selection of such cutoffs, in practice the parameters that we used for this study are most commonly accepted by investigators and reasonable because allow retention of most of the data in the array excluding only the most extreme and least reliable information. Validation of the data set through analysis of \"reference concordance\" ------------------------------------------------------------------------ To select from the data base the most reliable genes, we first assessed reproducibility by testing the level of concordance in 14 reference experiments. Reference concordance relies on the expectation that results obtained by repeatedly hybridizing the same test and reference material should perfectly collimate and the degree of deviation from such prediction estimates experimental variance. Concordance can be easily measured by periodically re-hybridizing the same test sample with the constant reference sample. We analyzed a matrix of 7 forward and 7 reciprocally-labeled replicate array experiments hybridized periodically every other 25 cDNA array slides. Reciprocal labeling was applied to measure labeling bias \[[@B30]\]. Data generated by reciprocal experiments were mathematically reversed into the same labeling direction for data analysis. Genes that were discordant due to labeling bias were identified by student\'s *t*test as those with ratios significantly different between the 7 forward and 7 reciprocal experiments after reversal of the reciprocal values (P \< 0.05) and the median ratio difference between the 7 forward and 7 reciprocal experiments was larger than 1.5 fold. The genes whose variances of the log ratio across all experiments were among the top 1 percentile of variance of all genes were identified as discordant due to hybridization bias. In addition, genes with more than 50% missing values were excluded. Overall 1,343 out of 16,738 were judged potentially discordant and were excluded from further analysis. The remaining 15,395 concordant genes were used for subsequent analyses. Endogenous reference gene identification ---------------------------------------- We studied 384 of 419 consecutive array experiments remaining after the exclusion of the 14 reference concordance experiments and 21 experiments judged of poor quality due to missing data or irregular distribution by M-A plot analysis. Selection of candidate endogenous reference genes followed these steps. First, the seven experiments used for the analysis of reference concordance in which labeling was done identically to the rest of the experimental samples (test labeled with Cy5 and reference with Cy3) were used to calculate experimental variance (Table [1](#T1){ref-type="table"}). Median of variances across the seven replicate arrays was calculated for all genes with average intensity in both channels \> 2,000. This parameter provides an estimate of the variance due to experimental variation (background variance) since theoretically no differences in gene expression should be noted by using the same material. Based on the assessment of the seven repeat experiments, we defined differential expression as \> 2 standard deviations (SD) from the mean of the 384 arrays which is equivalent to 1.46-fold change (± 0.549 in ~log2~ratio). Consequently, genes with fold changes \< 1.46 from zero across all 384 arrays were categorized as candidate endogenous reference genes and were ranked according to ascending values of SD of ~Log~ratio (Table [1](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 **384 experiments** **7 replicates** ------------------------ ----------------------- ------------------------- -------------------- --------------------- -------------------- --------------------- -------------------- cDNA FLJ40458 1571492 Hs.181346 2835 0.19 0.27 0.39 0.15 Unknown 301067 2571 0.1 0.27 0 0.21 [PTH]{.underline} [322051]{.underline} [Hs.37045]{.underline} [3380]{.underline} [0]{.underline} [0.28]{.underline} [-0.02]{.underline} [0.27]{.underline} KIAA1935 39938 Hs.300776 2359 0.16 0.28 0.01 0.19 ESTs 49313 Hs.395460 2079 0.15 0.28 0.08 0.03 cDNA FLJ30539 fis 49463 Hs.21489 2561 0.09 0.28 0.17 0.13 SDCCAG16 1576228 Hs.271926 2111 0.11 0.29 0.28 0.14 clone IMAGE:41799 278673 Hs.271721 3132 0.16 0.3 -0.04 0.1 PSIP2 289945 Hs.82110 2521 -0.07 0.3 -0.06 0.19 ETV2 1468722 Hs.194061 2646 0.09 0.31 0 0.28 ESTs 1571401 Hs.126999 2659 0.05 0.31 -0.04 0.08 [NEDD8]{.underline} [277660]{.underline} [Hs.75512]{.underline} [3875]{.underline} [0.08]{.underline} [0.31]{.underline} [0.1]{.underline} [0.2]{.underline} HIP2 486259 Hs.155485 2511 0.11 0.31 0.12 0.25 [HIRIP5]{.underline} [745314]{.underline} [Hs.430439]{.underline} [3407]{.underline} [-0.08]{.underline} [0.31]{.underline} [0.07]{.underline} [0.31]{.underline} ACTR8 156363 Hs.124219 2731 0.08 0.32 0.14 0.16 [GPLD1]{.underline} [293696]{.underline} [Hs.272529]{.underline} [3110]{.underline} [-0.03]{.underline} [0.32]{.underline} [-0.05]{.underline} [0.28]{.underline} KIAA0769 299128 Hs.19056 2148 0.14 0.32 0.05 0.2 CSNK1G2 346031 Hs.181390 2320 -0.03 0.32 -0.24 0.28 [RAB31]{.underline} [784150]{.underline} [Hs.223025]{.underline} [3479]{.underline} [-0.06]{.underline} [0.32]{.underline} [-0.36]{.underline} [0.36]{.underline} C14orf117 796100 Hs.103189 2008 0.12 0.32 -0.01 0.16 ATP6IP2 825077 Hs.183434 2159 0.02 0.32 0.28 0.32 [SNRPD3]{.underline} [897099]{.underline} [Hs.1575]{.underline} [3031]{.underline} [-0.04]{.underline} [0.32]{.underline} [-0.48]{.underline} [0.25]{.underline} Unknown 1292073 2745 -0.13 0.33 0.05 0.28 ESTs 22137 Hs.187406 3133 0.07 0.33 -0.01 0.21 ESTs 32782 Hs.443140 2464 -0.02 0.33 0.09 0.26 L3MBTL 43090 Hs.300863 2581 -0.08 0.33 -0.41 0.31 GNRH1 487071 Hs.82963 3294 0.05 0.33 0.23 0.23 clone 24629 mRNA 746258 Hs.142570 2645 0.05 0.33 -0.21 0.14 FLJ12998 826367 Hs.343627 2267 0.14 0.33 0.03 0.3 cDNA: FLJ23477 fis 1492329 Hs.145362 2142 0.09 0.34 0.35 0.34 HOXC4 1756945 Hs.50895 2581 0.03 0.34 -0.02 0.22 WSX1 1855887 Hs.132781 3259 -0.01 0.34 0.12 0.3 [HIST1H1A]{.underline} [1872543]{.underline} [Hs.150206]{.underline} [2523]{.underline} [0.07]{.underline} [0.34]{.underline} [-0.17]{.underline} [0.14]{.underline} TFAP2B 363144 Hs.33102 2510 0.05 0.34 0.67 0.35 TRIM31 509760 Hs.91096 2753 0.06 0.34 -0.08 0.2 FLJ00166 713191 Hs.43213 3134 0.08 0.34 0 0.22 [CDK5R1]{.underline} [757873]{.underline} [Hs.93597]{.underline} [2965]{.underline} [-0.12]{.underline} [0.34]{.underline} [-0.22]{.underline} [0.15]{.underline} PPP3CA 796730 Hs.272458 2605 -0.07 0.34 -0.36 0.16 PMS2L4 161373 Hs.278468 2534 -0.14 0.35 -0.21 0.22 DLGAP1 1758491 Hs.75814 2143 0.02 0.35 -0.24 0.2 ESTs 177884 Hs.14613 2054 0.01 0.35 -0.22 0.25 NRL 2364249 Hs.89606 2300 -0.04 0.35 -0.09 0.14 RAB36 281489 Hs.38772 3143 -0.08 0.35 0 0.12 RPC8 323603 Hs.353192 2367 -0.06 0.35 -0.33 0.24 ZNF177 33294 Hs.172979 2988 0.01 0.35 -0.21 0.25 LY6G5C 448136 Hs.246845 3591 -0.02 0.35 -0.44 0.28 RAD17 586844 Hs.16184 3797 -0.01 0.35 0.12 0.15 LOC253039 1033388 Hs.41181 3612 -0.02 0.36 0.31 0.08 IMPDH2 1582050 Hs.75432 2709 -0.01 0.36 -0.08 0.18 PRMT3 2074202 Hs.152337 2042 0.07 0.36 -0.06 0.21 [FZD6]{.underline} [214916]{.underline} [Hs.114218]{.underline} [3884]{.underline} [0.03]{.underline} [0.36]{.underline} [0.02]{.underline} [0.26]{.underline} AP3B2 47510 Hs.21022 2350 0.03 0.36 -0.23 0.31 NF2 769716 Hs.902 2887 -0.08 0.36 -0.11 0.31 SCP2 855395 Hs.75760 3218 -0.12 0.36 -0.1 0.19 [COL8A1]{.underline} [1472775]{.underline} [Hs.114599]{.underline} [4178]{.underline} [-0.06]{.underline} [0.37]{.underline} [0.05]{.underline} [0.18]{.underline} AKAP4 1643144 Hs.97633 2557 -0.03 0.37 -0.07 0.15 Unknown 1671546 3386 0.03 0.37 1 0.42 TCF8 178463 Hs.232068 2726 -0.04 0.37 -0.01 0.06 TNP2 1839038 Hs.2748 2737 0.08 0.37 0.11 0.24 PIWIL1 2329739 Hs.194712 2573 -0.08 0.37 -0.14 0.22 DBCCR1 47037 Hs.6090 2481 -0.01 0.37 -0.18 0.24 SLC21A3 289706 Hs.46440 2708 0 0.38 -0.11 0.27 ESTs 35105 Hs.403854 3126 0.03 0.38 0.23 0.18 cDNA FLJ34400 fis 511835 Hs.380035 3058 0 0.38 0.02 0.37 CNTN1 51640 Hs.143434 2479 -0.05 0.38 -0.11 0.3 TFCP2 843067 Hs.154970 3728 -0.06 0.38 0.16 0.34 C17orf35 510032 Hs.15196 2778 -0.02 0.39 0.09 0.28 [AFAP]{.underline} [488062]{.underline} [Hs.80306]{.underline} [3142]{.underline} [0]{.underline} [0.42]{.underline} [0.12]{.underline} [0.15]{.underline} MTP 731054 Hs.195799 2764 -0.01 0.42 -0.09 0.4 ::: Among these genes, we further selected candidate endogenous reference genes according to the following criteria: in at least 95% of the 384 arrays the ~Log2~ratio had to be within 2 SD from zero and the average intensities of both channels across all samples need to be higher than 2,000. The intensity parameter was added to ensure that the selected endogenous reference genes were expressed at a relatively high level in most tissues to mitigate excessive fluctuations in ~Log2~ratios occurring when a low value is applied as denominator. This is important when a reference gene is applied as a denominator in the equation used to normalize the ratio of other tests genes; in such cases robust gene expression decreases the range of ratios resulting from the analysis decreasing, therefore, the experimental variance. Sixty-nine genes fit these criteria. The range of mean ~Log2~ratio was from 0 to 0.15 and the SD of mean ~Log2~ratio was from 0.27 to 0.42. Analysis of 7 replicate array based on the same group of candidate endogenous reference genes demonstrated a similar distribution of mean ~Log2~ratio and SD of mean ~Log2~ratios (with exception of TFAP2B) suggesting that this variance could be attributed to predominantly experimental noise. Among these genes we further selected for validation 11 that had high mean fluorescence intensity (underlined in Table [1](#T1){ref-type="table"}). However, probe preparation and other technical considerations limited to analysis to only 8 of these genes which included: NEED8, HIRIP5, GPLD1, RAB31, SNRPD3, FZD6, COL8A1 and AFAP. \"Leave-one-out cross-validations\" were used to estimate the error \[[@B33],[@B34]\]. In each validation experiment one array was left out. The remaining 383 arrays were used to identify endogenous reference genes using the same parameters used for the original analysis (test/reference ratios \< 2 SD in at least 95% of the experiments and average intensity \> 2,000 in both channels). The median ~Log2~ratio of the endogenous reference gene in the left out array was then compared with the \"real\" normalization factor consisting of the median ~Log2~ratio for all genes in that array supposed to approximate a balanced expression of test and reference genes for that array. This procedure was repeated for each array 384 times. Only in 3 arrays out of 384 (2.3%) were found errors in endogenous reference gene selection (errors were considered median ~Log2~ratio of endogenous reference genes in the left out array \> 2 standard deviations apart from the \"real\" normalization factor in the same array (which is zero for normalized arrays). The expression of each candidate endogenous reference gene identified in this study was then compared with available information about the expression of the same genes in 12 normal human tissues reported by the Affymetrix HG-U95A-E probe sets <http://nciarray.nci.nih.gov/cards/index.html>. We found a good correlation between the expression patterns observed by us and that reported by the Affymetrix GeenChip array (data not shown). In addition, a cluster of endogenous reference genes suggested by Applied Biosystems for standardization during qRT-PCR was analyzed by evaluating the SD of their Log2 ratios across the 384 arrays. As data indicated in Table [2](#T2){ref-type="table"}, although some genes (bold) demonstrated mean SD ~Log2~ratios comparable to that of the endogenous reference genes identified in this study, the corresponding percentage of the arrays in which ~Log2~ratios were within 2 SD from zero ranged between 0.68--0.85 which is significantly lower than the frequency in which the same parameter fell for the 69 candidate house keeping genes identified here (\>95%). In addition, we looked at the variation in our arrays of genes classically used as endogenous reference genes including: ribosomal protein L32, HRPT, β-actin and tubulin-α3. In all cases the mean SD of ~Log2~Ratio was relatively larger than for the endogenous reference genes identified in Table [1](#T1){ref-type="table"} with the following respective values: 0.88 for Ribosomal protein L32, 1.13 for HRPT, 1.32 for β-actin and 1.42 for laminin-α3. Interestingly, however, the mean intensities were relatively higher for all of these genes compared with the ones identified in this study (7,369; 12,778, 43,794 and 42,241 for the four genes respectively) suggesting that these genes are expressed probably at higher concentrations in tissue and, therefore, although relatively variable in expression in different tissues, they have been a useful marker of RNA abundance when only rough estimates are required like, for instance, for Northern Blotting. ::: {#T2 .table-wrap} Table 2 **Gene Name** **Image ID** **Description** **Mean Int** **Mean logR** **SD logR** **% of arrays** --------------- -------------- --------------------------------- -------------- --------------- ------------- ----------------- NKTR 712460 NK-tumor recognition protein=cy 7319.55 0.62 0.79 0.52 NKTR 712460 NK-tumor recognition protein=cy 3048.28 **0.19** **0.43** **0.84** NKTR 712460 NK-tumor recognition protein=cy 8637.15 0.99 0.99 0.53 NKTR 2064497 natural killer-tumor recognitio 959.97 0.72 0.85 0.7 PPIC 882459 peptidylprolyl isomerase C (cyc 5034.82 2.11 1.45 0.44 PPIB 756600 peptidylprolyl isomerase B (cyc 6606.09 1.08 1.04 0.39 PPIL2 450661 peptidylprolyl isomerase (cyclo 3035.17 **0.16** **0.4** **0.85** PPIL2 2017652 peptidylprolyl isomerase (cyclo 850.93 **0.23** **0.48** **0.76** PPIG 809621 peptidyl-prolyl isomerase G (cy 3991.93 0.28 0.53 0.67 PPID 884500 peptidylprolyl isomerase D (cyc 5805.51 0.33 0.57 0.66 PPID 71154 peptidylprolyl isomerase D (cyc 5652.12 0.27 0.52 0.63 PPIH 767277 peptidyl prolyl isomerase H (cy 3277.26 0.28 0.53 0.73 PPIF 758343 peptidylprolyl isomerase F (cyc 15900.53 3.18 1.78 0.13 CYP2J2 454580 cytochrome P450, family 2, subf 2302.76 0.71 0.84 0.65 PPIA 2580290 peptidylprolyl isomerase A (cyc 25478.36 1.32 1.15 0.36 PPIA 1570861 peptidylprolyl isomerase A (cyc 5028.34 0.76 0.87 0.46 GAPD 50117 glyceraldehyde-3-phosphate dehy 38011.52 1.66 1.29 0.4 GAPD 530934 glyceraldehyde-3-phosphate dehy 10526.82 1.14 1.07 0.37 GAPD 530868 glyceraldehyde-3-phosphate dehy 5052.59 0.81 0.9 0.25 GAPD 755641 glyceraldehyde-3-phosphate dehy 1693.23 **0.17** **0.42** **0.84** PGK1 949939 Phosphoglycerate kinase 1 4452.92 **0.16** 0.4 0.85 B2M 878798 beta-2-microglobulin 34211.05 2.08 1.44 0.43 AMBP 2063390 alpha-1-microglobulin/bikunin p 1813.96 **0.15** 0.39 0.86 GUSB 2273001 glucuronidase, beta 6205.14 0.81 0.9 0.41 GUSB 276449 glucuronidase, beta 2241.73 **0.22** **0.47** **0.68** HPRT1 280507 hypoxanthine phosphoribosyltran 12778.31 1.29 1.14 0.35 TBP 280735 TFIID=TATA box binding protein 4065.4 **0.24** 0.49 0.73 TFR2 461750 transferrin receptor 2 2700.22 **0.19** **0.44** **0.84** TFR2 2408681 transferrin receptor 2 2958.88 **0.21** **0.46** **0.82** ::: Test gene expression estimates according to house-keeping gene selection: the transcriptome array ------------------------------------------------------------------------------------------------- To validate the usefulness of candidate endogenous reference genes in large sample populations we developed a new tool that we are planning to use in the future for validation of gene expression across extensive data sets. This tool displays cDNA libraries originated from different tissues or cell lines individually spotted on a solid surface. The principle of this technology is similar to RNA dot blot which uses RNA isolated from samples and transferred to membrane making, therefore, the transcriptome array a high-density dot blot. The labeled DNA probe of interest is hybridized to the immobilized complementary strain of mRNA. A reference gene hybridization will carried out simultaneously to estimate the relative expression of the gene of interest compared with the reference gene. The new tool we describe here, termed transcriptome array, utilizes cDNA generated from source mRNA for target immobilization to improve the stability of the immobilized targets and differentially fluorescence-labeled test and reference probes (RNA or double strained DNA) then can co-hybridize on to the same spots. Using a validated RNA amplification technology \[[@B30]\], large quantity of pure amplified RNA with proportional representation of source mRNA species could be generated from which cDNA could be obtained through a reverse transcription reaction. Because of the minimum amount of cDNA used for fabricating each transcriptome array (\<5 nano gram cDNA/spot) and the size of spots (100 um), the expression of a large number of genes can be theoretically analyzed on thousands of different samples simultaneously. Since the amount of cDNA spotted may vary according to the quality of the starting material and the efficiency of RNA preparation for each sample, absolute estimation of fluorescence from the hybridized probe is not informative of the expression of the given gene in each sample. Therefore, a reference system is necessary so that comparative expression of the test gene can be presented proportional to that of a consistently expressed gene across all samples. Therefore, interpretation of data derived from the transcriptome array relies on endogenous reference gene normalization. To test the usefulness of various endogenous reference gene we resorted to the well characterized expression of the melanocytic lineage-specific gene gp100/Pmel17 \[[@B31]\] that is expressed exclusively though heterogeneously in cells of melanocytic lineage \[[@B35],[@B36]\]. The assumption of this experiment was that in spite of its heterogeneity of expression in samples of melanocytic origin, overall the expression of gp100 should be higher in meloanocytic compared with non melanocytic samples with a high degree of significance. Rho-C has been associated with the metastatization process in melanoma but its specificity for melanocytic lineage is unknown \[[@B37]\]. The differentiation control element DICE is found in the 3\'-UTR of numerous eukaryotic mRNAs and there is no solid association between its pattern of expression and specific physio-pathological or developmental conditions \[[@B38]\]. Therefore, we compared the expression profile of these three genes in 829 cDNA libraries that included 106 melanoma cell lines, 127 melanoma metastases, 2 benign nevi (total of 235 melanocytic samples) and 593 miscellaneous samples containing a predominance of primary esophageal, renal and colon cancers paired with normal tissues from the same organ and a large number of circulating mononuclear cells. In addition, samples from most other normal and cancerous tissues were included although in smaller number (complete list of samples available upon request). As endogenous reference genes we chose β-Actin and new candidate genes identified by this study (see previous section). The test genes (gp100, Rho-C and DICE) were separately hybridized to the transcriptome array. Each gene labeled with Cy3 was co-hybridized with individual endogenous reference genes labeled with Cy5. We then compared ~Log2~Cy5 / Cy3 ratios between melanocytic and non-melanocytic tissues. gp100 was, as expected, expressed more in melanocytic lesions with high degree of significance no matter what gp100 / endogenous reference gene combination was used. The range of significance, however, varied extensively depending upon the gp100 / endogenous reference gene combination. This difference was considered representative of the ability of different endogenous reference genes to normalize for tissue-specific gene expression patterns (Figure [1](#F1){ref-type="fig"}). In details, all endogenous reference genes could segregate melanocytic from non melanocytic lesions with a high degree of significance (Figure [2](#F2){ref-type="fig"}; unpaired two sample *t*test p~2~-values ranged from 6 × 10^-8^to 3 × 10^-36^). There was, however, a big range in the discriminatory capacity among endogenous reference genes with NEDD8, RAB3 and FZD6 providing the highest predictive value (*t*test p~2~-value = 3 × 10^-36^, 1 × 10^-32^and 1 × 10^-21^respectively) and β-actin being one of the least reliable (*t*test p~2~-value = 1 × 10^-8^). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Comparison of gp100 expression normalized by different endogenous reference genes. Candidate endogenous reference genes were selected among those in Table 1. The gp100 probe (Cy3) was co-hybridized with one endogenous reference gene probe (Cy5) at the time to the transcriptome array. The transcriptome array included 235 cDNA libraries derived from samples of melanocytic lineage (maroon bar) and 594 cDNA libraries from samples of non melanocytic lineage (green bar). Melanocytic samples consisting in the overwhelming majority of melanoma metastases or melanoma cell lines while non-melanocytic libraries included a large collection of esophageal, kidney, colon and other cancers and several normal tissues or circulating mononuclear cells. The complete list of samples is available upon request. The expression of gp100 normalized with different endogenous reference genes was compared by unpaired two-tailed student *t*test and the endogenous reference genes were ranked according to the level of significance in their ability to discriminate between melanocytic and non-melanocytic lesions (data presented as the ~Log10~p~2~-value (shown in the boxes associated with individual graphs; for details see Figure 2). The distribution of the ~Log2~ratios for each individual cDNA library is shown for each gp100 / endogenous reference gene combination. In addition, results compiled using the average of the ~Log2~ratios for all the endogenous reference genes, the 5 and the 3 with the lowest p~2~-value are presented (orange bar graphs). :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Expression of gp100 by melanocytic and non melanocytic lesions normalized with different endogenous reference genes. Average gp100 / endogenous reference gene ~Log2~ratios for melanocytic and non-melanocytic lesions are shown together with the standard deviation (SD) standard error from the mean (SEM) and the Log~10~of the *t*-test p~2~-values when melanocytic and non-melanocytic lesions were compared. Also data derived from mathematically averaging the results obtained with all the endogenous reference genes, the ones yielding the best 5 individual p~2~-values and the best three are shown. The same data are presented visually in the bar graph below; in black are data derived with individual endogenous reference gene normalizations, in orange data derived by averaging results from different endogenous reference genes. Filled bares portrait data from melanocytic lesions, empty bars from non-melanocytic lesions. :::  ::: The large variation in the results obtained using different endogenous reference genes to normalize gp100 expression could have been due to a higher stability of the expression of some genes across all samples or to a differential expression of the endogenous reference genes themselves in melanocytic lesion. In the latter case, the better results obtained could be coincidental and not useful in other experimental situations. We, therefore, tested whether pooling results obtained with all endogenous reference genes independently of each predictive value in this controlled experimental situation could yield results as informative about gp100 pattern of expression as those obtainable with individual endogenous reference genes, particularly those that provided the best prediction. The advantage of this strategy is that it does not depend on previous knowledge of gene expression patterns for the selection of individual endogenous reference genes not applicable in conditions, in which the suitability of a gene for a given experimental situation is, contrary to gp100, is unknown. Thus, we compared the predictive value of the average of gp100 / endogenous reference gene ~Log2~ratios obtained with all endogenous reference genes and those obtained using the genes that provided the best five or the best three results (Figure [2](#F2){ref-type="fig"}). The use of pooled information from various endogenous reference genes appeared to stabilize the non-melanocytic ~Log2~ratios gp100 / endogenous reference. In fact, the SD of the~Log2~ratios gp100 / endogenous reference among non-melanocytic samples derived by pooling endogenous reference gene results were significantly lower than the SD obtainable with any individual endogenous reference gene (F-test). The basis for this test was that non-melanocytic lesions uniformly should not express gp100 and, therefore, the ~Log2~ratios for any gp100 / endogenous reference gene combination should be constant resulting in very low SD. SD for non-melanocytic lesions were much lower using pooled endogenous reference genes for the normalization (Figure [2](#F2){ref-type="fig"}) and this was true at extreme levels of significance (F-test value for SD when all endogenous reference genes were used for normalization compared with individual endogenous reference were 0, 1.5 × 10^-9^, 0, 5.6 × 10^-7^, 9.0 × 10^-5^, 0, 0, 0 and 0 for NEDD8, RAB3, FZD6, AFAP, COL8A, SNRPD, HIRIP5, β-actin and GPLD1 respectively). Similar results were obtained when results from the best 5 and best 3 endogenous reference genes were pooled together. In addition, even if pooling data together did not provide the highest level of discrimination compared to the best results obtained with selected individual endogenous reference genes, the capacity of pooled data sets to discriminate gp100 expression between melanocytic and non melanocytic samples was still very high (2 × 10^-22^; 4 × 10^-29^and 1 × 10^-34^respectively when all, the best 5 or the best 3 endogenous reference gene results were pooled together). Interestingly, different endogenous reference genes provided not only different levels of discrimination but also provided different estimates of gp100 expression in melanocytic lesions with gp100 / endogenous reference gene ~Log2~ratio above (NEED8, RAB3, FZD6, AFAP, SNRPD3, HIRIP5, GPLD1) or below (COL8A1, β-actin) 0 (Bottom graph, Figure [2](#F2){ref-type="fig"}). This, of course, has no impact on the ability to characterize patterns of expression of different genes in different tissues but rather suggests that the latter group of endogenous reference genes is expressed at higher copy number, therefore, biasing the data toward its own fluorescence channel. Simple adjustment in probe concentration could easily solve this problem. We then analyzed the expression of Rho-C and DICE whose pattern of expression in the transcriptome array cannot be predicted based on available information save for the notion that of the expression of Rho-C is related to the metastatic process of melanoma \[[@B37]\]. The pattern of Rho-C was characterized with all the endogenous reference genes and representative information is presented in Figure [3](#F3){ref-type="fig"}. Most endogenous reference genes yielded a pattern suggestive of a preferential expression of Rho-C in melanoma metastases. In addition, averages of Rho-C / endogenous reference gene ~Log2~ratios demonstrated clearly a specific expression of this gene in melanoma metastases. This pattern was not necessarily specific for melanoma as the other cancerous tissues spotted in the transcriptome array were obtained from primary lesions. Thus, it is possible that the preferential expression in melanoma metastases was due to the metastatic process rather than the melanocytic lineage. Interestingly, the melanoma cell line Mel-A375 did not constitutively express Rho-C as previously observed by Clark et al. \[[@B37]\]. In addition, none of the melanoma cell lines mostly derived from melanoma metastases demonstrated expression of Rho-C. This information suggests that Rho-C may be involved in the natural metastatic process *in vivo*but is not constitutively expressed *in vitro*. Finally, DICE did not demonstrate any specificity of expression and appeared variably expressed in all specimens independently of the endogenous reference gene used for normalization. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Relative expression of gp100, Rho-C and DICE in melanocytic and non-melanocytic samples. Test samples (Cy3) and endogenous reference genes (Cy5) were hybridized to the transcriptome array as described in Figure 1. Test gene / endogenous reference gene ~Log2~ratios are displayed as a bar graph for NEED8, β-actin and for the averages of test / endogenous reference gene~Log2~ratios when the three endogenous reference genes with the best discriminating power in separating gp100 expression between melanocytic (maroon bar) and non-melanocytic (green bar) lesions (NEDD8, RAB3 and FZD6) were used. Samples derived from melanoma metastases are shown by the orange bar. In the boxes the ~Log10~p~2~-value of significance of differences between relevant samples (se text) are shown; ns = non significant). :::  ::: Validation of data by protein analysis and quantitative real-time PCR (qRT-PCR) ------------------------------------------------------------------------------- We tested whether gp100 / endogenous reference gene ~Log2~ratios correlated with gp100 protein expression level as measured by intra-cellular FACS analysis (Figure [4](#F4){ref-type="fig"}). The gp100 / endogenous reference gene ~Log2~ratios were derived from the transcriptome arrays while protein expression was based on previous characterizations of the same cell lines \[[@B35],[@B39]\]. Overall, a good correlation (Pearson\'s correlation) was noted for gp100 values normalized with most endogenous reference genes. However, with the exception of NEDD8, the best correlation was obtained when pooled information was used from the best 5 and 3 candidate endogenous reference genes as defined before. Interestingly, RAB3 that scored very high as a predictor of gp100 transcriptional expression in melanocytic lesions and β-actin provided the worst correlation with protein expression. This data suggest that, although different endogenous reference genes may yield better predictive value than others in large data sets, it is likely that their reliability varies in different conditions and possibly the best results can be obtained by pooling information from several of them. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Correlation of gp100 expression estimated obtained by intracellular FACS analysis and by gene profiling with the transcriptome array. Cell lines with different level of gp100 expression by intra-cellular FACS analysis using the gp100-specific mAb HMB45 \[35\] were used for the analysis. The compiled information from the transcriptome array in which cDNAs from each cell line was spotted are shown. The Cy5/Cy3 (endogenous reference gene / gp100) ~Log2~ratios are shown for each endogenous reference gene evaluated. In addition, ratios derived by using the data from all the endogenous reference genes, the best 5 and best 3 according to figure 2 are also presented. Pearson\'s correlation for the data set is shown for each endogenous reference gene and the same information is summarized graphically at the bottom of the figure. MEF = mean equivalent of fluorescence \[35\]. :::  ::: We also tested by quantitative real-time PCR (qRT-PCR) the discriminatory capacity of endogenous reference genes in samples with known patterns of expression of gp100. We selected 6 previously characterized \[[@B35],[@B36]\] melanoma cell lines four of which were known to express (MEL-553B, MEL-1317, MEL-526, MEL-1102) and two known not to express (MEL-836, MEL-1195) gp100 at the transcriptional and protein level (Figure [5](#F5){ref-type="fig"}) \[[@B36],[@B39]\]. In addition, 4 fine needle aspirate (FNA) samples from melanoma metastases expressing gp100 \[[@B40]\] and a series of tumors (2 renal cell, 2 esophageal one gastric carcinomas and HELA cells line) known not to express gp100 were tested. Absolute expression of gp100 was measured and although the lesions supposed to be gp100 expressing appeared more positive than those supposed to be gp100 negative there was no absolute demarcation among the two groups (Figure [5a](#F5){ref-type="fig"}). A better differentiation between positive and negative lesions could be obtained when the various endogenous reference genes were used for the analysis or the combined values from all of them (Figure [5b](#F5){ref-type="fig"}). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Absolute and normalized expression of gp100 based on qRT-PCR. Four melanoma cell lines expressing gp100 (MEL-553B, MEL 1317, MEL 526 and MEL-1102) and 2 not expressing gp100 (MEL-836 and MEL-1195) protein by FACS analysis \[35\] (maroon horizontal bar) were tested for gp100 expression by qRT-PCR as previously described \[40\]. In addition four fine need aspirates of melanoma metastases all expressing GP100 by immunohistochemistry (orange horizontal bars) and two renal cell, two esophageal and one gastric cancer specimens all not expressing gp100 (green horizontal bar) were tested. Similarly the gp100 not expressing HeLa cell line was tested. Absolute gp100 copy number is shown in (a); normalized expression of gp100 is shown using the average of all endogenous reference gene results (b). The protein expression for the melanoma cell lines is presented as mean equivalent of fluorescence (MEF) \[35\] as well as immunohistochemical evaluation while the expression by FNA is shown in (c) based on immunohistochemical (HIC) evaluation as previously described \[46\]. Absolute copy numbers of gp100 estimated by qRT-PCR are also shown. :::  ::: Discussion ========== Although the concept of endogenous reference genes is appealing it may be unrealistic to expect that a gene could be equally expressed in all eukaryotic cells in all physiopathological conditions. Thus, the endogenous reference gene concept can only approximate such ideal biological behavior. In fact, numerous studies noted that the expression of the ostensible endogenous reference genes varies according to distinct physio-pathological situations and can be affected by experimental manipulation \[[@B2],[@B20],[@B24],[@B26],[@B27],[@B41]\]. For instance, the levels of expression of GAPDH, cyclophilin and β-actin fluctuate in different tissues, disease stage and are affected by cell behavior like proliferation \[[@B6],[@B15],[@B18],[@B20],[@B42]\]. In addition, endogenous reference genes expression varies markedly in cancer compared to normal tissue \[[@B17],[@B26],[@B41]\] suggesting that oncogenesis may influence their expression \[[@B25]\]. In this study, therefore, we looked for relative stable endogenous reference gene challenging a 17,000 cDNA clone data base against approximately 400 specimens that included a broad variety of normal and neoplastic tissues as well as cell lines. Indeed, few previously unnamed genes approximated the ideal endogenous reference gene by being expressed consistently in more than 95% of the specimens analyzed within a variance that was barely and insignificantly different from what could be expected from intrinsic experimental fluctuation (Table [1](#T1){ref-type="table"}). In fact, the threshold that we selected based on the variance observed by periodically repeating identical experiments was consistent with criteria suggested by others \[[@B43]\] who claimed that genes with fold changes greater than 1.5 can be classified as differentially expressed at \~95% confidence interval. This is consistent to our selection criterion that contained endogenous reference genes within a 1.46 fold maximal variation. Although these genes have not been proposed before as candidate endogenous reference genes, analysis of available data bases suggested that they are relatively stably expressed in various normal tissues (i.e. Affymetrix HG-U95A-E probe sets). Selection of these candidate endogenous reference genes provided useful information when the expression of a lineage specific gene such as gp100 was compared between relevant and irrelevant samples. In addition, gp100 expression normalized according to some of the new endogenous reference genes more closely aligned to lineage specificity than normalization performed using β-actin. This data, however, confirm that there is no such thing as a perfect endogenous reference gene. Although some endogenous reference genes demonstrated high power of discrimination their ranking was dictated by previous knowledge of the expected experimental results (gp100 being melanocytic lineage specific). However, gene expression analysis is done to determine such specificity and in most cases no previous knowledge of tissue specificity is truly available. One may argue that, in spite of the extreme variability in normalization power demonstrated by various endogenous reference genes, still significant differences could extracted between the melanocytic and non-melanocytic samples no matter what endogenous reference genes were used. This statement is correct in this experimental situation where hundreds of samples could be compared with the transcriptome array. However, in cases when only few samples are analyzed the difference in discriminatory capacity becomes critical. For instance, when randomly selected gp100 / endogenous reference gene ~Log2~ratios from 10 melanoma metastases were compared with those from 10 kidney specimens only NEDD8, RAb3, FZD6, AFAP, COL8, SNRPD and pooled endogenous reference gene-based normalizations yielded significant differences between the two groups while non significant values were obtained for normalizations based on HIRIP5, β-actin and GPLD1. The lowest limit of detection in the transcriptome array of specific probes in term of copy numbers is currently under investigation. Decreasing concentrations of β-actin alone printed the array spots as internal controls suggest that the detection range is from 30 ng / spot to 3 pg / spot when standard fluorescent intensity range parameters are applied. The higher limit of detection before saturation can be easily adjusted to avoid under estimates of highly expressed genes, therefore, increasing the dynamic range of detection. This is achieved by adjusting the fluorescent intensity right below the saturation for both channel and subsequently calculating the abundance of reference gene expression using serially diluted internal control spots as standard curve. With few exceptions, the expression of most genes can be estimated with relative accuracy using this semi-quantitative approach. Since in most cases endogenous reference genes are selected based on previous knowledge of their usefulness for a particular experimental situations, it is likely that the best consistency could be achieved in such circumstances when averages of several endogenous reference gene normalizations are used. This may represent a reasonable solution particularly if a cocktail of endogenous reference genes identically labeled could be used as reference within a single experiment. We are presently testing various combinations of lineage specific genes and pools of endogenous reference genes for hybridization to the transcriptome array and for qRT-PCR. Such attempts, however, have been quite disappointing so far as results obtained using mixture of endogenous reference genes do not yield information comparable to that achievable when results obtained with the same endogenous reference genes used individually are pooled together. This is possibly due to imbalanced density of various endogenous reference genes in different samples with predominant effects of some over others. Therefore, extensive work in the future will be aimed at equilibrating the interactions among endogenous reference genes in various experimental conditions to test whether mixtures of them could be used in multiplex for normalization of large expression analysis studies. Conclusions =========== Our observations, together with previous work by others indicate that much caution should be taken when using endogenous reference to normalize the expression of test genes. It is likely that previous analyses based on single endogenous reference genes might have been strongly biased by the individual selections and the interpretation of the results should looked at with caution. In particular, this study suggests that even when the most stably expressed genes in array experiments are used as endogenous reference genes, significant variation in test gene expression estimates may occur and the best normalization is achieved when data from several endogenous reference genes are pooled together to minimize minimal but significant variation among samples. We are presently optimizing strategies for the preparation of endogenous reference gene mixtures that could yield information comparable to that of data pooled from individual endogenous reference gene normalizations. Methods ======= The list of samples used for cDNA microarray hybridization or for preparation of the transcriptome array is available upon request. Most melanoma and renal carcinoma cell lines were generated at the Surgery Branch, National Cancer Institute (NCI), National Institutes of Health (NIH), Bethesda, MD and maintained in RPMI (Biosource International, Camarillo, CA) supplemented with 10% fetal calf serum (Biosource). Cell lines were harvested by Trypsin /Versene (Biosource) digestion. PBMC were collected by leukapheresis from six unrelated normal donors in the Department of Transfusion Medicine, Clinical Center, NIH and purified by Ficoll gradient separation. Surgically removed tumor tissues were collected from the Tissue Network (Philadelphia, PA), Surgery Branch and CPSB specimen bank, NCI, (Bethesda, MD). Fine needle aspiration (FNA) biopsies were obtained from patients with metastatic melanoma referred to the Surgery Branch, NCI for immunotherapy. Total RNA was isolated using Qiagen RNeasy kit and its quality and quantity was estimated using Agilent Bioanalyzer 2000 (Agilent Technologies, Palo Alto, CA). Target preparation for cDNA microarray hybridization ---------------------------------------------------- Total RNA extracted from test and reference samples was transcribed *in vitro*into anti-sense RNA (aRNA) and reverse-transcribed into fluorescence labeled cDNA for hybridization to 17,000 gene cDNA-based microarrays \[[@B30],[@B44]\]. Pooled PBMC were used to prepare reference aRNA to be co-hybridized in all experiments with test aRNA. cDNA targets were labeled with Cy3 (green) for reference material and Cy5 (red) for test material with the exception of reciprocal experiments. cDNA Microarrays ---------------- cDNA (UniGene cluster) microarrays were printed at the Immunogenetics Section, DTM, CC, NIH with a configuration of 32 × 24 × 23 containing 17,500 elements. Clones used for printing were selected from the Research Genetics RG\_HsKG\_031901 8 k clone set and the 9k RG\_Hs\_seq\_ver\_070700 40 k clone set. The 17,500 spots included 12,072 uniquely named genes, 875 duplicate genes and about 4,000 expression sequence tags. For a complete list of genes included in the Hs-CCDTM-17.5k-1px printing please visit our web site at <http://nciarray.nci.nih.gov/gal_files/index.shtml>. Transcriptome array ------------------- A collection of aRNA-based libraries was prepared from 960 frozen tissue samples or cell lines and individual aRNAs were reverse transcribed into cDNA in the presence of 1 μl of dN6 primer (8 μg/μl), 6 μl of first strand buffer, 3 μl of 10 mM dNTP, 3 ul of 50 mM dTT, 1.5 μl of Rnasin and 10 μl of aRNA (6--12 ug) with addition of 3 μl of Superscript II (BRL) in 5.5 μl volume after heating to 65°C for 5 minutes. Reactions were carried out in 96-well plates at 42°C for 90 minutes followed with addition of 1 μl Rnase H and heating for 20 min at 37°C. cDNA were further purified by utilizing CentriCept gelfitration plates to remove non incorporated primer and dNTP. Purified cDNA were transferred to 384-well plates and dried by speedvac. Samples were re-suspended in 13 μl of 3 × SSC followed by shaking at 3,000 rpm for 15 min. Reconstituted cDNA libraries from individual samples and spiked reference gene at different concentration were printed on to poly-L-lysine coated glass slides at the concentration of 0.5--1 μg/μl using the OmniGrid (GeneMachine. San Carlos, CA) printer and Telechem printing pins (TeleChem International, Inc. Sunnyvale, CA). Each 100 μm diameter spot was duplicated at a 250 μm distance. After complete exsiccation, slides were post-processed as described at <http://nciarray.nci.nih.gov/reference/Post_Processing.html>. Probe design Transcriptome array hybridization experiments ---------------------------------------------------------- Specific PCR primers for the amplification of each gene probe were designed using Primer 3 program <http://molbio.info.nih.gov/molbio/analysis.htm>. In order to generate single strand fluorescence labeled cDNA for hybridization, modified specific 5\' primers with an extension of the T7 promoter region were used for PCR amplification. Double strand PCR fragments (300--800 bp) were used as template for *in vitro*transcription to generate single stranded mRNA. 3 μg of amplified RNA were then reverse transcribed into cDNA in present of 4 μl of first strand buffer, 1 μl dN6 primer (8μg/μl; Boehringer Mannheim), 2 μl 10 X low T-dNTP (5 mM A, C and GTP, 2 mM dTTP), 2 μl Cy-dUTP (1 mM, Cy3 or Cy5), 2 μl 0.1 M DTT, 1 μl RNasin, 3 μg amplified mRNA in 8 μl DEPC H~2~O. After heating to 65°C for 5 minutes and cooling to 42°C, 2 μl of SSII was added and the labeling reaction was carried out at 42°C for 90 min. Probe purification and hybridization were performed as previously described \[[@B30],[@B44]\]. Quantitative real-time polymerase chain reaction (RT-PCR) --------------------------------------------------------- Primer and probe sets from each candidate genes were designed using the Primer Express 2.0 program (Applied Biosystems, Foster City, CA) and synthesized by BioSource. Taqman probes were labeled at the 5\' end with the reporter dye molecule 6-carboxy-fluorescein (FAM; emission λ~max~= 518 nm) and at the 3\' end with the quencher dye molecule 6-carboxytetramethyl-rhodamine (TAMRA; emission λ~max~= 582 nm). Total RNA was used for gene expression validation. Measurement of gene expression was performed using the ABI 7900HT sequence detection system (Perkin Elmer, Foster City, CA) as previously described \[[@B40],[@B45]\]. Messenger RNA from test samples were reverse transcribed into cDNA in the presence of oligo dT (12--18 mer). The qRT-PCR procedure was performed by alternating 2 minute cycles at 50°C and 10 minute at 95°C. Forty cycles involving denaturation at 95°C for 15 seconds and annealing/extension at 60°C for 1 minute in 50 μl volume with 1 × TaqMan Master Mix (PerkinElmer). Standard curves were generated for each gene with high and accurate PCR amplification efficiency as determined by the slope of the standard curves. Linear regression analysis of all standard curves documented in all cases an R value ≥ 0.99. The sequence of primer sets and probes used for each gene are shown in Table [3](#T3){ref-type="table"} with statistics for each standard curve. Standard curve extrapolation of copy number was performed for the gene of interest as well as an endogenous reference gene for each sample. To correct for concentration of starting material, normalization of samples was performed by dividing the copies of the gene of interest by copies of the reference gene as previously described \[[@B40]\]. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Quantitative RT-PCR primers and probes ::: ---------- --------------------------------------------------- -------------------------------------- AFAP (+)-TGTCAAGTTAAACCACTAATGTGTTGGT ***Y*= -3.105x + 43.726 R2 = 0.99** (-)-GGCATCCAAATTCTCCAAGAAA FAM-TGCTGCCTCTCCTGAGTAGGGTGGGT-TAMRA CDK5R1 (+)-TCCTACATGGGCAACGAGATC ***Y*= -3.315x + 45.983 R2 = 0.99** (-)-CCAAAAGGCCTCCTTGCA FAM -TACCCGCTCAAGCCCTTCCTGGTG -TAMRA COL8A1 (+)-CCGAGCTAACCGCACCTTT ***Y*= -3.143 x +43.294 R2 = 0.99** (-)-GTCTGCGGGTTGTAGTTCTGTCT FAM -AGTGAAGTTTAACAAACTGCTGTATAACGG -TAMRA FZD (+)-AGCCTCAAAGGTTCCACATCTC ***Y*= -3.515 x + 47.092 R2 = 0.99** (-)-AGGTCACTTCCAGTGTAACACAAATT FAM -TGAGAAAAGAGCAGGGAGGTGGTTGTCA -TAMRA GPLD1-1 (+)-CAGATTGAAGATTTCACTGCATTTC ***Y*= -3.653 x + 49.086 R2 = 0.99** (-)-CATCAAAATGCTCACCATGGA FAM -TCTGCCCACCTCTCTCATGCTGAATCAC -TAMRA GPLD1-2 (+)-CTTTTGCCTGTAGTAGTAAATTGCTTTTA ***Y*= -3.168x + 46.599 R2 = 0.99** (-)-AACTGGCCATATAAC CAAAGGTGTT FAM -TGAATGGTGTTTATTAAACCCTTATGGTCGATATTTCC-TAMRA HIRIP5 (+)-GCTGCCCTAGTTCAATCATTACTCT ***Y*= -3.160x + 43.313 R2 = 0.99** (-)-CGCCTTCTACCTCCGGAATAT FAM -- AAAAATGGAATTCAGAACATGCTGCA-TAMRA HIST1HIA (+)-AGGCGTCCTCCGTGGAA ***Y*= -3.143x + 43.294 R2 = 0.99** (-)-ATGCACCCGTTGCCTTAGTT FAM -- AGCCCGGCGCCTCAAAGGTG-TAMRA NEDD8 (+)-TGACCGGAAAGGAGATTGAGA ***Y*= -3.458x + 48.686 R2 = 0.99** (-)-CCACACGCTCCTTGATTCG FAM -- TGACATTGAACCTACAGACAAGGTGGA-TAMRA PTH (+)-GCATAACCTGGGAAAACATCTGA ***Y*= -3.063x + 45.215 R2 = 0.99** (-)-TGCACATCCTGCAGCTTCTT FAM -TCGATGGAGAGAGTAGAATGGCTGC-TAMRA RAB31 (+)-CCCCTGAAGGATGCTAAGGAA ***Y*= -3.303x + 48.517 R2 = 0.99** (-)-AGCATTTTTTGCACTTGTCTCAAC FAM -- ACGCTGAATCCATAGGTGCCATC-TAMRA SNRPD3 (+)-GGCACAGCTGGAGCAGGTAT ***Y*= -3.177x + 47.200 R2 = 0.99** (-)-CTTCAGCATGTCAGGCAAAATC FAM -- ATCCGTGGCAGCAAAATCCGC -TAMRA β-actin (+)-GGCACCCAGCACAATGAAG ***Y*= -3.127x + 44.733 R2 = 0.99** (-)-GCCGATCCACACGGAGTACT FAM-TCAAGATCATTGCTCCTCCTGAGCGC-TAMRA gp100 (+)GGTTCCTTTTCCGTCACCCT ***Y*= -3.282x + 45.987 R2 = 0.99** (-)CTCACCGGACGGCACAG FAM-ACATTGTCCAGGGTATTGAAAGTGCCGAGAT-TAMRA ---------- --------------------------------------------------- -------------------------------------- ::: Authors\' contributions ======================= PJ validated the system combining high-throughput data with gene-directed analysis by quantitative real-time PCR. YZ performed statistical analysis. YN, MCP, DN, VM and KS participated in the preparation and utilization of array experiments used for the identification of stably expressed genes. NH, HS and PRT provided a large library of samples necessary for the development of the transcriptome array. FMM supervise the project as principal investigator. EW provided leadership for the development of cDNA arrays, transcriptome array and for the overall conduct of the experiments.

PubMed Central

2024-06-05T03:55:47.426650

2004-8-13

PMC516028

Background ========== Gene expression profiling, along with other methods to evaluate the global changes in genomes, provides the opportunity to understand whole scale changes present in human biology. Yet the sheer mass of data presented by these techniques often makes subsequent analysis difficult. Techniques such as gene expression profiling may result in the identification of hundreds if not thousands of differentially expressed genes that may be associated with the biological process, but may also represent noise related to the biological and technical variation. In an economic environment where limited resources are available for the follow-up and validation of potential target genes methods must be provided for the prioritization and sorting of data. Previous methods have relied heavily on the mapping of metabolic pathways or transcription factor binding sites \[[@B1]-[@B5]\]. These processes rely on the premise that the metabolic pathways associated with a given disease are well delineated, or that groups of proteins with very similar structural or functional design are involved in the disease process. In situations where these assumptions may not be true, alternative methods for the sorting of the data are needed. Here we demonstrate an alternative approach using comparative genomics and animal models of human prostate cancer to sort and identify genes involved in the response of prostate cancer cells to the proposed chemopreventive agent Selenium \[[@B6],[@B7]\]. This process takes advantage of the continued sequencing of multiple animal genomes and the ability to produce gene expression profiles in multiple species. Through the use of these techniques one can leverage established animal models to identify genes associated with human disease processes, as is demonstrated here with the identification of Insulin-like growth factor-2 Binding protein 3 (IGFBP3) and retinoid-X-receptor alpha (RXRalpha). Results ======= Generation of common genes and homologs --------------------------------------- Sequence validated gene libraries for both the rat and human DNAs were obtained from Research Genetics (Huntsville, AL), and were supplemented with additional DNA samples obtained from the University of Iowa rat clone sequencing program \[[@B8]\]. The majority of the rat DNAs, and a subset of the human DNAs were resequenced by Dr. J. Quackenbush at TIGR through a joint Program in Genomic Applications consortium. The GeneBank accession numbers for the 19,200 individual human or rat clones present in the recent slide printings were used to query the NCBI Unigene database to return the associated Unigene IDs. Unigene IDs were returned for virtually all identified clones, and were placed in an Oracle database where they were compared with the downloaded NCBI Homologene dataset (build 106) of rat, mouse, and human homologues. Of the 19,200 clones, 5740 genes were identified with homologues present on both the rat and human slides. This homologue set was used for the subsequent comparisons across species. Similar global and prostate gene expression profiles between rat and human prostate cancer cell lines ----------------------------------------------------------------------------------------------------- We have sought to compare the rat and human prostate cancer transcriptomes in an effort to judge the degree of similarity between the two cell types. Because the use of differentially expressed genes would bias the comparison by eliminating the majority of genes that do not show any difference, we used the absolute level of expression for each gene and compared the rat and human genes for significant differences in absolute expression levels. In order to derive the absolute level of expression for individual genes in human or rat prostate cancer cells we used expression values derived from the associated self-self hybridizations performed for each cell line. The experiments were facilitated by the use of slides that have been quality controlled for the quantity of spotted target DNA through the use of a FITC label third dye \[[@B9]\]. These slides were subsequently imaged for FITC fluorescence and sorted based on the similar amounts of target DNA present on each slide \[[@B10]\]. Using the third dye quality control correlation coefficients of greater than 0.80 are routinely achieved between slide replicates \[[@B9]\]. In this manner comparisons of bound hybridized probe can be made across slides with a degree of confidence. RNA samples from cells were harvested, labeled, and homotypically hybridized to establish the baseline level of consistency within the hybridizations. Performing slice analysis on the normalized homotypic gene expression data across all the self-self hybridization slides within a species and retaining genes that demonstrated consistent expression patterns within two standard deviations of the mean expression value was performed to remove a degree of error from the technical replicates. Using the third dye as a baseline for comparison, these common expressed genes were then broken down into their component Cy3 or Cy5 expression vectors and used to build the transcriptomes for each gene using their absolute expression values. These transcriptomes were then used to compare expression values between the rat and human cell lines. These genes were annotated and gene homologues identified from the NCBI Homologene\[[@B11]\] dataset of rat-human homologues. Thus from a dataset of 5740 homologues, 2883 genes were found that were present within this experimental dataset and expressed in both the rat and human prostate cancer cell lines, and thus could be used for comparative genomics. These samples were processed using the Multiexperiment Viewer mircoarray statistical analysis and visualization program developed by TIGR \[[@B12]\]. Files were loaded and visualized for comparison across the 2883 common expressed genes in a self-organizing tree algorhythm \[[@B13]\] (figure [1](#F1){ref-type="fig"}) and analyzed for similarities in global expression patterns. The hierarchical clustering in self-organizing trees failed to demonstrate a pattern of clustering between species. T-test analysis \[[@B12],[@B14]\] between the human and rat cell lines identified 58 genes (2%) which demonstrated significantly different expression patterns between species (p \< 0.01 with Bonferroni correction). Thus in these comparisons, 2826 genes, or 98% of the genes examined, failed to demonstrate a statistically significant difference in expression between the human and rat prostate cancer cell lines. Using principle components analysis (figure [2](#F2){ref-type="fig"}, \[[@B12]\]) these studies can be visualized, and demonstrate while there is some clustering of the rat and human prostate cancer cell lines, the differences are not significant. Thus when comparing gene expression patterns in rat and human cell lines one will detect significant species-specific differences in expression in 1 out of every 50 genes, with the majority of the genes demonstrating similar expression patterns. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Gene expression profiles for human and rat prostate cancer cells. Clustering of the expressed genes in the human (LNCAP, DU145, PRO4, LN4, and PC3 derivatives) and rat (AT3, MatLyLu, and PAIII) prostate cancer cell lines based on the common homologs as defined within to NCBI Homologene database. Raw data files are available for review from the corresponding author. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Principal Components Analysis of Rat and Human Prostate Cancer Cell Lines. There is a clustering of the human (Pro4-purple, LN4-dark-blue, PC3S-light blue, PC3US-yellow) and rat (MatLyLu-red, AT3-magenta, PAIII-green) prostate cancer cell lines in the same quadrant. The degree of separation within the quadrant was not significant by T-testing. Each sample is presented in duplicate based on independent Cy3 and Cy5 vector profiles. :::  ::: The presence of a large quantity of genes whose expression may be related to general cellular functions, as opposed to prostate specific metabolism, could infuse a significant amount of homogeneity to the data. In the presence of such homogeneity it may be impossible to identify the true differences that are related to prostate cellular function, and thus the perceived similarities may be artifactual. To address this issue we sought to repeat the analysis using only prostate related genes. To generate a list of such genes we used cDNAs in eight normal human prostate cDNA libraries present in the NCI Cancer Genome Anatomy Project \[[@B15]\]. Generation of a list of common genes proved impossible, as the combination of more than four of the cDNA lists resulted in the number of common genes being reduced to zero. A similar result was obtained when one attempted to generate a list of commonly expressed genes across multiple different cancer cDNA libraries. As an alternative approach we developed a list of 12,008 expressed genes were identified based on their presence in at least one of the eight normal human prostate cDNA libraries. The human Unigene IDs for each of the expressed genes were then used to identify the associated rat homologues from Homologene \[[@B11]\] and yielded 2,269 homologous rat genes (18.9%), of which 1,319 (58.1%) had associated prostate cancer gene expression data. These 1,319 prostate expressed genes were then used to repeat the comparative genomics. Similar visual and clustering results were identified for the prostate transcriptomes. T-test analysis \[[@B12],[@B14]\] between the human and rat cell lines identified 30 prostate expressed genes (2%) which demonstrated significant differential expression between species (p \< 0.01 with Bonferroni correction, while 1,289 genes (98%) failed to demonstrate a significant difference in expression across species. Thus even when only prostate expressed genes are considered, similar results were obtained. Between the rat and human prostate cancer cell lines the patterns of expression are similar for 49 of 50 genes examined. Comparison of global and prostate specific differential gene expression profiles between rat and human prostate cancer cell lines treated with selenium ------------------------------------------------------------------------------------------------------------------------------------------------------- While global gene expression profiles appear to be similar between rat and human prostate cancer cell lines one wonders whether the response to specific physiologic stimuli may elicit similar transcriptional changes. If so, one may be able to infer a degree of homology in their biological response to the stimuli. This has already been observed on a physiological level for the rat models of prostate cancer. For example, rat and human prostate cancers respond very similarly to chemotheraputic and environmental agents including hormonal agents (both respond), cyclophosphamide (neither respond), high fat diets (increased incidence), and soy isoflavones (decreased incidence) \[[@B16]-[@B22]\]. In an effort to evaluate these similar biological responses we have compared the transcriptomes between rat and human prostate cancer cell lines treated with the proposed prostate cancer chemopreventive agent Selenium. Samples from the human PC3 and rat PA-III cell lines were treated with Selenium and examined for differential gene expression profiling. These two cell lines were chosen based on their similar biologic characteristics, as both cell lines were derived from androgen independent metastatic tumors, and thus represent tumors with similar biologic potential \[[@B23],[@B24]\]. The cells were treated with twenty-five micromolar Selenium for either 6 hours or 5 days, to identify both immediate changes in gene transcription or changes related to the long term exposure to Selenium. Due to our interest in prostate cancer we have attempted to choose a form and concentration of Selenium that would be reflected in the ongoing prevention trials such as the SELECT prostate cancer prevention trial \[[@B25],[@B26]\]. In this trial patients receive Selenium in the form of Selenized baker\'s yeast. Previous HPLC and electrospray mass spectroscopy studies have demonstrated that 85% of the Selenium in yeast is present as selenomethionine \[[@B27]\]. Selenomethionine has previously been used in *in-vitro*studies of prostate cancer cells\[[@B28],[@B29]\]. These studies demonstrated an inhibition of prostate cancer cell proliferation over a broad range of concentrations, while an IC50 and/or decreased expression was seen at concentrations above 70 micromolar selenomethionine. To avoid the general effects of cell inhibition or cell death while focusing on the effect of Selenium we chose a lower concentration of 25 micromolar selenomethionine. These changes, while not resulting in increased cell death, did cause decreased cell division and increased doubling time in both species (data not shown). Common rat and human homologous genes demonstrating differential expression by greater than two standard deviations were identified and included 1123 genes after 6 hours and 1053 genes after 5 days of exposure to Selenium. When the expression patterns of these genes were compared across species by T-test and principle component analysis as outlined above 713 genes (25%) were found to have statistically significant differences in expression between species (p \< 0.01 with Bonferroni correction). Thus when comparing rat and human samples, while the majority of the gene expression changes are similar, in at least one in four genes (p = 0.75) one can detect significant species specific differences in expression alteration when cells are treated with Selenium. Yet similar physiologic changes (decreased cellular proliferation, increased cell death) were observed in both species. These changes represent the desired physiologic changes one would expect for the chemopreventive effects of Selenium, and could be dissected by examining the common transcriptional changes seen in both species with respect to Selenium. Combined differential expression patterns for selenium responsive genes identify common gene pathways ----------------------------------------------------------------------------------------------------- Because some of the differences in the rat and human prostate cancer cell line transcriptomes may be related to confounding variables such as culture methods, cell passage number, or time in culture, an effort was made to focus on genes that are common, and as such may define the similar Selenium based cell proliferative changes. The subsets of 1123 and 1053 differentially expressed genes (6 hours and 5 days respectively) were analyzed for genes that demonstrate similar changes in expression with respect to Selenium across species. Of these differentially expressed genes, 291 and 309 demonstrated up-regulation in rat and human cells at 6 hours and 5 days respectively. Likewise, 261 (6 hours) and 216 (5 days) demonstrated down-regulation in the presence of Selenium. When these subsets were further analyzed to identify genes with similar levels of up or down-regulation (defined as ratio differences within 0.2 units of each other) 81 genes were identified at 6 hours and 73 at 5 days (table 1-see [additional file 1](#S1){ref-type="supplementary-material"}). These genes included 40 ESTs or genes with limited associated data, and 90 defined genes with associated gene data. Twenty-four of the genes were common to Selenium treatment at both 6 hours and 5 days. Additional information related to these genes was obtained using the GeneInfo data mining tool. This tool was developed by the authors (MWD, XW, HL, GZ) to allow for the rapid identification of supplemental data from the biomedical literature related to genes of interest. In brief, the tool allows one to cut and paste a list of genes based on either Unigene or Genebank IDs and search PubMed for associated references based on annotations of the associated gene names. Additional search terms can be stipulated by the user based on their knowledge of the biological process or in response to results received from the previous search. Results are returned in a table that lists the number of references that met the search criteria and provides a hyperlink to the associated references for either downloading or viewing. In this way the user is allowed to direct queries in an open manner based on their own experience or unpublished data. In this manner searches were conducted using the list of genes and the search terms \"prostate cancer\", \"Selenium\", and \"apoptosis\" (table 1-see [additional file 1](#S1){ref-type="supplementary-material"}). IGFBP3 and RXR-alpha are expressed in the prostate, induced by selenium, and downregulated in prostate cancer ------------------------------------------------------------------------------------------------------------- Of the 154 genes identified with similar cross-species differential expression changes with respect to Selenium, two genes were identified that had unique features based on their associated references and interrelated functions. These genes, IGFBP3 and RXR-alpha were both up-regulated with respect to Selenium and could be used to suggest a model for Selenium action in prostate cancer. PXR-alpha is upregulated in both rat and human prostate cancer cells at 5 days in response to Selenium. Likewise, IGFBP3 is upregulated after six hours of Selenium treatment in both species. These two genes both contained Medline references with respect to prostate cancer, but had not yet been implicated in Selenium action. Western blotting performed on the human prostate cancer cell line PC3 with respect to Selenium validated the bioinformatically identified expression data (figure [3](#F3){ref-type="fig"}). To confirm the role of these two proteins in the prostate immunohistochemical studies on prostate cancer tissue microarrays were performed to identify IGFBP3 and RXR-alpha in both normal, nodular hyperplasia (benign prostatic hypertrophy), high grade prostatic intraepithelial neoplasia (HGPIN), invasive carcinoma, and metastatic prostatic carcinoma (table [2](#T1){ref-type="table"}). These studies demonstrate that both IGFBP3 and RXR-alpha are expressed in the normal human prostatic epithelium (figure [4](#F4){ref-type="fig"}, table [2](#T1){ref-type="table"}). IGFBP3 is also expressed in the prostatic basal cells. Patterns of expression were predominantly nuclear, a finding that has been described for both proteins \[[@B30]\]. In addition, staining for IGFBP3 was also noted in the prostatic stroma, consistent with IGFBP3\'s associated function as a secreted protein. Decreased levels of IGFBP3 was noted in prostatic cancers when compared to normal prostate epithelium (p = 0.0044). Along with this decreased expression there was a distinct shift in the protein localization nuclear to cytoplasmic was observed (p \< 0.00001), and in cases where expression was still present, there were decreased numbers and intensity of cell staining. IGFBP3 expression was similar in HGPIN, invasive carcinoma, and metastatic carcinoma. The level and pattern of IGFBP3 expression in nodular hyperplasia was similar to that seen in normal prostate tissues, and significantly different from the expression seen in cancer samples (p = 0.0036 and p \< 0.00001 respectively). RXR-alpha expression was also significantly downregulated in prostate cancer when compared to normal prostate epithelium or nodular hyperplasia (p \< 0.0001), and was similar to that seen in HGPIN and metastatic carcinoma. RXR-alpha expression was consistently nuclear in the samples studied, and while the intensity of staining was similar, in the remaining positive cancer cases there were decreased numbers of cells staining (8.6 +/- 12.6% in malignant epithelium vs 20.0 +/- 25.5% in normal epithelium). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Expression of IGFBP3 and RXR-alpha with respect to Selenium. Western blotting reveals an induction of RXR-alpha or IGFBP-3 protein after Selenium treatment of human PC3 prostate cancer cells (arrows, upper row). Western blotting of immunoprecipitations from rat PAIII cells (bottom row) reveal RXR-alpha in immunoprecipitated IGFBP3 extracts (right panel) and IGFBP-3 in immunoprecipitated RXR-alpha extracts confirming and extending the reported interactions between the human proteins\[40\]. :::  ::: ::: {#T1 .table-wrap} Table 2 ::: {.caption} ###### Expression of IGFBP3 and RXRalpha in Prostatic Epithelium ::: Normal Prostate Nodular Hyperplasia HGPIN Prostate Cancer Metastatic Cancer ------------------------------- ----------------- ----------------------- -------------------- ----------------------- ------------------- IGFBP3 Positive cases 105 62 49 202 25 Negative cases 5 1 9 36 8 Statistics (comparison) p = 0.0036 (cancer) N.S. (cancer) p = 0.0044 (normal) N.S. (cancer) IGFBP3 Intensity (avg+/-std) 2.47 +/- 0.70 2.49 +/- 0.65 2.57 +/- 0.82 2.74 +/- 0.56 2.79 +/- 0.49 Percentage cells (avg+/- std) 8.3 +/- 13.5 7.5 +/- 12.5 8.8 +/- 15.2 4.4 +/- 6.6 8.5 +/- 12.6 Nuclear cases 92 59 40 94 8 Cytoplasmic cases 22 6 18 152 11 Statistics (comparison) p \< 0.00001 (cancer) p = 0.065 (cancer) p \< 0.00001 (normal) N.S. (cancer) RXRalpha Positive cases 92 58 35 112 16 Negative cases 10 3 31 125 19 Statistics (comparison) p \< 0.00001 (cancer) N.S. (cancer) p \< 0.00001 (normal) N.S. (cancer) RXRalpha Intensity (avg+/-std) 2.73 +/- 0.51 2.78 +/- 0.50 2.83 +/- 0.38 2.76 +/- 0.49 3 +/- 0 Percentage cells (avg+/- std) 20.0 +/- 25.5 23.2 +/- 25.7 8.4 +/- 12.5 8.6 +/- 12.6 4.2 +/- 4.6 Nuclear cases 92 58 35 107 16 Cytoplasmic cases 2 0 6 9 0 Statistics (comparison) N.S. (cancer) N.S. (cancer) N.S. (normal) N.S. (cancer) ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Expression of IGFBP3 and RXRalpha in human prostate tissues. Immunohistochemical staining for IGFBP3 is present as brown staining in normal prostate (A) and prostate cancer (C). Similarly RXRalpha expression is present in normal prostate (B) and lost in prostate cancer (D). All images recorded at 100× magnification. :::  ::: Discussion ========== Leveraging cross-species bioinformatics in the prioritization of gene data -------------------------------------------------------------------------- Through the use of cross-species comparisons of the number of differentially expressed genes to be examined after 6 hours and 5 days of Selenium treatment was dropped from 9453 and 7768 to 1123 and 1053 respectively, an 87--89 percent reduction of the sample size. Even with the use of multiple timepoints, the number of differentially expressed genes was only reduced in a single species study to 5934, less than half. By using comparative genomics the final dataset was reduced to 154 genes, providing a greater than 100 fold enrichment of the data. Thus by leveraging the additional biological species the ability to reduce the final analysis pool was substantial. This process only works if the species used have biological relevance to the disease in question. The choice of rat prostate cancer cell lines was made based on their use as an animal model for the study of prostate cancer \[[@B31]\]. The animal systems have been extensively used in the study of hormonal carcinogenesis, and in particular have been of value as a model of environmental and dietary effects on prostate cancer \[[@B18]-[@B20]\]. Previous studies have identified similar effects of rat animal models and prostate cancer cell lines to soy based diets \[[@B17]-[@B19]\], high fat diets \[[@B20]-[@B22]\], hormonal chemotherapeutics (Pollard, personal communication) and standard chemotherapy \[[@B32],[@B33]\]. While comparative gene expression profiling has been performed, this has usually been through cross-species hybridizations to leverage RNA studies in species where sufficient expressed transcripts in a given species have not been identified for the production of species-specific gene expression slides, in particular for microbial genomes \[[@B34]-[@B37]\]. Thus the approach taken here leverages the production of species-specific gene expression profiles along with the increasing amount of gene homolog data generated by the sequencing of additional animal genomes. It is expected that with future genome efforts additional cross-species studies will be possible that leverage the knowledge of additional animal models in the study of disease. Similarities in prostate cancer transcriptomes across species ------------------------------------------------------------- For both overall and prostate expressed genes, we have failed to identify a significant difference in the transcriptomes between rat and human prostate cancer cell lines. This general similarity in transcriptomes may be due to the inherent biological similarities of the cell lines and/or their underlying biological origin. While the studies sought to utilize prostate cancer cell lines with similar biological potentials (established cell lines all derived from metastases) the degree of diversity present within the samples may account for some of the residual differences still identified. In addition, the extended period of time that these cell lines have been used has allowed for the continued in-vitro evolution of the cells, and could possibly extend those genomic differences. Yet the common clustering of the rat and human cell lines together suggests there are still significant similarities in their biological potential. This is also demonstrated by the similar biological potential of the cell lines when treated with a given stimulus, in this example Selenium. This parallels the similar physiological properties observed in the rat models of human prostate cancer. Based on these features we demonstrate that it is possible to identify functionally significant genes related to Selenium response by using comparative genomics. These findings also support the use of animal models in the study of human prostate cancer by suggesting that there is enough inherent genomic similarity that valuable insights may be gained from animal systems. Comparative genomics identifies functionally significant genes with respect to selenium chemoprevention ------------------------------------------------------------------------------------------------------- A true test of the profiling method is the identification of genes that have a functional significance to the experimental system. In this case we have identified a series of genes, which when examined with additional data mining techniques, identifies genes with associated roles related to apoptosis (IGFBP3, RXRalpha, dynamin-2), antioxidant protection (selenoprotein N, peroxiredoxin I, zinc metalloprotease, glutathione S transferase), cell cycle (CDC26-anaphase promoting complex, kinetochore associated protein), and protein balance (proteasome subunit beta-4, ubiquitin conjugating enzyme). In addition, the ability to sort the identified genes by their associated biomedical literature allowed the focus to shift to IGFBP3 and RXRalpha. Retinoids, through the retinoid X receptor, have been shown to induce the expression of IGFBP3 \[[@B38]\]. In concert these two proteins act to induce apoptosis in cancer cell lines \[[@B39]\]. In particular, recent data has shown that these proteins work in synergy to enhance apoptosis in prostate cancer, and that there is a physical interaction between these two proteins in prostate cancer cells\[[@B40]\]. Further validation and confirmatory data is presented here that demonstrates the selenium induced expression and interaction between both RXRalpha and IGFBP3 in prostate cancer cells, along with their expression in normal prostate epithelium and subsequent down-regulation in malignant prostatic epithelium. This allows one to pose a model by which the restoration of IGFBP3 and RXRalpha levels by Selenium treatment may lead to the disruption of prostate tumorigenesis. This model is testable, and if validated, would present not only a mechanism by which Selenium may exert its effect, but provide a biomarker for assaying the effect of Selenium supplementation in the ongoing prostate cancer prevention clinical trials. Conclusions =========== Using gene profiling on highly controlled spotted cDNA arrays we have demonstrated that similar baseline and selenium induced gene expression profiles can be identified between rat and human prostate cancer cells. This has allowed us to filter our gene expression data to identify genes whose transcriptional response to Selenium is similar across species, and by so doing focus our discovery process on specific common physiologic pathways. Two such proteins, RXR-alpha and IGFBP-3, which may be located in a common pathway, have been identified as dysregulated in human prostate cancers. This provides further support that the cross-species methods employed here can identify genes with roles in human prostate cancer. Methods ======= Cell culture and selenium treatment ----------------------------------- Cell lines were received from ATCC, Rockford, MD, (LNCap, DU-145, MatLyLu, AT3), from Drs. Paul Lindholm and Andre Kadjacsy-Balla (LN4, Pro4, PC3, PC3-NI(PC3US), PC3-I(PC3-S)), or Dr. Morris Pollard and Mark Suckow (PA-III). These cells were cultured in RPMI (DU-145) or DME medium supplemented with 10% fetal calf serum, 10 mM glutamine, and 10 mM sodium pyruvate, and passaged 1:8 or 1:10 when the cells reached 70--80% confluence with trypsin-EDTA. For the Selenium studies PC3 or PAIII cells from a single cell stock were seeded at 1 × 10EE4 cells per ml and grown to 50% confluence at which time the culture medium was changed to either standard growth medium (above) or medium supplemented with twenty-five micromolar Selenium (Seleno-DL-methionine, Sigma cat\# S3875, St. Louis MO). The cells were then cultured for an additional 6 hours or 5 days. Cells that reached 80% confluence prior to the five day timepoint were split using trypsin-EDTA and replated in either control or selenium-containing medium for the duration of the experiment. Cells were monitored for viability and cell growth with parallel growth curves conducted in triplicate, this data demonstrated the previously described \[[@B41],[@B42]\] decrease in cellular proliferation (data not shown) observed in the presence of Selenium. RNA isolation and quantitation ------------------------------ RNA was isolated from cells using Trizol (Invitrogen cat \# 15596018, Carlsbad, CA) and subsequently examined for quality using agarose gel electrophoresis and Gelstar nucleic acid stain against known RNA standards and failed to demonstrate significant degradation based on the presence of high molecular weight RNA species, and intact 28s and 18s ribosomal RNA bands. DNA library preparation and amplification ----------------------------------------- Sequence-verified rat and human libraries (Research Genetics, Huntsville, AL, and University of Iowa cDNA clone set, IA), consisting of 41,472 human clones and 36,000 rat clones were used as a source of probe DNA. A subset of 200 randomly selected clones were chosen from these libraries, resequenced locally, and demonstrated clone accuracy of 92%. We have opted to reformat libraries from 96 to 384-format for culture growth/archiving, PCR, purification, and printing. This has reduced the number of plates of our 41,472 human clone library from 432 to a more manageable 108, and the rat clone library from 375 to 94. The library was reformatted and subsequently manipulated using slot pin replicator tools (VP Scientific, San Diego, CA). Cultures were grown in 150 ul Terrific Broth (Sigma, St. Louis, MO) supplemented with 100 mg/ml ampicillin in 384 deep-well plates (Matrix Technologies, Hudson, NH) sealed with air pore tape sheets (Qiagen, Valencia, CA) and incubated with shaking for 14--16 hr. Clone inserts were amplified in duplicate in 384-well format from 0.5 μl bacterial culture diluted 1:8 in sterile distilled water or from 0.5 μl purified plasmid (controls only) using 0.26 μM of each vector primer {SK865 5\'-fluorescein-GTC CGT ATG TTG TGT GGA A-3\' and SK536: 5\'-fluorescein-GCG AAA GGG GGA TGT GCT G-3\'} (Integrated DNA Technologies, Coralville, IA) in a 20 μl reaction consisting of 10 mM Tris-HCl pH8.3, 3.0 mM MgCl~2~, 50 mM KCl, 0.2 mM each dNTP (Amersham, Piscataway, NJ), 1 M betaine, and 0.50 U *Taq*polymerase (Roche, Indianapolis IN). Reactions were amplified with a touchdown thermal profile consisting of 94°C for 5 min; 20 cycles of 94°C for 1 min, 60°C for 1 min (minus 0.5° per cycle), 72°C for 1 min; and 15 cycles of 94°C for 5 min; 20 cycles 94°C for 1 min, 55°C for 1 min, 72°C for 1 min; terminated with a 7 min hold at 72°. PCR reactions analyzed for single products by 1% agarose gel electrophoresis analysis. Products from replicate plates were pooled and then purified by size exclusion filtration using the Multiscreen 384 PCR filter plates (Millipore, Bedford, MA) to remove unincorporated primer and PCR reaction components. Forty wells of each 384-well probe plate were quantified by the PicoGreen assay (Molecular Probes, Eugene, OR) according to the manufacturers instructions. After quantification, all plates were dried down, and reconstituted at 150 ng/μl in 3% DMSO/1.5 M betaine. Array slide fabrication ----------------------- A single printing array containing 19,200 elements (human) or 2 arrays of 9,600 (rat), were printed on poly-L-lysine coated slides prepared in-house (1--2 arrays/slide) as previously described \[[@B9]\]. Printing was conducted with a GeneMachines Omni Grid printer (San Carlos, CA) with 16 or 32 Telechem International SMP3 pins (Sunnyvale, CA) at 40% humidity and 22°C. To control pin contact force and duration, the instrument was set with the following Z motion parameters, velocity: 7 cm/sec, acceleration: 100 cm/sec^2^, deceleration: 100 cm/sec^2^. All slides were post-processed using the previously described nonaqueous protocol\[[@B9]\]. Slide coating was performed as described previously \[[@B43]\]. Image files on all arrays were collected after blocking (fluorescein), and again after hybridization (Cy3 and Cy5) with a ScanArray 5000 (GSI Lumonics, Billerica, MA). Experimental design and bioinformatics based data analysis ---------------------------------------------------------- The experimental design utilized two biological replicates for each comparison with each replicate incorporating a Cy3/Cy5 dye flip. In addition, self-self hybridizations were performed for each sample to ensure experimental accuracy and evaluate expression bias. Comparisons were organized in a loop design for either human or rat prostate cancer cell lines, or were run as two-sample comparisons of baseline untreated control and Selenium treated cells. Array image TIFF files were analyzed with Gleams software (Nutec Sciences, Atlanta, GA). Additional TIFF file analysis, data normalization, clustering, and principle components analysis was performed using the Spotfinder, MIDAS and MultiExperiment Viewer Software from The Institute for Genomic Research (TIGR, Rockville, MD, \[[@B44]\], \[[@B12]\]) and used default values set in the MCW Practical Guide to TIGR Software Use (M. Datta, unpublished). In brief, image expression data was used as channel intensity minus background and intensity thresholds were set at a value of 300. Images were analyzed as dye flip pairs normalized using MIDAS with LocFit based LOWESS normalization and slice analysis set at two standard deviation cutoffs and a sample data population of 500 \[[@B45]\]. Samples were then averaged across two dye flip replicate pairs with removal of zero/dropped values using locally developed averaging software from the BEAR microarray suite (M. Datta, submitted). These final averaged values were subsequently annotated using the BEAR suite annotator and used for pattern identification and correlation with gene homologs. Homologous genes were identified from the NCBI homologene database ftp files <ftp://ftp.ncbi.nih.gov/pub/HomoloGene/> and parsed using local scripts and databases present in the Bioinformatics Program,\[[@B46]\]. Additional data mining to identify references in the biomedical literature associated with specific genes and user chosen search terms was performed using the locally developed GeneInfo data tool (M. Datta, submitted). Raw data files, along with analyzed data subsets are available for use and study and can be obtained via a secure ftp site after contacting the corresponding author <mdatta@mcw.edu>. Protein purification, western blotting, and immunoprecipitation --------------------------------------------------------------- Protein extracts were prepared and immunoprecipitations and/or western blots made from five day twenty-five micromolar Selenium treated or control PC3 or PAIII prostate cancer cell lines as described previously\[[@B47]\]. In brief, ten micrograms of total protein were run on pre-cast 12% reducing SDS PAGE gels (Bio-Rad Labs, Hurcules, CA) and transferred to PVDF membranes. After blocking with caseine blocking buffer (Bio-Rad Labs, Hurcules, CA) the PVDF membranes were incubated with either anti-RXR-alpha or anti-IGFBP-3 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) at 200 μg/ml dilutions, washed, and incubated with anti-rabbit secondary antibody (2 μg/ml) and developed with ECL Chemiluminescence (cat. RPN2108, Amersham Biosciences, Piscataway, New Jersey). Immunoprecipitations were carried out using 200 microgram samples of total cellular protein, which after preclearing with protein A agarose beads was sequentially incubated with either anti-RXR-alpha (1 μg/ml) or anti-IGFBP-3 (1 μg/ml) antibodies, washed, incubated with anti-rabbit protein A agarose beads, washed, and the protein pellet western blotted with the complimentary antibody (anti-IGFBP-3 or anti-RXR-alpha, respectively), and developed with ECL Chemiluminescence. Tissue microarray production, immunohistochemistry, and analysis ---------------------------------------------------------------- After expedited institutional review board approval normal prostate tissues and prostate cancer samples were obtained from de-identified discarded patient specimens. The formalin-fixed paraffin embedded specimens were prepared as 5 micron sections. Tissue microarrays were prepared from donor tissue blocks as 0.6 mm cores in 12 (4 × 4) or (5 × 5) grids with between 192 to 300 samples and used in the preparation of 5 micron sections. Immunohistochemistry was performed using primary rabbit polyclonal antibodies to the insulin-like growth factor binding protein 3 (IGFBP3, 1:300, Santa Cruz Biotechnology, Santa Cruz, CA), or retinoic-X-receptor alpha (RXR-alpha, 1:800, Santa Cruz Biotechnology, Santa Cruz, CA) using methods previously described \[[@B48],[@B49]\]. In brief, endogenous peroxidase from deparaffinized sections were blocked with Methanol/Acetic acid, and after treatment with blocking serum (ABC kit, Pierce Biotechnology, Rockford, IL) samples were incubated for 30 minutes with either anti-IGFBP3 (1:300) or anti-RXRalpha (1:600). Sections were subsequently washed, and incubated with anti-rabbitt secondary antibody conjugated to horseradish peroxidase and counterstained with Mayers hematoxalin. Antigen retrieval (90 C waterbath for 10 minutes) was used for RXRalpha. Positive controls for each antibody included nuclear staining in Sertoli cells \[[@B50]\] and lymphocytes\[[@B51]\]. Positive staining was recorded and scored on a 0--2 scale (0 = no staining, 1 = staining that does not obscure the hematoxalyn counterstain, 2 = staining that obscures the hematoxalyn counterstain). Evidence of positive staining was recorded as presence of staining (yes/no) or percent of epithelial or basal cells staining (number of cells staining over total number of cells). Patterns of staining (nuclear, cytoplasmic, membranous, diffuse extracellular) were also recorded. All samples were analyzed and recorded by two separate personnel, including a trained urologic pathologist (MWD, BM). Statistical analysis was performed using Chi-squared probability analysis. Abbreviations ============= None declared. Authors contributions ===================== M.W.D. was responsible for the conception and implementation of this project in association with P.J.T., M.S., and M.P. H.L., X.W., and G.Z. were actively involved in the programming, database construction, and testing of the software. M.S. and M.H. were responsible for spotted cDNA construction, hybridization, and experimental analysis along with M.W.D. Cell culture, western blots, immunoprecipitations, and selenium treatments were performed by M.S. with assistance by B.M. Tissue microarray staining and analysis was performed by M.W.D., R.D., T.B., and B.M. All the authors reviewed and accepted the final version of the paper. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 Table 1, Word document, Table of the genes identified in the selenium gene expression studies. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ The authors would like to acknowledge the support of NCI grant R21CA098032 to MWD and the Milwaukee Breast Cancer Showhouse Foundation Award to MWD in support of this work.

PubMed Central

2024-06-05T03:55:47.435265

2004-8-20

PMC516029

Background ========== The thymus is a bilobate organ derived from embryonic endoderm and mesoderm differentiation and is located just above the heart (reviewed in \[[@B1]\]). It is the primary organ for maturation of T cells. This process involves the interaction between developing thymocytes and thymic stromal cells. Thymic stromal cells which forms the thymic architecture, have been classified according to their anatomical localization. They encompass a very diverse array of cell types, including cortical and medullar epithelial cells, fibroblasts, macrophages and dendritic cells (reviewed in \[[@B2]\]). Stromal cells control the differentiation of haematopoietic precursors derived from the liver or the bone marrow into T lymphocytes: T-cell differentiation is defined by the acquisition of maturation markers such as CD4, CD8 and the T-cell receptor complex (TCR), which conditions the reactivity of immature thymocytes with thymic stromal cells. The early thymocyte progenitors entering the thymus do not express T-cell-specific molecules, such as CD3, the alpha or beta-chain of the TCR, or the CD4 and CD8 molecules. These CD4^-^CD8^-^cells, referred to as double-negative (DN) cells, then become CD4^+^CD8^+^, the so-called double-positive (DP) stage, and then progressively acquire TCR molecules. The final maturation of T-cells involves the selective loss of either the CD4 or the CD8 molecules to generate fully mature single-positive (SP) cells with cytotoxic/suppressor or helper/regulator function, respectively. During this process, the TCR-mediated positive and negative selection of T cells ensures the selection of a diverse TCR repertoire able to react with foreign peptide presented by autologous major histocompatibility complex (MHC) molecules, but tolerant to self-antigens. This property renders the thymus an attractive site for manipulation of T-cell tolerance. To date, results on tolerance induction via direct manipulation of the thymus have been scarce (reviewed in \[[@B3]\]). However, previous studies using intrathymic (IT) injection of pancreatic islet cells \[[@B4]\], soluble antigens \[[@B5]\] or adenoviral vectors \[[@B6],[@B7]\] have shown that induction of tolerance to foreign antigens in non-immunosuppressed animals is feasible. Since the production and maturation of thymocytes may be a life-long process, a major drawback to the utilisation of soluble antigens or adenoviruses is their short-term expression in the thymus \[[@B8]\]. Indeed, modulation of the selection process should stop upon the disappearance of the antigen, which might be a problem for long-term tolerance induction. Due to their ability to infect resting cells and to stably integrate into the genome, lentiviral vectors represent powerful new tools for long-term expression of a given transgene *in vivo*\[[@B9]\]. Lentiviral vectors have been used successfully *in vivo*to infect hepatocytes and muscle cells \[[@B10]\], antigen-presenting cells \[[@B11],[@B12]\], as well as cells of the central nervous system \[[@B13]\]. We reasoned that lentiviral vectors might be better suited than adenoviral vectors for long-term IT expression of a foreign gene. We therefore investigated the pattern of infection of a ubiquitous lentiviral vector after IT injection. We report herein that thymic stromal cells are massively and persistently infected. Developing thymocytes exhibit a significant but low level of infection. Moreover, we show that IT injection of a lentiviral vector encoding the cognate antigen in TCR-transgenic (Tg) mice leads to negative selection of developing thymocytes. Results ======= Intrathymic injection of lentiviral vectors results in the efficient and persistent infection of thymic stromal cells --------------------------------------------------------------------------------------------------------------------- We used a concentrated viral stock of the LvPGK-GFP vector (2.10^9^TU~143B~/ml) to inject between 7 × 10^7^to 1.2 × 10^8^infectious units in the thymus of normal C57Bl/6 mice. Infected cells could readily be detected at day 5 post-injection by direct examination of the thymi under a UV microscope (figure [1](#F1){ref-type="fig"}). Of note, transduced cells could still be observed at 1 (figure [1](#F1){ref-type="fig"}) and 3 months (data not shown) post injection. Most of the transduced cells had a fibroblastic morphology. To monitor a possible passage of the vector through the bloodstream, we checked for the presence of transduced cells in the liver, which is the primary target organ after intravenous injection of lentiviral vectors \[[@B14]\]. We indeed observed numerous eGFP^+^cells in the liver (figure [1](#F1){ref-type="fig"}), suggesting a significant leak into the circulation upon IT injection of up to 30 μl of the vector. Nevertheless, our results demonstrate efficient and persistent infection of thymic stromal cells upon IT injection of a lentiviral vector. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### ***In vivo*expression of eGFP after intrathymic injection of the LvPGK-GFP vector.**D5: day 5 post-injection localisation of transduced cells around the injection site (arrow) under visible and UV-lights (upper picture). Fibroblast-shaped cells are predominantly transduced (UV-light only) (lower panel). D30: day 30 post-injection expression of eGFP in the thymus (visible + UV-light) (upper panel) and in the liver (UV-light only) (lower panel). CTRL: Thymus (upper panel) and liver (lower panel) pictures from control mice injected IT with PBS examined for background fluorescence under visible and UV-lights. Magnifications are indicated in the lower right corner of each picture. :::  ::: Induction of negative selection after intrathymic injection of lentiviral vectors --------------------------------------------------------------------------------- We next wanted to investigate whether thymic stromal cells infected with a lentiviral vector would be able to mediate negative selection of developing thymocytes. We thus injected a lentiviral vector encoding the HA of the Influenza virus, or eGFP as a control, into the thymus of SFE-Tg mice expressing a TCR specific for a HA protein peptide and for which a clonotypic antibody recognizing the transgenic TCR (clone 6.5) is available \[[@B15]\]. Six days after injection, we analyzed the thymocytes by flow cytometry after thymi dilacerations. The frequency of 6.5^+^cells within CD4SP and CD8SP thymocytes is shown in figure [2A](#F2){ref-type="fig"} for a representative experiment. Intrathymic injection of the LvPGK-HA vector resulted in a diminution in the frequencies of 6.5^+^cells within CD4SP by a factor of 2 and by a factor of 5 in CD8SP cells with an almost complete disappearance of 6.5^hi^cells in the latter subset. Of note is that the intensity of TCR transgenic expression was reduced in CD4SP cells (figure [2A](#F2){ref-type="fig"}). Overall, we observed a 3.5-fold decrease in the total numbers of thymocytes expressing the specific TCR in mice injected with the LvPGK-HA vector compared to the LvPGK-GFP vector (figure [2B](#F2){ref-type="fig"}). Our results altogether demonstrate that within a week after intra thymic injection of the LvPGK-HA vector, infected thymic stromal cells efficiently mediated negative selection of HA-specific thymocytes. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Negative selection of developing thymocytes(A)**TCR transgenic expression within thymic CD8SP (white) and CD4SP cells (grey) identified by the anti-clonotypic monoclonal antibody 6.5 in SFE-Tg mice six days after IT injection of 40 to 60 ng p24 of the LvPGK-GFP lentiviral vector (n = 2) or of 3.5 to 6 ng p24 of the LvPGK-HA vector (n = 3). Shown are representative profile of two independent experiments. Numbers indicate the frequency of 6.5^+^cells **(B)**Absolute counts of HA-specific thymocytes six days after intra thymic injection of LvPGK-GFP or LvPGK-HA lentiviral vectors. These figures were obtained based on the percentages of total 6.5^+^thymocytes determined by flow cytometry as shown in (A). Statistical analysis was performed using Student\'s t-test. :::  ::: Intrathymic injection of lentiviral vectors results in low level infection of immature thymocytes ------------------------------------------------------------------------------------------------- We next investigated whether developing T-cells would be infected upon IT injection of the lentiviral vector. Five days post-injection of the LvPGK-GFP vector into the thymus of a normal mouse, very few eGFP^+^cells could be detected by flow cytometry within the thymocytes obtained from dilacerated thymi (figure [3](#F3){ref-type="fig"}). Most of these cells were CD3^-^cells, and belonged to the DN and DP subsets, showing that infected cells were mostly immature. Interestingly, at day 30 post-injection, we observed more mature eGFP^+^cells that expressed CD3 and that belonged to the subsets of CD4SP and CD8SP for more than half of them. This result indicates that infection *per se*did not interfere with the normal process of T-cell development. Moreover, we observed a similar repartition of CD4/CD8-expressing cells in non-infected eGFP^-^cells (data not shown). Collectively, these results show that immature thymocytes can be infected by *in situ*lentiviral infection. However, we were unable to clearly detect infected T-cells in the spleen of injected mice, likely due to their small representation within the pool of mature lymphocytes in the absence of a selective advantage for the transduced thymocytes. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **eGFP expression in developing thymocytes.**Total thymocytes were stained with anti-CD4, anti-CD8 and anti-CD3 monoclonal antibodies. Upper panels: saline-injected control mice (CTRL) and LvPGK-GFP-injected mice at two different time points after injection (D5 and D30) are shown. Lower panels: The profile of CD4/CD8 expression is shown within gated eGFP^+^cells. :::  ::: Discussion ========== We report herein that IT injection of a lentiviral vector results in the predominant infection of thymic stromal cells, and to a low level infection of thymocyte progenitors. Significant infection of liver cells was also detected. This observation is reminiscent of what was observed by DeMatteo et al. with adenoviral vectors \[[@B7]\]. Together with the fact that liver cells are main targets of IV-injected lentiviral or adenoviral vectors, this suggest that a significant leak into the circulation does occur upon IT injection of viral vectors. Our *in situ*analysis suggests that thymic epithelial cells represent the vast majority of infected cells, and studies are underway to more precisely define the cells targeted by IT lentiviral injection. Whatever the proportion of cortical, medullar epithelial cells, or thymic dendritic cells that are transduced, we show here that this results in an antigen presentation that efficiently mediates negative selection of specific thymocytes. This is not due to the injection of a \"crude\" preparation of viral supernatant that could have non-specifically affected T cell differentiation. Indeed, we injected 10 times lower amounts of p24 from the LvPGK-HA vector than of the LvPGK-GFP vectors, suggesting that negative selection of HA-specific thymocytes was a direct effect of HA expression by thymic stromal cells. This is further supported by the analysis of the frequencies of 6.5^+^thymocytes which shows deletion of 6.5^hi^cells within CD8SP cells, an MHC class-II restricted population in these TCR-transgenic mice \[[@B15]\]. Down modulation of the transgenic TCR and deletion was observed within CD4SP cells. This is reminiscent of the results obtained recently by Trani et al. which showed that intra thymic delivery of increasing dose of the HA peptide in SFE TCR-Tg mice resulted in the down regulation of the transgenic TCR \[[@B16]\]. Therefore, deletion and/or receptor down regulation may act in concert in the negative selection of HA-specific CD4SP cells in SFE transgenic mice. A very low infection of developing thymocytes was detected. This is not surprising as (i) the multiplicity of infection (ratio of number of infectious units over number of total cells in the thymus) was estimated to be lower than 0.4 and (ii) lentiviral transduction of murine T cells is far less efficient than of human T cells \[[@B17]\]. Since the actual volume that can be injected in a mouse thymus is however limited, we used the highest MOI achievable with our concentrated vectors. It would be interesting to assess if the use of vectors with higher infectious titres or of repeated injections would lead to a better transduction of thymocytes. It should be stressed though that at day 5 after injection, the infected cells represented immature thymocytes not expressing CD3 molecules. At later time points, infection was detected in more mature thymocytes. We believe that this result may have important implications for *in vivo*gene therapy of severe combined immunodeficiencies (SCID) affecting T-cell development (reviewed in \[[@B18]\]). Indeed, most of these diseases are due to monogenic mutations and concerns immature thymocytes, such as in the T^-^B^+^NK^-^deficiencies linked to the common cytokine receptor gamma-c \[[@B19]\], or to the ZAP-70 protein tyrosine kinase \[[@B20]\]. Our results open the possibility of correcting these developmental blocks through IT delivery of a lentiviral vector expressing a functional molecule. For this particular application, it would be important to avoid transgene expression in the thymic stroma. The use of our recently described T-cell specific lentiviral vector represent an attractive possibility towards this end \[[@B21]\]. Given the tremendous proliferative potential of T cells and the selective advantage that will be provided by the transgene, even a low number of transduced cells should result in a significant T cell reconstitution. This is best exemplified by a unique case of X-linked severe combined immunodeficiency in which a reverse mutation occurred in a single early T cell precursor. It was determined that at least 1,000 T cell clones with unique T cell receptor-beta sequences were generated from this precursor and that this diversity seems to be stable over time and provides protection from infections in vivo \[[@B22]\]. Furthermore, our preliminary results show that intra thymic delivery of the ZAP-70 gene by mean of a T-cell specific lentiviral vector in ZAP-70-deficient mice results in the restoration of T-cell development (submitted). The presently described *in vivo*approach may represent an alternative to gene therapy protocols using cumbersome haematopoietic stem cell manipulation *ex vivo*prior to their reinfusion *in vivo*. Conclusions =========== Results presented herein may have important implications for the experimental and therapeutic manipulation of the immune system, and notably for tolerance induction and the correction of SCID. Methods ======= Mice and intrathymic surgery ---------------------------- C57Bl/6 mice were obtained from Charles River/IFFA Credo Laboratories (Les Oncins, France) at 6 weeks of age and were used at 8 to 10 weeks-old. SFE TCR-Tg mice \[[@B15]\] were bred in our own animal facility and were used at 6 to 10 weeks-old. Intrathymic surgery was performed after anesthetic treatment of animals with 40 mg/kg of Pentobarbital (Sanofi-Synthelabo, Gentilly, France). Mid-incision of the lower neck was performed to gain access to the trachea. Incision of the sternum was performed on the first two ribs and gently pulled aside to view the thymus. A single injection of 10 to 30 μl was performed using 0.3 ml Terumo insulin syringes (VWR, Fontenay-sous-bois, France). Lentiviral vector construction, production, concentration and quantification ---------------------------------------------------------------------------- The plasmid encoding the lentiviral vector pRRLsin.PPT.hPGK.GFPpre (LvPGK-GFP) has been described elsewhere \[[@B23]\] (kindly provided by L. Naldini (University of Torino Medical School, Torino, Italy). To construct the plasmid encoding the hemaglutinin (HA) protein of the Influenza virus, BamHI and SalI restriction sites at the 5\' and 3\' ends, respectively, were added to the cDNA of the HA protein of Influenza virus (H1N1) in the pCIneoHA plasmid (provided by Genethon, Evry, France) by PCR using the Taq polymerase (Invitrogen, Cergy-Pontoise, France). Total PCR products were cloned into the TA vector (Invitrogen), checked for sequence integrity and digested with BamHI/SalI. The plasmid pRRLsin.PPT.hPGK.GFPpre was digested with BamHI/SalI (New England Biolabs, Beverly, Massachussets, USA) to remove eGFP. After ligation, the plasmid pRRLsin.PPT.hPGK.HApre, hereafter referred to as LvPGK-HA, was obtained. To produce lentiviral vectors, a total of 4.10^6^293T-cells were co-transfected with the transfer vector, the packaging and the envelope plasmids in 10-cm dishes using the calcium phosphate method as described \[[@B21]\] in DMEM supplemented with serum and antibiotics (Lifetechnologies, Gaithersburg, Maryland, USA). Lentiviral supernatants were collected at 18, 42 and 66 hrs post co-transfection in serum-free DMEM supplemented with antibiotics and L-glutamine, and concentrated by ultrafiltration using either the Ultrafree-15 or the Centricon Plus-80 filter devices according to the manufacturer instructions (Millipore, Bedford, Massachussets, USA). Briefly, supernatants were applied to the filter devices and spun at 2000 g for 20 min. at 20°C. Concentrated supernatants were aliquoted and kept at -80°C until use. Viral stocks of the LvPGK-GFP lentiviral vector were titered on 143B cells as previously described \[[@B21]\]. Viral stocks of the LvPGK-GFP and LvPGK-HA vectors were also quantified using a *gag*p24 ELISA (Zeptometrix, Buffalo, New York, USA). Microscopy and images treatment ------------------------------- Whole thymus or liver were excised from injected mice and placed in PBS 1X in a 6-well plate. Pictures of the whole organ were acquired using a DP-11 numeric camera coupled with the CK-40 inverted microscope equipped with a mercury lamp (Olympus France S.A, Rungis, France). Images were processed using Adobe Photoshop (Adobe Systems Inc., San Jose, California, USA). Flow cytometry -------------- Thymi were dilacerated between two frosted slides in 1X PBS supplemented with 3% Fetal Calf Serum (Lifetechnologies). Cell suspensions were numerated and 10^6^cells were stained with the following monoclonal antibodies (Becton Dickinson Biosciences, le Pont de Claix, France): CD4-APC (allophycocyanin), CD8-CyCr (Cychrome) and either CD3 or pan beta-chain of the TCR-PE (phycoerytrin) or purified anti-clonotypic TCR for the HA peptide SFERFEIFPK presented by MHC class II I-E^d^(clone 6.5) (\[[@B15]\]) followed by biotinylated anti-rat IgG2b and streptavidin-FITC. Data were collected on a FACScalibur (BD Biosciences) and analysed with FlowJo software (TreeStar, Ashland, Oregon, USA). Abbreviations ============= DN: double negative DP: double positive SP: single positive IT: intra thymic IV: intra venous HA: hemaglutinin MOI: multiplicity of infection SCID: severe combined immunodeficiencies Authors\' contributions ======================= GM devised and realised the experiments, analysed the data and wrote the manuscript. DK conceptualised the study and edited the manuscript. Acknowledgements ================ We would like to acknowledge Dr H. von Boehmer for providing us SFE TCR-Tg mice. The contributions of Enguerran Mouly for production of the lentiviral vector LvPGK-GFP, Béatrice Levacher for construction of the LvPGK-HA lentiviral vector, Mélanie Dodille and Nadège Carrié for production and titration of the LvPGK-HA vector were essential to this work. The authors are also undebtful to Dr B. Salomon and to Dr N. Taylor for suggestions and critical reading of the manuscript. This work was supported by the Centre National de la Recherche Scientifique (CNRS) Agence Nationale de Recherche sur le SIDA (ANRS), SIDAction, Association Française contre les Myopathies (AFM) and the Fondation pour la Recherche Médicale (FRM).

PubMed Central

2024-06-05T03:55:47.439794

2004-8-19

PMC516030

Background ========== Periodicity of DNA in exons, with the period being equal to 3 nucleotides, has been well known for some time \[[@B1]-[@B6]\]. This periodicity reflects correlations between nucleotide positions along coding sequences \[[@B7]\], which is caused by the asymmetry in base composition at the three coding positions \[[@B8]\]. This periodicity has also been suggested as a reading-frame monitoring device during translation, due to interrupted periodic patterns matching with frame shifts downstream where the periodic pattern returns \[[@B9]\]. The triplet code has undergone evolution itself, from the earliest form of the triplet code to what exists today. The universal DNA periodicity observed in exons suggests a (**RNY)~n~**pattern (**R**= **A**or **G**, **Y**= **C**or **U**, **N**= any base), which probably was inherited from the earliest mRNA sequences \[[@B10],[@B11]\]. In this study comparisons between real and simulated coding sequences were used in attempt to better understand the cause of the DNA periodicity. The only data used by the simulation program were codon usage frequencies from real species. Thus the simulated coding sequences had frequencies of codons very similar to real species. The major difference, however, was a random position of codons in simulated sequences. The periodicity of exons, as well as other coding statistics can be an additional tool for exon prediction programs \[[@B7]\]. The distance between two types of nucleotides is counted, and a period is determined by the distance between the similar frequencies. For example, if there is a nucleotide **A**at one point in a sequence, and other **A**\'s are more common when there are 2, 5, 8 and so on nucleotides between them, a period of three can be determined \[[@B7]\]. Additional methods of finding periodicity include Fourier analysis \[[@B12]\], the length shuffle Fourier transform algorithm \[[@B13]\], autocorrelation functions \[[@B14]\] and distance analysis \[[@B15]\]. We applied two models of an evolutionary algorithm \[[@B16],[@B17]\] to quantify DNA periodicity. Thus the second objective of this study was investigation of a new method for quantification of DNA periodicity. Results ======= Periodicity of DNA in exons --------------------------- As mentioned in the Background, periodic 3-nucleotide pattern has been known for eukaryotic exons for some time. We studied a question whether DNA periodicity similar to that observed in exons can be simulated in computer experiments utilising codon usage frequencies (CUF) of real species as the only source of information. The computer program GENERATE, which was used in these experiments, composed artificial coding sequences using CUF of several species as the only source of information. Thus despite random choice the frequencies of codons in simulated sequences were very similar to the real CUF. As an example of these experiments, Figure [1A](#F1){ref-type="fig"} shows distribution of **Adenine**nucleotides in real *Drosophila melanogaster*exons (phase 0) and in simulated (Figure [1B](#F1){ref-type="fig"}) coding sequences (phase 0) created by GENERATE using *D. melanogaster*CUF. To avoid any significant influences of splicing signals, *D. melanogaster*exons aligned at the 5\' end start from the 10^th^nucleotide (4^th^codon). In the simulated coding sequences periodicity is also highly pronounced and the periodicity patterns observed in *D. melanogaster*exons and simulated sequences are nearly identical (Figure [1A](#F1){ref-type="fig"} &[1B](#F1){ref-type="fig"}). Other studied species *C. elegans*and *H. sapiens*despite significant differences in **AT**and **GC**content also show high similarity in the periodicity pattern between exons and simulated sequences (data are not shown). Periodicity of other nucleotides was also observed and it shown high similarity in both DNA of real exons and simulated sequences (data are not shown). The obvious conclusion following from this study is that CUF, which was the only source of information for the simulated coding sequences, is the crucial factor determining periodicity. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **A. Adenine**periodicity in *Drosophila melanogaster*exons (N~exons~= 32,760); **B. Adenine**periodicity in simulated coding sequences (N~genes~= 13,065) created by GENERATE using *D. melanogaster*CUF; **C. Adenine**periodicity in simulated coding sequences (N~genes~= 13,065) created by GENERATE using equal frequencies of all non-stop codons and *D. melanogaster*frequencies of stop codons and **D. Adenine**periodicity in simulated coding sequences (N~genes~= 13,065) created by GENERATE using equal frequencies of all codons including frequencies of stop codons. The length of coding sequences was determined by the program. :::  ::: DNA periodicity in simulated coding sequences was dramatically reduced in the experiments where the frequencies of all non-stop codons were made equal (Figure [1C](#F1){ref-type="fig"}). This observation strongly supports the conclusion that codon usage frequencies determine DNA periodicity in exons. A very light periodicity of **Adenine**and **Thymine**(data are not shown) was caused by the fact that 3 stop codons and the corresponding combinations of nucleotides were present in the simulated coding sequences in different and much lesser frequencies than other codons. **Cytosine**, which is not a component of any stop-codon, does not show periodic pattern at all because frequencies of **Cytosine**containing codons were equal to frequencies of all other non stop codons. Finally, when frequencies of all codons, including stop codons, were made equal, no periodicity was observed in the simulated sequences (Figure [1D](#F1){ref-type="fig"}). Thus the computer simulations lead to a firm conclusion that 3 nucleotide periodicity observed in DNA of exons is determined by codon usage frequencies. The triplet nature of genetic code is rather responsible for the length of the period but not periodicity itself, as some people might think. Quantification of DNA periodicity using an evolutionary algorithm ----------------------------------------------------------------- Data sets on frequency of the nucleotides and all 16 dinucleotides at each location were constructed for *Caenorhabditis elegans*, *Drosophila melanogaster*and *Homo sapiens*phase 0 exons. The frequencies of **Adenine**and dinucleotide pair **AG**are shown in this paper as an example. Two models were used to fit to these data, with the key objective of describing periodicity in the data. Inspection of Figures [2](#F2){ref-type="fig"} to [5](#F5){ref-type="fig"} shows that periodicity is quite apparent, and that the frequencies between peak and trough frequencies are generally consistent, but sometimes trending in value. Two models were used to accommodate this shifting pattern. The first model is: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Curves of best-fit for **Adenine**in phase 0 exons compared to actual frequencies. Pink points represent frequencies of nucleotide A in phase 0 *C. elegans*, *D. melanogaster*and *H. sapiens*exons aligned at the 5\' end. Blue points represent the best-fit curve for the data points in an ideal situation, from position 20--100. The scales of the graphs were altered to provide better contrast between the data points. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Curves of best-fit for dinucleotide **AG**in phase 0 exons compared to actual frequencies. Pink points represent frequencies of AG in phase 0 *C. elegansD. melanogaster*and *H. sapiens*exons aligned at the 5\' end. Blue points represent the best-fit curve for the data points in an ideal situation, from position 20--100. The scales of the graphs were altered to provide better contrast between the data points. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Fixed period best-fit curves for **Adenine**in phase 0 exons compared to actual frequencies. Pink points represent frequencies of nucleotide A in phase 0 exons aligned at the 5\' end. Blue points represent the best-fit curve for the data points in an ideal situation, from position 20--100. The scale of the graph was altered to provide better contrast between the data points. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Fixed period best-fit curves for dinucleotide **AG**in phase 0 exons compared to actual frequencies. Pink points represent frequencies of dinucleotide **AG**in phase 0 exons aligned at the 5\' end. Blue points represent the best-fit curve for the data points in an ideal situation, from position 20--100. The scale of the graph was altered to provide better contrast between the data points. :::  :::  \- where  is predicted frequency at nucleotide position *i*, and *b*~1~to *b*~5~are parameters to be estimated from data. Because of irregularities in frequencies close to the 5\' end of exons, nucleotide position 20 was take as position *i*= 1. The component *b*~1~+ *b*~2~*i*fits an overall linear trend in frequencies, independently from finer-scale periodicity. For some data sets it could be useful to include a quadratic term for *i*. Parameter *b*~3~gives the amplitude of periodic waves. As *2π*radians describes a full cycle, periodicity is given by parameter *b*~5~. However, if *b*~5~differs from exactly 3, this generates a shift in phase that is linear with nucleotide position. This shift combines with static phase shift parameter *b*~4~to fit the relative frequencies in adjacent groups of three locations. This is not an ideal model, in that parameter *b*~5~does not cleanly describe periodicity, but it proves to work well in practice. The second model is:  \- where the value of Offset depends on nucleotide position I as follows: if i \< b4 Then Offset = b6 else if i \< b4 + b5 Then Offset = b7 else Offset = b8 Thus, in this case, three regions are defined by parameters *b*~4~and *b*~5~, and Offset is defined within region by parameters *b*~6~and *b*~7~. Model 2 fixes periodicity at 3 nucleotides, but allows for different patterns of relative frequency in regions chosen by the data. The analysis task for both models is to find values of the *b*parameters that give a close fit between the real frequencies *Y*and the predicted frequencies . The criterion used for this was the sum of squared errors across nucleotide position:  Best-fitting *b*parameters were found using a form of evolutionary algorithm (Differential evolution, \[[@B17]\]) with modifications to improve robustness following \[[@B18]\]. To test if the above method could identify a quantifiable period in exons, several tests were performed. All phase 0 exons were extracted from the EID database, described in Material and Methods. This procedure dramatically enhanced the visible periodicity compared to exons of all three phases (Figure [1A](#F1){ref-type="fig"}.). Once the phase 0 exons were separated, they were aligned at the 5\' end and data for all four single nucleotide frequencies were analysed using both methods. In addition to all single nucleotide frequencies, the dinucleotide frequency of **AG**was run through the analysis for both dynamic period determination best-fit curve, and the static period of 3 best-fit curve. This was done for the three studied species, *C. elegans*, *D. melanogaster*and *H. sapiens*for phase 0 exons. Several other dinucleotide pairs also shown clear periodicity patterns in exons (data are not shown). **AG**was used in this paper as an example. Introns were only run through the analysis for dynamic period determination and did not show clear and stable periodicity. Phase 0 exons -- dynamic period determination model (Model 1) ------------------------------------------------------------- Curves of best-fit were created for the four separate nucleotides and the dinucleotide **AG**for *C. elegans*, *D. melanogaster*and *H. sapiens*phase 0 exons. These curves were created using model 1, in an attempt to find a periodicity within the given data. The first few positions of exons are frequently under different selection pressures, which do not always conform to the same pressures as the remainder of the exon. It was for this reason that the algorithm was run starting from position 20 to position 100 in exons. Table [1](#T1){ref-type="table"}, shows DNA periodicity in exons of *C. elegans*, *D. melanogaster*and *H. sapiens*as they were determined by the analysis. The amplitude of the periodicity was measured as the variation from the center-point of the sine curve. As mentioned earlier, the criterion value of goodness of fit is the sum of squared deviations between observed frequencies and frequencies predicted by the model with the prevailing parameters. This means that as the analysis runs through its generation cycles, it finds better fitting curves as it goes along and replaces the previous curve of best-fit. The criterion value is a reflection of this process, in that as it gets closer to zero, the closer the curve of best-fit represents actual data. The data show that in all cases, the determined period is very close to 3 in exons for all three species for all nucleotides and the dinucleotide pair **AG**. The range of periods goes from a low of 2.990391 in *C. elegans*nucleotide **C**, a difference from 3 is \~0.0096, to a maximum of 3.019349 in *C. elegans*dinucleotide pair **AG**, a difference from three of \~0.0193. The criterion values of goodness of fit are also very low for exons, with the largest among them at 0.0093573 in *H. sapiens*nucleotide **C**. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Periods of best-fit curves in phase 0 aligned exons. Periods and criterion values of phase 0 *C. elegansD. melanogaster*and *H. sapiens*exons aligned at the 5\' end. The four single nucleotide as well as a single dinucleotide pair, **AG**were studied. ::: **Species** **Nucleotides** **Best-Fit Period** **Amplitude** **Criterion Value** ------------------- ----------------- --------------------- --------------- --------------------- *C. elegans* **A** 2.995367 ± 0.02007 0.0019226 **C** 2.990391 ± 0.01583 0.0019602 **G** 3.005597 ± 0.08199 0.0047119 **T** 3.014848 ± 0.06247 0.0049716 **AG** 3.019349 ± 0.02077 0.0031441 *D. melanogaster* **A** 3.002192 ± 0.10128 0.0055213 **C** 3.000776 ± 0.07056 0.0043703 **G** 2.998296 ± 0.09142 0.0036970 **T** 2.999109 ± 0.06115 0.0049037 **AG** 2.999744 ± 0.05038 0.0012143 *H. sapiens* **A** 3.000390 ± 0.07898 0.0087358 **C** 2.997380 ± 0.05627 0.0093573 **G** 2.998682 ± 0.08669 0.0014469 **T** 2.998730 ± 0.06627 0.0065035 **AG** 2.995860 ± 0.03872 0.0025106 ::: A comparison between the best-fit curves for nucleotide **A**and actual frequencies of nucleotide **A**in exons of the three studied species can be seen in Figure [2](#F2){ref-type="fig"}. Figure [3](#F3){ref-type="fig"} shows a similar comparison for the dinucleotide pair **AG**. Best-fit curves for nucleotides **C**, **G**and **T**are not shown. The blue points on the graph represent data points predicted by fitting model 1 to actual data, represented by pink points. As can be seen in the graphs, the blue best-fit curves in both Figure [2](#F2){ref-type="fig"} and Figure [3](#F3){ref-type="fig"} both closely follow the pink line for actual data, which confirms that the best-fit line quite accurately portrays actual data. Phase 0 exons -- static period 3 model (Model 2) ------------------------------------------------ The data were then fitted to model 2, keeping the period fixed at 3. With this algorithm, the ideal best-fit curve would remain nearly in the same pattern. Again phase 0 exons were used as sample data. Figure [4](#F4){ref-type="fig"} shows the best-fit curve for nucleotide **Adenine**in *C. elegansD. melanogaster*and *H. sapiens*phase 0 exons with the period fixed at three. Figure [5](#F5){ref-type="fig"} shows the best-fit curve for the dinucleotide pair **AG**in 0 exons in the same species. In both sets of graphs, pink points represent actual frequencies of nucleotides and the dinucleotide pair, **AG**, while the blue points represent the optimized curve of best-fit for these frequencies. It is clear from the graph that keeping the period fixed at exactly three in exons does not detract from the accuracy of the curve of best-fit. The curve of best-fit is remarkably similar to the actual data points. Discussion ========== The fact of DNA periodicity in exons, as well as the lack of periodicity in introns, is known for some time \[[@B1]-[@B6]\]. \"Such a periodic pattern reflects correlations between nucleotide positions along coding sequences (that is, the probability of finding a nucleotide at a given position in a coding sequences is not independent of the nucleotide occurring at some other even distant position). The correlations arise, in turn, because of the asymmetry in base composition at the three codon positions in coding sequences\" \[[@B7],[@B8]\]. The simulation experiments described in the paper support such conclusion and provide a clear proof that frequency of codon usage is the key cause for DNA periodicity in exons (Figure [1](#F1){ref-type="fig"}). We have shown that simulations, which utilized only codon usage frequencies data, produced an exceptionally good match to periodicity observed in real exons. As soon as frequencies of all codons are set as equal, DNA periodicity in exons entirely disappears. It is reasonable to think that the asymmetry in base composition, studied by Guigó \[[@B7]\], might be caused by the codon usage frequency. The results presented in this paper also demonstrate effectiveness of the evolutionary algorithm and the both models used to identify a periodicity pattern in exons. Although periods, which are seen in Table [1](#T1){ref-type="table"}, are not precisely equal to 3, they are very close. This minor discrepancy is a result of the analysis compensating for slight changes in the pattern of frequencies over nucleotide position. When the period is fixed at exactly 3, and the program allows for change-over points where the curve of best-fit is adjusted to better suit the data, the curves of best-fit still closely match the actual data points (see Figures [4](#F4){ref-type="fig"} and [5](#F5){ref-type="fig"}), revealing that the period of 3 is not simply coincidental when it is allowed to be determined by the program. As it can be seen in Figures [2](#F2){ref-type="fig"},[3](#F3){ref-type="fig"},[4](#F4){ref-type="fig"},[5](#F5){ref-type="fig"} the amplitude of variation is much more narrow for *C. elegans*than for the two other species under consideration. Introns do not show any specific period that can be determined by the analysis (results are not shown). Although the analysis does produce a period for each data set given, these periods are not consistent with each other, and the predictions do not fit the data well. As introns are not composed from codons, this is an additional indication supporting the conclusion that CUF determine periodicity pattern in exons. Since only exons show a strong periodicity of three, this type of analysis can be in principle used as an additional component of exon finding tools. Such possibility was already considered \[[@B7]\]. Unfortunately, the methods discussed here being very effective in quantifying DNA periodicity in a set of many sequences, are not sensitive enough for a single sequence. Further modifications of the approach are necessary before it can be used in exon prediction programs. Conclusions =========== Conclusion can be drawn that DNA periodicity in exons is determined by codon usage frequencies. It is essential to differentiate between DNA periodicity itself, and the length of the period equal to 3. Periodicity itself is a result of certain combinations of codons with different frequencies typical for a species. The length of period equal to 3, instead, is caused by the triplet nature of genetic code. The models and evolutionary algorithm used for characterising DNA periodicity are proven to be an effective tool for describing the periodicity pattern in a species, when a number of exons in the same phase are analysed. Methods ======= Exon-Intron Database -------------------- Information relevant to *C. elegans*, *D. melanogaster*and *H. sapiens*, was extracted from the exon-intron database (EID), which was compiled in the W. Gilbert laboratory, Department of Molecular and Cellular Biology, Harvard University \[[@B18]\]. The database contains protein-coding intron-containing genes. From the version of the database that we used, the following data were extracted: *C. elegans*14,836 genes and 98581 exons; *D. melanogaster*13,361 genes and 58,801 exons, *H. sapiens*7150 genes and 47908 exons. exScan ------ This program calculates frequencies of nucleotides or any combination of nucleotides in a database. exScan align all exons in the database at either the 5\' or 3\' end. The program then searches the exons for given sequences and give a summary of the sequences found. exScan was used in order to obtain the frequencies for nucleotides and dinucleotide pairs in each position of exons aligned at the 5\' end. A summary of the program\'s operation follows: \- Command line for exScan selects the database to be used for searching. \- The string(s) to be searched are also entered into the command line. \- ExScan then aligns the exons by the 5\' end and searches each exon for all matches to the search strings. \- The output of the program will provide the number of matches for each search string entered at every position along the aligned exons. \- These numbers are then converted into frequencies. ExScan is written in the C++ programming language; its full description and the program itself are available upon request ### GENERATE The program simulates a required number of coding sequences, using as an input file CUF for a particular species and a generator of random numbers. Inclusion of stop-codon in a coding sequence terminates the gene. There are options, which allow establishing a minimal and a maximal length of genes as well a shape of gene length distribution. GENERATE was used in this study to show how CUF alone could create periodicity, even in randomly created sequences, which do not code for any real protein. The procedure when running GENERATE follows: \- GENERATE accepts as input a file containing usage frequencies of all 64 codons. These codon usage frequencies for different species were taken from the database located at <http://www.kazusa.or.jp/codon/>. Thus, despite random choice the frequencies of codons in simulated sequences were very similar to the real CUF. \- These frequencies are then used to construct a requested number of artificial genes, with the codons chosen randomly based on their frequencies. \- Artificial genes all start with **ATG**, and will terminate once a stop codon is randomly chosen. \- The artificial genes can then be used as a separate database for analysis with exScan as above. GENERATE was written in the C++ programming language. Description of the program and program itself are available upon request. Differential Evolution ---------------------- The specific method used for fitting the two periodicity models was Differential Evolution (DE). As DE is a widely applicable method of general utility for optimization, the reader is directed to Storn and Price \[[@B13]\] for detailed description and example computer code. The concept is outlined here: \- A population of candidate solutions is established. Each population member is constituted by a randomly sampled set of *b*parameters and is characterized by its *fitness*(its value on the prevailing objective function, the sum of squared errors across nucleotide position). \- For each population member, a challenger is constructed. If this challenger has superior fitness, it will replace the population member in the next *generation*. A challenger is constructed as follows: \- Three other population members are chosen at random. We can label these as *a*, *b*and *c*. Each parameter is then addressed in turn. With probability *CR*(*CR*= 0.4 was adopted) the parameter is simply taken from the population member that the challenger is challenging. Otherwise, a new parameter value is constructed as the value for member *a*plus *F*times the *difference*of the values for *b*and *c*. For this application, *F*= 0.4, except *F*= 1 every fourth generation and F = 2 every seventh generation, to help avoid local optima. In addition, mutation independent from differences between other solutions was invoked periodically, also to help avoid local optima. \- Successful challengers replace their respective population members, and, together with surviving members constitute a new generation with higher mean fitness. The process continues over sufficient generations to achieve convergence close to an optimal solution, with the most fit solution being chosen. Competing interests =================== None declared. Authors\' contributions ======================= SE conducted the data analysis and was involved in drafting the manuscript. FE designed and wrote the code for the C++ computer programs. BK assisted with the manuscript and wrote the code for modeling and fitting periodicity. AR drafted the manuscript, performed statistical analysis, conceived the study and participated in its design and coordination. All authors read and approved the final manuscript.

PubMed Central

2024-06-05T03:55:47.441866

2004-8-18

PMC516031

Background ========== Quiescent cells exposed to growth factors respond by expressing a variety of immediate early genes (IEG) that do not need new protein synthesis for their expression \[[@B1]\]. Serum or growth factor induced expression of many of these immediate early genes, such as c-fos, egr1, cyr61 and pip92, is dependent on a sequence element in their promoter termed the Serum Response Element (SRE). This sequence element contains an A/T rich core flanked by an inverted repeat and is also known as the CArG box (CC(A/T)~6~GG). The CArG box is specifically bound by Serum Response Factor (SRF) \[[@B2]-[@B4]\]. Both the SRE and SRF are required for the serum inducibility of these genes since microinjection of SRE oligonucleotides or anti-SRF antibodies blocked induction in NIH3T3 cells \[[@B5]\]. In addition, mutation of the SRE blocked serum induction of reporter genes containing immediate early gene promoters and SRF null ES cells were defective for immediate early gene activation \[[@B6],[@B7]\]. Although the immediate early genes are so named because of their rapid inducibility after growth factor treatment, different kinetics of expression have been observed among the immediate early genes. Expression of the proto-oncogene c-fos peaks at around 30 minutes after stimulation whereas the peak expression of SRF mRNA occurs after 90--120 minutes \[[@B8],[@B9]\]. Thus SRF has been characterized as a \"delayed\" IEG although its expression is still independent of new protein synthesis. Activation of SRF by growth factors occurs through at least two mechanisms -- the TCF and RhoA pathways \[[@B10],[@B11]\]. Serum or growth factor induction leads to the phosphorylation of p62TCF by MAP kinases. TCF is a ternary complex factor that binds to both SRF and flanking sequences of the SRE. TCF binding to the SRE requires the prior binding of SRF as well as an adjacent TCF binding site. TCF is encoded by three ets-related genes, Elk1, SAP1 and SAP2/Net \[[@B12]\]. An additional pathway that activates SRF is through activation of the small GTPase RhoA \[[@B11]\]. Activated RhoA induces the expression of SRE reporter genes while inhibition of RhoA blocks serum induction. RhoA also causes the formation of stress fibers and the use of actin filament inhibitors and actin mutants suggests that actin treadmilling can control SRE activation \[[@B13],[@B14]\]. The RhoA effectors mDia and ROCK appear to be involved in regulating both actin treadmilling and SRF activation \[[@B15],[@B16]\]. This has led to a model whereby free G-actin inhibits SRF activation and this inhibition is relieved when G-actin levels are depleted by their polymerization into actin filaments. However, mutants of RhoA have been identified that are defective for SRF activation but still cause the formation of stress fibers and vice versa, suggesting that a pathway exists for RhoA activation of SRF independent of stress fiber formation \[[@B17],[@B18]\]. The use of pathway specific inhibitors has suggested the presence of two types of SRF target genes -- those that are largely dependent on the Rho-actin pathway and those that are largely dependent on the MAP kinase pathway \[[@B19]\]. Use of reporter genes suggests that some promoters may be activated by both pathways, but that activation by the Rho pathway is only apparent after mutation of the TCF pathway \[[@B20]\]. We and others have recently identified a family of SRF-specific transcriptional co-activators -- the MKL family, comprised of MKL1 and MKL2, that function downstream of the RhoA pathway \[[@B21]-[@B26]\]. MKL1 and 2 are widely expressed genes related to the heart and smooth muscle-specific SRF co-activator myocardin \[[@B21],[@B22],[@B24]-[@B29]\], thus making them candidate proteins in the signal transduction pathway of immediate early genes. Indeed, experiments using dominant negative proteins and RNA interference have shown that they are required for the induction of the immediate early genes srf, vinculin and c-fos \[[@B21],[@B23]\]. Serum induction of SRF and vinculin was dependent on MKL1 both at the early and late time points of induction, while serum induction of c-fos showed a modest MKL-dependence at the later time points of serum induction and was largely independent of MKL at the early time points of induction (30 minutes) \[[@B21]\]. Consistent with this, Gineitis et al. identified c-fos as a gene whose serum induction is largely dependent on the MAPK-TCF pathway \[[@B19]\]. Overexpression of MKL1 has different effects on specific SRF target gene promoters, as some were strongly activated (e.g. smooth muscle α-actin, SM22) while others were poorly induced (e.g. c-fos, egr-1)\[[@B21]\]. The lower activation of the c-fos promoter is not due to the c-fos SRE per se, since reporter genes containing the c-fos SRE on a minimal promoter were strongly activated by MKL1 \[[@B21],[@B27]\]. Chromatin immunoprecipitation experiments have shown that MKL1 binds to the promoters of the cyr61, srf and vinculin genes but not egr-1 or c-fos after swinholide treatment which leads to the activation of the Rho-actin pathway \[[@B23]\]. Moreover SAP-1 (a TCF factor) bound to the egr-1 promoter but not to the srf, vinculin or cyr61 promoters suggesting that TCF and MKL1 binding to target promoters might be mutually exclusive. Indeed, in vitro gel mobility shift assays suggest that MKL1 and Elk-1 bind to the same region of SRF \[[@B23]\]. Nevertheless, it is still unclear how MKL family factors distinguish SRF target genes. It is possible that there are specific CArG box sequences, that TCF sites adjacent to SREs inhibit activation or that other flanking elements control sensitivity to MKL activation. As a first step towards a better understanding of this differential target gene selection and to demonstrate the importance of the MKL coactivators in immediate early gene induction, we sought to identify the group of SRF target genes that were dependent on the RhoA-MKL pathway. We performed microarray analysis using cell lines expressing dominant negative MKL1 (DN-MKL1) to identify MKL-dependent and -independent serum-induced SRF target genes. The DN-MKL1 protein is effective in blocking both MKL1 and 2 activation of SREs, thus solving the problem of redundancy of these proteins in the cell \[[@B21],[@B24]\](unpublished data). As a control, DN-MKL1 did not affect TCF-pathway induction of reporter genes \[[@B21]\]. Hence, cell lines expressing DN-MKL1 can be used to elucidate the target genes of SRF that require MKL1/2 for their activation. We have further searched the target promoters for common regulatory elements, in particular for perfect or variant CArG box sequences to identify promoters that are more likely to be direct targets of SRF-MKL activity. Results ======= Dominant negative MKL1 inhibits serum induction of SRF target genes ------------------------------------------------------------------- We have utilized a C-terminal deletion mutant of MKL1 (a.a. 1-630) as a dominant negative mutant. We previously found that this mutant can prevent activation of SRE reporter genes by MKL1 and MKL2 overexpression \[[@B21]\](unpublished data). This DN-MKL1 mutant can bind to SRF but lacks a transcriptional activation domain. Expression of DN-MKL1 reduced serum induction of a c-fos reporter gene that lacks a TCF site and strongly blocked RhoA activation of this reporter \[[@B21]\]. We generated a cell line stably expressing DN-MKL1 in NIH3T3 cells in order to more easily look at its effect on endogenous gene expression. We found a strong effect on serum-induced expression of the known SRF target genes vinculin and the SRF gene itself \[[@B21]\]. However, binding of dominant negative MKL1 to SRF could titrate other SRF complexing proteins such as the TCF factors Elk1 and SAP-1/2, since they all bind to the MADS box DNA binding domain of SRF \[[@B23]\]. Thus there is a possibility that DN-MKL1 could block the serum induced expression of target genes by indirectly blocking the TCF pathway. We previously found that dominant negative MKL1 does not significantly inhibit tetradecanoyl phorbol acetate (TPA) activation of SRE reporter genes, which functions through MAPK phosphorylation of TCF \[[@B12],[@B21]\]. We sought to further confirm this for endogenous gene targets. Serum starved cells were induced with serum or TPA for 60 minutes and the mRNA from these cells were assayed for the expression of the endogenous target genes c-fos and junB by real-time PCR (Fig. [1](#F1){ref-type="fig"}). The DN-MKL1 cell line showed a modest decrease in serum induced expression of c-fos, consistent with our previous RNase protection measurements \[[@B21]\]. A larger decrease was observed for serum induction of junB, however there was no significant change in TPA induced expression of either c-fos or c-jun in the DN-MKL1 cell line (Fig. [1](#F1){ref-type="fig"}). Hence, DN-MKL1 does not block the TCF pathway and effectively distinguishes the MKL and TCF pathways. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Effect of serum and TPA on gene expression in WT and DN-MKL1 cell lines.**Serum starved cells were treated with new born calf serum (20%) or TPA (100 ng/ml) for 1 hour and relative mRNA levels for c-fos and jun B were measured by quantitative real-time PCR using the SYBR green method. Data are represented as the relative fold activation ± the standard deviation of the induced cells compared to the serum starved WT cells. :::  ::: Microarray analysis identifies MKL dependent genes with varying kinetics of expression -------------------------------------------------------------------------------------- In order to more fully characterize the temporal program of serum induced gene expression and the role of MKL, we performed microarray analysis on wt and DN-MKL1 NIH3T3 cells that had been serum-starved and induced with serum for 0, 30, 60 or 120 minutes. These times should distinguish early vs. \'delayed\' immediate early genes. Mouse Affymetrix oligonucleotide arrays (MOE430A) containing 14,824 non-redundant probes were used for hybridization. All data points were done in triplicate and hierarchical clustering was performed using dChip software analysis. We found 229 genes that showed a significant variation across the 8 samples and used the results with these genes for unsupervised clustering \[[@B30]\]. The hierarchical clustering analysis shows classes of differentially regulated genes which we have designated classes 1 to 10 (Fig. [2](#F2){ref-type="fig"}). We were particularly interested in serum-induced genes. However, one class was constitutively repressed in DN-MKL1 cells (\#1) and another was constitutively activated (\#5). Further, one class was repressed by serum induction (\#4) though there was no significant effect of DN-MKL1. This leaves seven classes of genes that are serum-induced with different kinetics and dependency upon MKL1. Classes 2, 3 and 8 were induced early at the 30 minute time point while classes 6 and 7 had maximal induction at 120 minutes. Classes 9 and 10 had peak induction at the intermediate time point of 60 minutes. Serum induced expression of classes 2, 7 and 9 as well as many of the genes in class 8 were particularly MKL-dependent while the other serum-induced genes (classes 3, 6 and 10) were predominantly MKL-independent. The positions of known SRF target genes that we previously studied are indicated (Fig. [2](#F2){ref-type="fig"}). The vinculin and srf genes fall into the MKL-dependent classes 7 and 9 consistent with our previous results \[[@B21]\]. Serum-induced expression of the c-fos gene was predominantly independent of MKL1. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Hierarchical clustering of gene expression.**RNA from WT and DN-MKL1 expressing cells induced with serum for the indicated times was used to probe an Affymetrix mouse chip with 14,824 non-redundant probes. dChip software was used to identify genes with significant variation across samples. 229 such genes were used for clustering analysis. Different classes of similarly regulated genes are indicated to the right and discussed in the text. The respective positions where the previously characterized MKL target genes srf, vinculin and c-fos fall are indicated by the arrows. Expression scales ranging from -3 (blue) to +3 (red) fold are indicated at the bottom of the figure. :::  ::: To analyze the serum-inducible genes more carefully, we further dissected the data using more stringent criteria to identify MKL-dependent genes at each time point. We filtered for genes that were serum-induced ≥ 2 fold in WT cells and whose expression in DN-MKL1 cells was at least 35% less than in WT cells (Tables [1](#T1){ref-type="table"}, [2](#T2){ref-type="table"} and [3](#T3){ref-type="table"}). We picked points that fell within the 90% confidence intervals for both fold serum induction and decrease in DN-MKL1 cells. At the 30 minute time point, 12 out of the 52 serum-inducible genes were MKL dependent. Similarly, 6 out of 59 and 18 out of 123 genes were MKL dependent at the 60 and 120 minute time points, respectively. The genes previously shown to contain SRF binding sites are indicated. Consolidating these results to eliminate multiple listings of genes induced at different time points, we have identified 28 out of 150 serum-inducible genes as MKL-dependent (Table [4](#T4){ref-type="table"}). A list of the 122 serum inducible, MKL-independent genes is shown in the supplemental table. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### List of genes induced by serum at 30 minutes in WT cells and reduced in DN-MKL1 cells. ::: **Gene** **Locus Link** **Fold change** **Lower CI** **Upper CI** **% decrease in DN** **known SRE** ---- ---------------------------------------- ---------------- ----------------- -------------- -------------- ---------------------- --------------- 1 similar to Hs Mitogen inducible gene 6 74155 **18.67** 12.81 26.3 **54** 2 coagulation factor III 14066 **8.05** 5.44 14.29 **47** 3 retinoblastoma inhibiting gene 1 19649 **6.45** 4.97 8.90 **63** 4 tropomyosin 1, alpha 22003 **5.92** 4.18 8.55 **63** **YES** 5 pleckstrin homology-like domain 21664 **5.82** 3.67 11.48 **65** 6 epiregulin 13874 **5.77** 3.15 9.27 **71** 7 leukemia inhibitory factor 16878 **4.84** 2.79 7.65 **57** 8 serum response factor 20807 **4.7** 3.25 6.40 **60** **YES** 9 tribbles homolog 1 211770 **3.87** 2.81 5.10 **50** 10 aortic alpha actin-2 68377 **3.77** 2.94 5.06 **52** **YES** 11 fos-like antigen 1 14283 **3.02** 2.27 3.89 **44** **YES** 12 CDC42 effector protein 3 260409 **2.47** 2.19 2.81 **46** ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### List of genes induced by serum at 60 minutes in WT cells and reduced in DN-MKL1 cells. ::: **Gene** **Locus Link** **Fold change** **Lower CI** **Upper CI** **% decrease in DN** **known SRE** --- ---------------------------------- ---------------- ----------------- -------------- -------------- ---------------------- --------------- 1 serum response factor 20807 **6.26** 4.66 8.20 **59** **YES** 2 enigma homolog 56376 **4.75** 3.63 6.28 **61** 3 adrenomedullin 11535 **4.15** 3.28 5.17 **57** 4 retinoblastoma inhibiting gene 1 19649 **3.92** 2.68 5.75 **50** 5 CDC42 effector protein 3 260409 **3.18** 2.40 4.38 **48** 6 adenosine A2b receptor 11541 **2.56** 2.00 3.24 **40** ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### List of genes induced by serum at 120 minutes in WT cells and reduced in DN-MKL1 cells. ::: **Gene** **Locus Link** **Fold change** **Lower CI** **Upper CI** **% decrease in DN** **known SRE** ---- ------------------------------------------------- ---------------- ----------------- -------------- -------------- ---------------------- --------------- 1 leukemia inhibitory factor 16878 **32.79** 16.31 54.48 **82** 2 epiregulin 13874 **21.48** 13.61 32.57 **54** 3 interleukin 6 16193 **7.84** 4.11 12.98 **90** 4 inhibitor of DNA binding 3 15903 **7.34** 5.86 9.45 **51** 5 snail homolog 1 20613 **6.14** 4.17 9.00 **55** 6 serum response factor 20807 **5.83** 4.39 7.61 **53** **YES** 7 hexokinase 2 15277 **5.43** 3.52 7.51 **52** 8 TGFB inducible early growth response 21847 **5.39** 3.88 7.42 **53** 9 enigma homolog 56376 **5.37** 4.26 6.93 **57** 10 Jun-B oncogene 16477 **5.25** 3.12 7.93 **61** **YES** 11 vinculin 22330 **4.54** 3.32 6.10 **52** **YES** 12 expressed sequence AA939927 99526 **4.38** 3.35 5.79 **76** 13 B-cell translocation gene 2, anti-proliferative 12227 **3.58** 2.34 6.09 **63** 14 adrenomedullin 11535 **3.30** 2.79 3.91 **47** 15 methionine adenosyltransferase II, alpha 232087 **3.28** 2.70 3.95 **43** 16 ELL-related RNA polymerase II 192657 **3.15** 2.34 4.56 **39** 17 zyxin 22793 **2.70** 2.09 3.37 **46** 18 transmembrane 4 superfamily member 10 homolog 109160 **2.32** 2.11 2.56 **39** ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### List of genes induced by serum at 30, 60 or 120 minutes whose induction is MKL-dependent or -independent. ::: **MKL-Dependent genes (28)** --- ---------------------------------------- ---- ------------------------------------------- ---- -------------------------------------------------- **Known SRF target Genes** **Other Genes** **Other Genes** 1 Jun-B 6 adrenomedullin 18 Adenosine A2b receptor 2 serum response factor 7 B-cell translocation gene 2 19 Coagulation factor III 3 fos-like antigen 1 8 CDC42 effector protein 3 20 Leukemia Inhibitory factor 4 tropomyosin 1, alpha 9 enigma homolog 21 Retinoblastoma inhibiting gene 1 5 vinculin 10 epiregulin 22 TGFβ inducible early growth response 11 hexokinase 2 23 Tribbles Homolog 1 12 inhibitor of DNA binding 3 24 Aortic alpha actin-2 13 interleukin 6 25 expressed sequence AA939927 14 pleckstrin homology-like domain, family A 26 transmembrane 4 superfamily member 10 homolog 15 similar to Hs Mig6 27 methionine adenosyltransferase II, alpha 16 snail homolog 1 28 ELL-related RNA polymerase II, elongation factor 17 zyxin **MKL-Independent genes (122)** **Known SRF target genes** 1 cysteine rich protein 61 (**Cyr61**) 2 thrombospondin 1 3 FBJ osteosarcoma oncogene (**c-fos**) 4 FBJ osteosarcoma oncogene B (**FosB**) 5 cysteine rich protein 1 (**Crp1**) 6 prostaglandin-endoperoxide synthase 2 7 early growth response 1 (**Egr1**) 8 early growth response 2 (**Egr2**) ::: Correlation of microarray and real time PCR data ------------------------------------------------ To verify the results obtained from the microarray experiments, we checked seven genes by real time quantitative PCR using the SYBR Green method. There was a good, but not perfect correlation, of these methods (Fig. [3](#F3){ref-type="fig"}). Four MKL-dependent genes, jun B, srf, interleukin 6 and epiregulin, were also found to be MKL-dependent by real time PCR. Similarly, two MKL-independent genes, cyr61 and egr-1, were also MKL-independent by real time PCR. Finally, c-fos expression was only moderately reduced in DN-MKL1 cells whether measured by microarray or real time PCR (Fig. [3G](#F3){ref-type="fig"}). This inhibition was not high enough to meet the cut-off of 35% reduction for our list of MKL-dependent genes (Fig. [4](#F4){ref-type="fig"}). This result contrasts somewhat with our measurements of c-fos mRNA by RNase protection. We previously found that c-fos mRNA was significantly reduced in DN-MKL1 cells at the 60 minute time point, though not the 30 minute point. Since there was inhibition noted by each method, this difference may reflect the sensitivity of each method to detect precise changes in expression. There were also some differences between the microarray and real time PCR results such as inhibition of Jun B expression at the 30 minute point by the PCR method but not by microarray (Fig. [3A](#F3){ref-type="fig"}) and the greater serum induction of egr-1 mRNA measured by real time PCR (Fig. [3F](#F3){ref-type="fig"}). Nevertheless, the general levels of induction and DN-MKL1 inhibition were similar by both methods. This is reflected in the Pearson\'s correlation coefficient of 0.92 for a comparison of all of the data attained by each method shown in Fig. [3](#F3){ref-type="fig"}. The correlation coefficient ranged for 0.85 to 0.99 for the data with any one gene. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Correlation of microarray and real-time PCR data**. The expression pattern of select serum-induced MKL-dependent and -independent genes was determined by quantitative real-time PCR (right) and compared to the microarray results (left) for the indicated genes. WT or DN-MKL1 cells were induced with serum for the indicated times before isolation of RNA. The results derived from the microarray hybridizations are the averages of triplicates while the real-time PCR measurements are the averages of at least duplicates ± the standard deviation. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Known CArG boxes in serum inducible genes.**The upper panel lists the positions and sequences of the known CArG boxes of the MKL-dependent or -independent genes. The bases that differ from the CArG box consensus sequence are in bold. The bottom panel shows the multilevel consensus sequence that was derived from each of these groups of CArG boxes. Below the consensus sequence is the simplified position-specific probability matrix that specifies the probability of each possible base appearing at each position in an occurrence of the motif multiplied by 10. \'a\' denotes a probability that is almost or equal to 1. The consensus sequence is the best match to the CArG boxes oriented on either strand. :::  ::: Promoter analysis of MKL-dependent and -independent serum-inducible genes ------------------------------------------------------------------------- We compared the serum inducible genes to known SRF target genes to identify those genes with SRF binding sites. Comparing to a list of SRF targets genes \[[@B31]\](J.M. Miano, personal communication) we found that five of 28 MKL-dependent genes had known CArG box sites while eight of 122 MKL-independent genes had known sites (Table [4](#T4){ref-type="table"}). We further analyzed the promoters of the MKL-dependent genes to identify SRF binding sites that would potentially be direct targets of MKL. Promoter sequences for the serum inducible MKL-dependent or -independent genes were extracted from the Database of Transcription Start Sites (DBTSS) \[[@B32]\]. This database contains exact information of the genomic positions of the transcriptional start sites (based on full length cDNAs) and the adjacent promoters for 6,875 mouse genes. The upstream sequence (-1000 to +200 relative to the transcription start site) of each of the candidate genes was searched for the consensus SRF binding site (CCWWWWWWGG, where W is A or T) with or without one base mismatch allowed. Of the 28 MKL-dependent genes, upstream sequences were available for 20 genes in the DBTSS database while sequence was extracted for 75 of the 122 MKL-independent genes. There were 17 exact matches to the CArG box in these 95 serum inducible promoters (18%) (Table [5A](#T5){ref-type="table"}). Given multiple matches in some genes this resulted in 11 genes (12%) having exact matches. These frequencies were similar in the MKL-independent and -dependent classes of serum-inducible genes but significantly higher than the frequencies found in searching the entire DBTSS collection of promoters as indicated by the p values in Table [5A](#T5){ref-type="table"}. The p value for the MKL-dependent promoters is high because of the small sample size, however the frequency of exact CArG box matches is similar to the other serum-inducible genes. ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Search of MKL-dependent and -independent serum-inducible promoters for SRF binding sites. ::: **A. Exact matches to CArG box** ------------------------------------------------------- ---------------------- ---------------- --------- ----------- -------- -------------------- ----------- -------- ---------------------- ----------- -------- genes in micro-array genes in DBTSS matches frequency p genes with matches frequency p genes w. \>1 matches frequency p Total genes 14824 6875 419 0.061 401 0.058 15 0.002 All Serum-inducible 150 95 17 0.179 0.0001 11 0.116 0.0218 3 0.032 0.0010 MKL-indep, serum-inducible 122 75 14 0.187 0.0002 9 0.120 0.0293 2 0.027 0.0101 MKL-dep, serum-inducible 28 20 3 0.150 0.1193 2 0.100 0.3245 1 0.050 0.0392 **B. Matches to CArG box allowing one base mismatch** Total genes 14824 6875 6707 0.98 4230 0.615 1744 0.254 All Serum-inducible 150 95 103 1.08 0.155 63 0.663 0.1958 25 0.263 0.4576 MKL-indep, serum-inducible 122 75 72 0.96 0.839 47 0.627 0.4682 15 0.200 0.8886 MKL-dep, serum-inducible 28 20 31 1.55 0.005 16 0.800 0.0666 10 0.500 0.0155 The promoters from indicated groups were searched for exact matches with the CArG box sequence CC(A/T)~6~GG (A) or allowing for one base mismatch (B). Frequency indicates the number of matches divided by the number of DBTSS genes in each category. The p values were calculated based on a binomial distribution except for the value for matches in (B) (column 6 from the left) where the numbers were too large and the Poisson distribution of the frequency of matches per base was used. ::: We also searched for promoters with more than one CArG box match since multiple SREs have been reported in several SRF target genes and this could provide a mechanism to distinguish responses. While there were more genes with greater than one CArG box match in the serum-inducible genes than in the DBTSS promoters, the numbers of MKL-independent vs. -dependent genes with greater than one match were too low to show a statistical significance (Table [5A](#T5){ref-type="table"}). The search for SRF binding sites allowing a one base mismatch was more problematic since there were a high rate of matches in the DBTSS data set, an average of .98 hits per gene (Table [5B](#T5){ref-type="table"}). There was not a significant increase in this rate in the MKL-independent promoters, however there was an increase in the MKL-dependent promoters to 1.55 hits per gene which was found to be statistically significant (p = 0.005; Table [5B](#T5){ref-type="table"}). This difference in the MKL-dependent set was also apparent when considering the increased frequency of genes with matches and of genes with greater than one match (Table [5B](#T5){ref-type="table"}). The increased frequency of the exact CArG box sequence in the upstream region of many of the MKL-dependent genes gives further confidence that SRF coordinately regulates many immediate early genes. However, the absence of a CArG box does not preclude a gene from being a direct target due to the flexibility of enhancer positioning. The regulatory site may be upstream or downstream of the 1200 bp we have analyzed. For instance, in the case of the jun B gene it has been shown that a CArG box downstream of the gene mediates its response to serum \[[@B33]\]. In addition not all SRF target sites are perfect CArG boxes. Several of the known SRF target sites contain single base changes (Fig. [4](#F4){ref-type="fig"}). While our search allowing a single mismatch resulted in an overly high frequency of matches, there was a significantly higher rate of matches in the MKL-dependent promoters suggesting that many of these sites may be used as direct SRF targets. We compared the CArG boxes more closely since previous work with altered SREs suggested that changes in the A/T core of the SRF can affect sensitivity to the TCF-independent (i.e. RhoA-MKL) pathway \[[@B34]\]. The known CArG boxes were aligned from each class of genes, either MKL-dependent or -independent (Fig. [4](#F4){ref-type="fig"}). We chose not to analyze the predicted CArG boxes since there were few exact matches in the MKL-dependent promoters and because there was too high a rate of matches allowing one mismatch such that many of the matches are likely false positives. The MEME motif discovery tool (see Methods) was used to arrive at a multilevel consensus sequence based on a position specific probability matrix. The matrix specifies the probability of each possible letter appearing at each position in an occurrence of the motif. MEME also takes into account both orientations of the sequence to arrive at the consensus sequence. The consensus sequences for the two sets are quite similar although there are some minor differences and the sample size is relatively small. We have compared the sequences of the known SRF target genes to determine whether flanking sequences can explain the differential sensitivity to DN-MKL1. We first compared sequence flanking the known CArG boxes. The MEME motif discovery tool was used to identify common elements in 30 bp flanking each side of the CArG boxes. There were no statistically significant matches in the sequences from MKL-dependent of -independent genes. We might have expected to see a TCF site, however this site is very short and flexible. TCF binds to a short site near an SRE and its binding is stabilized by binding to SRF. A consensus of (C/A)(C/A)GGA(A/T) was found, however the orientation and position relative to the SRE was flexible and only the GGA sequence was absolutely required \[[@B35]\]. Visual inspection of the flanking sequence to the known CArG boxes suggested a potential TCF site in all of the sequences except for that of cyr61. We used two methods to compare the full promoter sequences of the MKL-dependent and -independent, serum-inducible genes for common elements that might be required for their induction. We first used the MEME tool to compare the promoters (-1000 to +200) derived from the DBTSS database. As in Table [4](#T4){ref-type="table"}, 20 MKL-dependent promoter sequences and 75 MKL-independent promoter sequences were extracted. The MEME tool could only compare 50 of these promoters at a time such that 50 MKL-independent promoters and 50 random promoters were compared. While MEME identified some common elements within each of these three classes of promoters, none of the elements identified among the MKL-dependent of -independent promoters appeared to be specific since similar elements were identified in the random promoter set. A second method we used was to look for common regulatory elements was oligonucleotide analysis which has been used in yeast to identify regulatory sites \[[@B36]\]. This method looks for enriched oligonucleotide frequencies in a group of genes. We used the entire set of 6875 DBTSS mouse promoter sequences to calculate the expected background frequencies of hexamers. The most specific matches we found were in the set of 75 MKL-independent promoters. The related oligonucleotides GGAGGG, CCGGAG and CGGAGA were enriched with significance values (E-value) of 1.4 × 10^-4^to 6.9 × 10^-3^. These oligonucleotides were not enriched in a test case of 60 random promoter sequences. The related sequence GGGAGG, however, was similarly enriched in the MKL-dependent promoters albeit with less statistical significance (E = 0.8). It is intriguing that these oligonucleotides are similar to the TCF consensus binding site (C/A)(C/A)GGA(A/T). The most significantly enriched oligonucleotide in the 20 MKL-dependent promoters was CCGCGC with an E value of 0.078. This lower significance may partially reflect the smaller size of this data set but also may reflect the lower significance of this enrichment. Given the relatively small number of serum inducible genes and present computational tools, it appears that more careful experimental mapping of the sequence elements in MKL-dependent and -independent genes will be required to identify the elements that determine their sensitivity to the MKL pathway. Discussion ========== Serum inducible genes have far ranging effects in many physiological processes such as proliferation, wound healing, migration, and tissue remodeling \[[@B37]\]. SRF has been implicated in the expression of many serum inducible genes, particularly immediate early transcription factors, but is also required for the expression of many muscle-specific genes \[[@B31]\]. SRF is activated by different co-activators to control its diverse set of target genes in different tissues \[[@B27],[@B38]-[@B45]\]. Here, we have elucidated the target genes that are dependent on one particular SRF co-activator family -- the MKL family. Since dominant negative MKL1 can block activation by all of the members of the MKL family, MKL1 and 2, as well as myocardin (unpublished data)\[[@B21],[@B24]\], we have used cell lines expressing dominant negative MKL1 to determine the role of the MKL family in serum-induced gene expression patterns. Cluster analysis of the microarray data identified different classes of genes that showed significant variation across the samples. Two classes of genes were either constitutively activated or constitutively repressed. These could be indirect targets of the SRF-MKL pathway or could suggest functions for MKL apart from its role as a transcriptional co-activator of SRF. Among the serum inducible genes, we identified a significant fraction, 28 of 150, that are MKL-dependent. This is not to say that the other 122 genes may not be sensitive to MKL under certain conditions. Rather, they must at least have additional mechanisms for serum induction in NIH3T3 cells. c-fos was identified as an MKL-independent gene although we previously found by RNase protection that its induction was significantly reduced in cells induced with serum for 60 minutes \[[@B24]\]. We did observe a small decrease in c-fos expression in our microarray experiments, consistent with the RNase protection results, however it did not pass our stringent statistical criteria for MKL-dependence. Serum induction of some of the MKL-independent immediate early genes may also be independent of SRF. For example, the c-jun gene utilizes other sequence elements for serum induction \[[@B46]\]. The TCF factors, elk-1, SAP1 and SAP2, are the main candidates besides MKL factors for mediating serum induction of SRF target genes. Further expression profiling will be required with inhibitors of the SRF and TCF pathways to better characterize the pathways used by the 150 serum inducible genes described here. Many of the serum-inducible genes identified here have been previously described as immediate early genes either as single genes or in microarray experiments \[[@B37],[@B47],[@B48]\]. The groups of genes include immediate early transcription factors and cytokines, as previously noted \[[@B37]\], but also a broad array of other types of genes. SRF null mice die in utero due to the defective formation of a mesodermal layer and SRF null ES cells are defective in focal adhesion and migration \[[@B49],[@B50]\]. We have also observed that the DN-MKL1 cell line is significantly less adherent than the control line (unpublished data). It is therefore interesting that a number of our serum inducible genes are involved in cytoskeletal structure and adhesion. These include vinculin, tropomyosin, zyxin, thrombospondin, tenascin, integrin α5, and transgelin. It will be interesting to determine whether lowered expression of these genes in SRF or MKL deficient cells leads to the changes in their adhesive properties. Of these proteins, only vinculin is a previously known MKL and SRF target gene \[[@B19],[@B21],[@B51]\]. Our microarray analysis found that zyxin is also an MKL-dependent gene and zyxin expression was previously found to be decreased in SRF null cells \[[@B50]\]. Sequences required for MKL-dependent and -independent genes ----------------------------------------------------------- We analyzed the promoter sequences of the MKL-dependent and -independent genes to identify elements that might determine this sensitivity. Exact matches to a simple CArG box consensus, CC(A/T)~6~GG, resulted in matches in only about 12% of the serum-inducible genes with similar proportions in the MKL-dependent and -independent classes. While some functional SRF binding sites contain a single base mismatch to the consensus, a broader search for CArG boxes allowing a mismatch resulted in a high number of matches in a broad promoter database. There was a modest, statistically significant increase in the matches in the MKL-dependent promoters suggesting that some of these sites are real SRF targets. Other SRF target sites may lie outside of the promoter regions we have searched (-1000 to +200). This still leaves the possibility that induction of many of the immediate early genes is independent of SRF such as we have found for the c-jun gene \[[@B46]\]. Microarray analysis of serum-inducible genes with inhibition of SRF activity is required to show which genes are SRF dependent. Possible mechanisms for determining sensitivity to MKL are the sequence of the CArG box, flanking sequence or the context of other promoter elements. Our comparison of the known CArG boxes from MKL-dependent and -independent promoters did not show a strong difference. Similarly we were not able to identify differences in the flanking sequence. Analysis of the more full promoter sequences has also not yielded clear common regulatory elements to date. One set of oligonucleotides, containing the sequence GGAG, was found in both the MKL-dependent and -independent promoters at a significantly higher frequency than in the full promoter database. The significance of these sites remains to be determined though it is notable that they are similar to the TCF consensus site. One possible mechanism for distinguishing MKL target genes is the presence of a TCF site next to a CArG box. This would allow for TCF binding and activation by a MAP kinase pathway instead of the MKL pathway. Since the binding sites for TCF and MKL on SRF are similar \[[@B23]\], TCF binding could preclude MKL activation. In fact, mutation of the TCF site in the c-fos promoter allows it to be activated by a serum induced, actin filament-dependent pathway presumably through RhoA and MKL \[[@B20]\]. In addition, we found that mutation of the TCF site in the c-fos SRE results in significantly higher activation by MKL1 (Bo Cen and R.P., unpublished results). Nevertheless, c-fos SRE elements with TCF sites are still activated by MKL1 and myocardin \[[@B21],[@B27]\] such that there may be additional elements that determine sensitivity to MKL1 activation. Chromatin immunoprecipitation experiments have shown that MKL1 can bind to the SRF, vinculin and cyr61 promoters in cells treated with the actin inhibitor swinholide A \[[@B23]\]. This is consistent with our identification of SRF and vinculin as MKL target genes, but we found that induction of cyr61 was MKL-independent. It is possible that this difference is due to our inducing with serum rather than swinholide A, since serum will induce alternative pathways such as the MAP kinase pathway. Chromatin immunprecipitations showed that the TCF factor SAP1 bound to the egr1 promoter, possibly explaining its independence of MKL, but no binding of SAP1 was observed for two other MKL1-independent genes egr-1 and cyr61 \[[@B23]\]. Thus, TCF binding may not always explain the lack of requirement for MKL factors. Conclusions =========== Our results indicate that a subset of serum inducible genes is dependent upon the MKL family for its induction. This genomic classification of MKL-dependent and -independent serum-inducible genes is a significant step for characterizing which pathways are required for induction of each cellular immediate early gene. Methods ======= Gene expression analysis ------------------------ NIH3T3 cells stably expressing dominant negative MKL1 (a.a. 1-630)(DN-MKL1) or containing the vector, pBabePuro, were previously described \[[@B21]\]. The cells were serum starved in Dulbecco\'s modified Eagle\'s medium (DMEM) with 0.2% new born calf serum (NCS) for 24 hours and then induced with 20% NCS for 30, 60 and 120 minutes. Total RNA from these cells was prepared using Trizol (Invitrogen) and then purified using RNeasy columns (Qiagen). Total RNA (8 μg) was used for first-strand cDNA synthesis using T7-Oligo-dT primers and Powerscript reverse transcriptase (Invitrogen) for 1 hour at 42°C. This was followed by second-strand synthesis for 2 hours at 16°C using RNase H, *E. coli*DNA polymerase I, and *E. coli*DNA ligase (Invitrogen). The obtained double-stranded cDNA was then blunted by the addition of 20 units T4 DNA polymerase and incubation for 5 min at 16°C. The material was then purified by phenol:chloroform:isoamyl alcohol extraction followed by precipitation with ammonium acetate and ethanol. The cDNA was then used in an in vitro transcription reaction for 6 hours at 37°C using a T7 in vitro transcription kit (Affymetrix) and biotin-labeled ribonucleotides. The obtained cRNA was purified on an RNeasy column. The eluted cRNA was then fragmented by incubation of the products for 30 min in fragmentation buffer (40 mM Tris-acetate, pH 8.1, 100 mM KOAc, 30 mM MgOAc) at 95°C. The fragmented labeled RNA (15 μg) was hybridized to an Affymetrix mouse MOE430A Genechip at 45°C according to the manufacturer\'s protocol \[[@B52]\] and stained with streptavidin-phycoerythrin. The chips were then scanned with an Affymetrix Genechip Scanner GS300. All the data sets were done in triplicates from serum stimulation of cells to scanning of the microarrays. The percent of genes with significant expression (\"present\" calls) ranged from 50.4 to 56.5% for each microarray. The 3\':5\' ratio for a control actin gene ranged from 1.0 to 1.31 for each of the 24 microarrays scanned. The scaling factor to account for differences in probe labeling and general changes in hybridization in each microarray was scaled to a target intensity of 250 (using Affymetrix Microarray Suite 5.0 software) and ranged from 1.273--5.542. dChip analysis -------------- The dChip software \[[@B30],[@B53]\] was used for the normalization and calculation of model based expression indexes (MBEI) after pooling the replicate arrays and for the hierarchical clustering analysis of the data from the Affymetrix Gene chips. Normalization was based on a large set of probes determined iteratively to be invariant across the different microarrays. After normalization each array has a similar overall signal \[[@B30]\]. The initial analysis was performed using the perfect match (PM)-only model and genes were filtered according to the criteria of a) co-efficient of variation (standard deviation/mean) across the eight conditions must be between 0.50 and 10.00, b) a gene must be called \'Present\' in ≥ 20% of the arrays used, c) the variation of the standard deviation/mean for replicate arrays for a single condition must be between 0 and 0.5, and d) the expression level must be \>= 100 in at least one of the time points. Hierarchical tree clusters were then generated using this filtered gene list. Expression values (MBEIs) at each time point were determined with a standard error of the mean. These values were used to calculate the fold change with 90% confidence intervals. The confidence intervals were derived from the standard errors and fold changes based on a χ^2^distribution model with one degree of freedom \[[@B30]\]. For the lists of serum inducible genes the low confidence interval value was required to be greater than two. For the inhibition in the DN-MKL cells, the low confidence interval value was required to be at least 35% less than in the WT cells for the gene to be designated MKL-dependent. Real time PCR ------------- Total RNA was isolated from the DN-MKL1 and vector containing NIH3T3 cells as described for the microarray analysis and 1 μg was used for first strand cDNA synthesis (Powerscript Reverse Transcriptase, BD Biosciences) using oligo-dT primers according to the manufacturer\'s protocols. One fiftieth of the reverse transcription reaction was included in a 20 μl PCR reaction. For a quantitative analysis, SYBR green PCR technology was used (Applied Biosystems). Real-time detection of the PCR product was monitored by measuring the increase in fluorescence caused by the binding of SYBR green to double-stranded DNA with an ABI PRISM 7000 Sequence Detector. To calculate relative quantification values a threshold cycle (C~t~), the cycle at which a statistically significant increase in fluorescence occurs, was derived from the resulting PCR profiles of each sample. C~t~is a measure of the amount of template present in the starting reaction. To correct for different amounts of total cDNA in the starting reaction, C~t~values for an endogenous control (Acidic ribosomal phosphoprotein P0) were subtracted from those of the corresponding sample, resulting in ΔC~t~. The ΔC~t~value of the serum starved wt sample was chosen as the reference point and was subtracted from the ΔC~t~value of all the other samples, resulting in ΔΔC~t~. The relative quantification value is expressed as 2^-ΔΔCt^giving the relative difference of the serum induced points compared to the serum-starved cells for expression of a particular gene. All real time PCR data sets shown are the results of two independent mRNA preparations and amplifications. Amplification of only a single species in each PCR reaction was determined by checking for a dissociation curve with a single transition. Promoter sequence analysis -------------------------- The Database of Transcriptional Start Sites (DBTSS)\[[@B32],[@B54]\] was used to extract the promoter sequences (-1000 to +200) of all the available full length cDNAs and RefSeq entries based on the start site of the cDNA as +1. This mouse data base was searched for SRF binding sites using the consensus CCWWWWWWGG, where W is A or T, with or without allowing for one base mismatch. The CArG box sequences of the known SRE regulated genes were derived from the literature. The motif discovery tool MEME (Multiple EM for Motif Elicitation) \[[@B55],[@B56]\] was then used to derive a multilevel consensus sequence based on a position specific probability matrix. The matrix specifies the probability of each possible nucleotide appearing at each possible position in an occurrence of the motif. MEME takes into account both orientations of the sequence to arrive at the consensus sequence. Sequences flanking the known SREs (30 bp on each side) were extracted from Genbank and MEME was also used to compare these sequences. A control set of 10 upstream sequences from the same genes (70 bp in length) was also searched for comparison. Promoter sequences (-1000 to +200) of the serum inducible genes were also searched for common elements with MEME. Promoters for 20 MKL-dependent, 50 MKL-independent and 50 random promoters were searched. These promoter sequences were additionally searched by oligonucleotide analysis for enriched hexamers \[[@B36],[@B57]\] except that 20 MKL-dependent and 75 MKL-independent were searched. The background frequency of oligonucleotides was set using promoter sequence (-1000 to +200) for all 6875 mouse promoters in DBTSS. A control searching 60 random promoters also did not show significant enrichment of specific oligonucleotides. Authors\' contributions ======================= AS carried out all experimental sections of the paper. AS and RP conceived of the study and participated in its design and coordination. Both authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 **List of serum inducible MKL-independent genes**This Microsoft Excel spreadsheet file contains a list of genes that were serum inducible (\>2-fold) at either the 30, 60 or 120 minute time points and that satisfied the 90% confidence interval criteria for fold-change using the dChip software. The expression of these genes were reduced by less than 35% in the DN-MKL1 cells at each time point. The MKL-dependent genes (reduced by more than 35%) shown in Tables 1 to 3 were removed from this list to give 122 MKL-independent genes. The Affymetrix probe set information, the gene name, Genbank Accession number and Locus link identifier are shown for each of the genes. For each time point the fold induction in serum-treated wt NIH3T3 cells compared to serum-starved wt cells (\'fold change\') and its upper and lower confidence intervals are shown. The time points where serum induction was deemed significant for each gene are marked by asterisks (\*) in the \'filtered\' columns. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ This work was supported by National Institutes of Health Grant CA50329 to R.P. We thank Harmen Bussemaker for help in promoter sequence analysis.

PubMed Central

2024-06-05T03:55:47.444446

2004-8-25

PMC516032

Background ========== The frequency of patient visits to general/family practice is affected by a range of factors including patient characteristics, physician/practice factors and broader issues such as access to services. A major contributor to the variance in visit rates to general/family practitioners is the complexity of health problems experienced by the patient. Those patients who attend general practice frequently report poorer health than those who attend less frequently \[[@B1],[@B2]\]. Not all health problems however, have an equal effect on patient visit rates to general practice. Numerous studies have found that for a subset of patients, the presence of psychological health problems is associated with frequent attendance at medical services \[[@B3]-[@B6]\]. Higher depression scores and increased levels of health anxiety have been found among patients who attend general practice frequently \[[@B3]\], and practitioners are more often involved in managing psychosocial issues for frequent attenders than other patients \[[@B1]\]. Frequent attendance at general practice has also been found to be associated with an increased use of other health services such as out of hours services and \"inappropriate attendance\" at accident and emergency wards \[[@B6]-[@B8]\]. The general practitioner (GP) also influences patients\' attendance rates through doctor-initiated visits and some practitioners attract a larger proportion of frequent attenders than others \[[@B9]\]. Remuneration may also affect GP behaviour. For example, GPs paid a flat rate capitation per patient may be more motivated to manage their own workload by reducing patient return visits \[[@B10]\]. Positive aspects to increasing general practice attendance includes recalling patients for chronic conditions as part of a program of structured care \[[@B11],[@B12]\]. Patients who attend the GP frequently often receive improved continuity of care \[[@B9]\], while those who attend general practice infrequently may be receiving less than optimal care, especially where geographical barriers to access result in lower utilisation of general practice services \[[@B13],[@B14]\]. The available evidence on visit frequency to general/family practice comes from a range of settings in several countries. There are differences in general practice between countries that may affect patterns of patient visits and some studies are based on a limited number of practices, which may affect the generalisability of the findings \[[@B3],[@B15]\]. Although the association between frequent medical visits and increased psychological health problems has been well-researched, the effect of other specific chronic conditions on general practice visit rates is less clear. In Australia there are few studies that examine the relationship between patient morbidity, social and demographic factors and the frequency of visits to general practice. We therefore investigated patient sociodemographic and morbidity factors associated with patient self-reported visit frequency among general practice patients in Australia, as part of an ongoing national study of general practice activity. Methods ======= The study reported here is part of the Bettering the Evaluation and Care of Health (BEACH) program, a national study of general practice activity in Australia. The BEACH method has been described in detail elsewhere \[[@B16]\]. In summary, BEACH is a continuous cross-sectional study, which commenced in April 1998. A random sample of approximately 1,000 GPs is recruited throughout each year in a rolling sample. The GPs are selected from a sampling frame of all Australian GPs who claimed 375 or more general practice (A1) items from the Medicare Benefits Schedule in the previous quarter. Each GP provides details for 100 consecutive patient encounters. Design ------ This paper is based on a sub-sample of GPs who participated in BEACH between August 2001 and March 2002. For a subset of 30 out of the 100 encounters, the GP recorded, in addition to the problems managed at the encounter, any other major health problems of the patient that had not been managed at the current encounter. GPs were instructed to include chronic illnesses requiring ongoing care, past problems that affect present and future care and social problems that influence health. As many as 12 extra problems could be recorded per patient and when added to the maximum of four problems managed at the encounter this allowed up to 16 health problems to be recorded per patient. The GP recorded problems in free text, which were secondarily classified according to the International Classification of Primary Care (ICPC-2) using an extended vocabulary of terms \[[@B17],[@B18]\]. ICPC-2 includes components for diagnoses and symptoms, and process codes (e.g. \"Test results\", \"Prescription renewal\"). Process codes were excluded from counts of health problems, unless they clearly identified the nature of the underlying condition. Synonymous or related problems were grouped according to standard ICPC-2 categories to ensure that problem categories included all patients with the disorder \[[@B16]\]. Morbidity variables ------------------- Mental health (particularly depression), asthma, diabetes, osteoarthritis and cardiovascular disease are five National Health Priority Areas that are major contributors to mortality, morbidity and health service costs in Australia \[[@B19]\]. They are also chronic health problems commonly managed in general practice \[[@B20]\]. In addition hypertension, back complaint and oesophageal disease are among the most common problems managed in general practice \[[@B20]\]. We therefore focussed on these health areas of particular relevance to general practice when investigating the association between morbidity and frequency of patient visits to a GP. Number of annual visits to general practice ------------------------------------------- The GP asked each patient in the sub-sample to recall how many times he/she had visited a GP in the previous 12 months. Demographic variables --------------------- Commonwealth concession health care cards are available to people on limited incomes and entitle the holder to health services at greatly reduced cost. In 1998, around one third of Australians aged 15 years and over held a concession health care card \[[@B21]\]. Concession health care card status was included in the analysis as a marker of socioeconomic status \[[@B22]\]. The Remote Index of Australia (ARIA) was used to classify patients according to geographical remoteness and accessibility of services \[[@B23]\]. Patients were classified by patient residential postcode into three broad categories: those from areas that were \"highly accessible\" to services, those from areas that were \"accessible/moderately accessible\" and those from areas \"remote/very remote\" from services. These three categories were used to investigate the association between geographical barriers to access and the frequency of patient visits to a GP. Statistical analysis -------------------- The patient sample was a single stage cluster design with the GP as the primary sampling unit. Observations recorded by the same GP were therefore not independent and the statistical analysis adjusted for the correlation between patients within each cluster. We used procedures in SAS software V8.2 that adjust the standard error for the intra-cluster correlation \[[@B24]\] and all reported p-values and confidence intervals in both the descriptive and multivariable analyses include this adjustment. Multiple regression ------------------- Number of annual visits to a GP was the main outcome of interest. After checking the linear relationship between number of visits and other ordinal variables (age and number of health problems), multiple regression was performed in two stages to identify predictors of visit frequency. 1\) Demographic predictors We performed multiple regression, with self-reported annual GP visits as an ordinal outcome to identify the independent demographic predictors of the frequency of visits to a GP after controlling for morbidity. We fitted the morbidity covariate as the total number of major health problems recorded for the patient. 2\) Morbidity predictors We fitted separate regression models to estimate the effect of specific morbidity on visit frequency, after adjusting for patient age, sex, other significant demographic predictors and the number of other health problems of the patient. The morbidities of interest were chronic problems most commonly managed in general practice, specifically depression, anxiety, back problems, diabetes, hypertension, asthma, ischaemic heart disease, and osteoarthritis \[[@B16]\]. We interpreted the partial regression co-efficient of morbidity in each model as the mean difference in annual GP visits between patients who had the problem versus those who did not, when other factors were kept constant. A large number of models was fitted, therefore an alpha of 0.01 was used as the test of significance to reduce type I error. Results ======= Three hundred and seventy nine GPs participated in the sub-study. Six hundred and eighteen patients (5.4%) did not answer the question on visit frequency, giving a final sample of 10,755 respondents. Patient respondents and non-respondents were not significantly different in mean age, sex distribution or mean number of health problems recorded. While all states of Australia were well represented in the GP sample, the sample had a smaller proportion of GPs aged less than 35 years and was somewhat more urban than the population of Australian GPs (Table [1](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Comparison of general practitioner (GP) characteristics for the sample and the population of Australian general practitioners in 2001. ::: **GP sample** **Australian population of GPs**^(a)^ ----------------------------- --------------- --------------------------------------- N 379 17,534 Male 61.4% 67.5% Age group \<35 years 8.2% 12.1% 35--44 29.3% 27.1% 45--54 36.7% 32.0% 55 + 25.9% 28.8% Urban/Metropolitan practice 76.0% 72.6% Solo practice 16.1% Not available 5 plus GPs in practice 45.2% Not available ^(a)^defined as practitioners who claimed the equivalent of 1,500 general practice (A1) Medicare items of service in 2001 (Medicare Benefits Schedule unpublished data, Australian Department of Health and Ageing) ::: Patients aged less than 25 years were under-represented in the sample relative to all patients who had visited a GP in Australia at least once in 2001, while patients aged 65 years and older were over-represented (Table [2](#T2){ref-type="table"}). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Characteristics of the patient sample versus the general practice patient population in Australia in 2001. ::: **Sample n (%)** **Australian population of general practice patients^(a)^%** --------------------------------------------- ------------------ -------------------------------------------------------------- Female patient 6,404 (59.5) 53.4 Age group \<15 years 1,360 (12.7) 19.9 15--24 1,005 (9.4) 13.1 25--44 2,875 (26.9) 29.7 45--64 2,677 (25.0) 23.8 65--74 1,274 (11.9) 7.6 75+ 1,514 (14.1) 5.9 Address in \"highly accessible\" location 8,702 (83.3) Not available Speaks language other than English at home. 890 (8.3) Not available Holds concession health care card 4,448 (41.4) Not available ^(a)^Defined as persons who claimed at least one general practice (A1) Medicare item in 2001 (Medicare Benefits Schedule unpublished data, Australian Department of Health and Ageing) ::: The majority of patients in the sample (83.3%) were from locations where services were \"highly accessible\", and 8.3% spoke a language other than English at home. The mean patient self-reported visit rate for the sample was 8.8 (95% CI: 8.3--9.2) visits to a GP per year. On average the GPs recorded 2.9 morbidities per patient, and half (51.1%) of the recorded problems were being managed by the GP at the current consultation (results not tabled). There was no difference in self-reported visit frequency between the sexes (p = 0.11) (results not tabled). The mean visit rate for children aged less than 5 was 5.8 visits per year, which fell to 4.3 for children 10--14 years of age, then increased linearly after age 15 years (p \< .0001). There was a simple linear increase in frequency of visits with increasing number of recorded health problems (p \< .0001). Therefore age was fitted as a categorical variable and number of problems as a numeric variable in the multiple regression models. After adjusting for age and sex, there was no difference in visit frequency according to language spoken at home (Table [3](#T3){ref-type="table"}). After adjusting for age, sex, remoteness and total number of health problems, concession health care card holders made on average 2.6 more visits per year to a GP than non-card holders (p \< .001, Table [4](#T4){ref-type="table"}). After adjusting for age, sex, concession health care card status and total number of health problems, patients from remote/very remote locations made on average 2.3 fewer visits per year to a GP than patients from locations where services were highly accessible (p \< .001). ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Mean number of recorded problems (crude rate) and differences in mean annual visits to a GP (crude rates and adjusted for age and sex) by patient demographics ::: **Annual visits** -------------------------------------------------- ----- ------------------- ------ --------- Highly accessible (8,702) 2.9 9.1 9.0 Moderately accessible (1,599) 2.7 7.7 7.9 0.05 Remote/very remote (145) 2.5 5.8 6.5 \<.0001 Concession health care card holder (4,448) 3.5 11.4 10.7 \<.0001 Non health care card holder (6,307) 2.4 6.8 7.5 Speaks language other than English at home (890) 2.8 9.0 9.1 0.369 Speaks English at home (9,865) 2.9 8.7 8.8 Male (4,285) 2.7 8.5 N/A N/A Female (6,404) 2.9 8.9 N/A ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Multiple regression model of sociodemographic predictors of visit rates and number of recorded health problems ::: **Partial coefficient^(a)^(95%CI)** **p-value** ---------------------------------------- ------------------------------------- ------------- Patient sex (ref: male) 0.0 (-0.6;0.5) 0.88 Patient age (ref: \<1 year) 1--4 years 2.2 (1.5;2.9) \<.001 5--14 -0.2 (-0.8;0.4) 0.46 15--24 0.5 (-0.1;1.2) 0.08 25--44 2.2 (1.5;3.0) \<.001 45--64 1.9 (1.1;2.8) \<.001 65--74 2.0 (0.9;3.1) \<.001 75+ 2.7 (1.4;4.1) \<.001 Accessibility (ref: Highly accessible) Moderately accessible -1.0 (-2.1;0.2) 0.12 Remote/very remote -2.3 (-3.1;-1.5) \<.001 Concession health care card holder 2.6 (1.8;3.3) \<.001 Each extra health problem 1.6 (1.3;1.8) \<.001 ^(a)^Interpreted as the change in mean number of annual visits after adjusting for all other variables in the model. ::: Morbidity and visit rates ------------------------- The mean annual visit rates for patients with common chronic disorders were well above average (Table [5](#T5){ref-type="table"}). After adjusting for demographic predictors and total number of health problems, patients with anxiety made on average 2.7 more visits to a GP per year (p = 0.003), those with depression 2.2 more visits (p \< .0001) and those with back problems 2.4 more visits (p = 0.009). Ischaemic heart disease, diabetes, asthma, oesophageal disease, osteoarthritis or hypertension did not affect patient visit rate beyond that expected from the patients\' demographics and overall number of health problems. ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### The effect of specific morbidity on annual visits to a GP (Crude rates, rates adjusted for age, sex, concession health care card, remoteness and number of other problems) ::: **Crude rates** **Effect on visit rate (adjusted)** ------------------------------- ------------------------ ------------------------------------- --------------------------------------- ------------- **Health problem^(a)^(n)** **Mean annual visits** **No. other problems** **Morbidity coefficient^(b)^(95%CI)** **p-value** Anxiety (381) 14.1 3.4 2.7 (0.9;4.5) 0.003^†^ Depression (1,100) 13.1 3.2 2.2 (1.4;3.1) \<.0001^†^ Back complaint (578) 12.8 2.9 2.4 (0.6;4.2) 0.009^†^ Diabetes (706) 13.7 3.8 0.9 (-0.1;1.9) 0.08 Ischaemic heart disease (585) 14.3 4.3 0.6 (-0.7;1.9) 0.38 Asthma (876) 10.5 2.8 0.3 (-0.4;1.0) 0.40 Oesophageal disease (636) 12.9 4.1 -0.2 (-1.2;0.8) 0.65 Osteoarthritis (859) 13.0 4.1 -0.4 (-1.3;0.5) 0.39 Hypertension (2,094) 11.8 3.5 -0.4 (-1.2;0.4) 0.30 ^(a)^Each line represents a separate regression model (adjusted for age, sex, concession health care card holder, remoteness and number of problems). ^(b)^Interpreted as the mean difference in number of annual visits between patients with the specific problem versus those without after adjusting for age, sex, concession health care card holder, remoteness and overall number of health problems. † significant p \< .01 ::: Discussion ========== As would be expected the number of health problems recorded for the patient was a major predictor of frequency of attendance at general practice. However this study demonstrated that specific chronic disorders had differential effects on visit frequency. Anxiety, back complaints, and depression had the greatest effects on increasing patient annual visits to a GP. The relationship between psychological health problems and increased medical visits has been demonstrated across many settings \[[@B1]-[@B3],[@B5]\]. The current study further demonstrates the generalisability of this relationship. However the interpretation of this finding in the Australian context is less clear. Patients with depression may be visiting GPs more frequently because GPs fail to recognise and treat the root causes of the patient\'s poor mental health \[[@B5]\]. Alternatively, above average visits for depression and anxiety may be appropriate if the practitioner has initiated the return visit for ongoing management. GPs in Australia have been targeted to improve the detection and management of depression among general practice patients \[[@B25]\], so higher visit rates for some patients with depression could reflect a more active intervention strategy by their GPs. Other research has reported a link between unexplained back pain and frequent attendance at medical services \[[@B26]\]. In the current study back complaints included problems described as symptomatic back pain as well as diagnosed disc and nerve problems \[[@B18],[@B20]\]. It is unclear the degree to which the frequent visits of patients with back complaints may be due to somatisation \[[@B27]\] and how much is explained by the need to manage acute/intractable pain from organic causes. Other research indicates that patients with diabetes make more general practice visits than average \[[@B28]\] and structured care programs might be expected to result in above average visit frequency for patients with diabetes \[[@B12],[@B29]\]. The current analysis controlled for other health and demographic factors and found that the higher visit rates of patients with diabetes was more related to the high number of health problems reported by the patient rather than to extra visits related to diabetes per se. Regression models revealed similar explanations for the visit rates of patients with asthma, ischaemic heart disease, osteoarthritis, oesophageal disease and hypertension: visit rates were mostly explained by the overall level of health problems experienced by the patient rather than to any increased demand for health care related to the specific morbidity. As expected, patients from remote areas reported visiting general practice less frequently relative to their health needs, confirming that geographic access to services remains a significant barrier to meeting the health care needs of patients from remote locations \[[@B13],[@B14]\]. Concession health care card holders have more chronic and psychological health problems managed in general practice relative to the rest of the Australian general practice population \[[@B22]\]. In this study concession health care card holders reported a higher than average number of visits even after accounting for their number of health problems. This relatively high visit rate coupled with high levels of ill-health places concession health care card holders among the most frequent attenders at general practice. Shorter consultation length is associated with lower socioeconomic status in Australia \[[@B30]\]. Perhaps concession health care card holders are partly compensating for less time spent with the GP at each visit by visiting more often. This study had the advantage of randomly sampling clusters of patients across a wide range of practices in all states and regions of Australia. A further advantage was having a qualified medical practitioner to record patient morbidity and the use of the medical record to help validate patient recall of GP visits. The methodology of this study was limited by using patient self-report of visit frequency and by sampling patients at the point of consultation. The average number of GP visits among Australians who visit a general practitioner at least once in a 12 month period is known to be six visits per year, increasing with increasing age (Australian Health Insurance Commission, unpublished data), while the self-reported visit rate in the current study was 8.8 visits per year. Because the patients were sampled at the consultation, older patients who attend general practice more frequently had a greater chance of being sampled and this was reflected in the older age distribution of the patient sample. Even allowing for this selection bias, the mean annual self-reported visits was somewhat higher than expected given the sample\'s age structure, indicating some overestimation by patients in their recall of the number of GP visits. One Australian study that validated patient self-report against actual Medicare benefit claims found that patients who overestimate recent use of medical services are \"telescoping\" real but less recent events into the nominated time period and are in fact higher users of health services over the long term \[[@B31]\]. In the current study the number of reported visits increased with age and number of morbidities as expected. Therefore even if somewhat overestimated, self-reported visit frequency had sufficient face validity to allow a comparative analysis of the effect of morbidity and demographic factors on the relative frequency of visits to a GP. Conclusion ========== This study has demonstrated that number of visits to a GP by patients in Australia is largely explained by the number of health problems the patient is experiencing. However psychological health problems and sociodemographic factors have differential effects on visit rates. GPs need to be alert to the particular psychological and social needs of patients who are attending frequently with multiple health problems. Equity of access, in terms of equal utilisation of services, continues to be an issue for patients living in remote locations. Competing interests =================== None declared. Authors contributions ===================== SK designed and performed the statistical analysis, researched and wrote the main draft. HB directed the study, participated in the design and assisted in writing and reviewing the final article. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2296/5/17/prepub> Acknowledgements ================ The authors wish to thank the GPs who participated in the study and all BEACH staff who made this work possible. The organisations who contributed financially to the conduct of the BEACH study in 2001--2002 were: the Australian Department of Health and Ageing; AstraZeneca Pty Ltd (Australia); Aventis Pharma Pty Ltd; Roche Products Pty Ltd, Janssen-Cilag Pty Ltd; Merck Sharp and Dohme (Australia) Pty Ltd. The General Practice Statistics and Classification Unit is a collaborating unit of the Australian Institute of Health and Welfare.

PubMed Central

2024-06-05T03:55:47.450740

2004-8-20

PMC516033

Background ========== The incidence of esophageal adenocarcinoma has increased more rapidly than any other form of cancer since the 1970s and now represents the majority of esophageal neoplasms in the West \[[@B1]\]. Barrett\'s esophagus is the replacement of native squamous mucosa by specialized intestinal metaplasia and is known to be the major risk factor for the development of adenocarcinoma via the metaplasia-dysplasia-neoplasia sequence \[[@B2]\]. Other risk factors for the development of esophageal adenocarcinoma include a lengthy and severe history of gastroesophageal reflux disease (GERD), increased body mass index, male gender and Caucasian race \[[@B3]-[@B5]\]. Recent studies, however, have detected Barrett\'s esophagus in nearly equal numbers of older white men whether or not they reported GERD symptoms \[[@B6],[@B7]\]. Recommendations, therefore, regarding screening and surveillance of patients at risk for esophageal adenocarcinoma are controversial \[[@B2],[@B8]\]. Detection of dysplastic Barrett\'s esophagus or mucosal adenocarcinoma is important because it allows the opportunity to intervene prior to the development of invasive neoplasia. Unfortunately, no medical or surgical GERD treatment has been consistently and convincingly demonstrated to prevent the development of adenocarcinoma \[[@B5]\]. Traditionally, high grade Barrett\'s dysplasia and mucosal adenocarcinoma are treated by surgical resection of the esophagus \[[@B9]\]. As endoscopic methods of screening and surveillance have become more widely applied, curative esophagectomy has produced increasing numbers of long-term survivors \[[@B10],[@B11]\]. Coincident with this trend has been the appearance of several reports describing the development of metachronous esophageal adenocarcinoma after undergoing curative esophagectomy for the primary tumor \[[@B12],[@B13]\]. These reports raise questions regarding the role of endoscopic surveillance in these patients. The aim of the current study was to evaluate patients who had undergone complete, presumably curative, esophageal resection for early Barrett\'s adenocarcinoma or high grade dysplasia in order to determine how frequently, and over what time period, they developed recurrent Barrett\'s esophagus or adenocarcinoma. Methods ======= After approval by the Mayo Foundation\'s Institutional Review Board for Research, the electronic medical records of Mayo Clinic patients in Jacksonville, Florida, were searched to find all patients who had undergone esophagectomy for cure at the Mayo Clinic surgical facility, St. Luke\'s Hospital, Jacksonville, Florida, since 1995. This time period was chosen to coincide with the routine availability and clinical use of pre-operative staging with endosonography in our institution. The records of these patients were reviewed for pre-operative and post-operative staging results including computed tomography and endosonography studies. In addition, endoscopic, surgical studies and histopathological studies were studied. Specifically, the surgical specimens were reviewed to ensure that the esophagectomy specimen, including lymph node sampling, was adequate and the proximal margin was completely free of Barrett\'s metaplasia, dysplasia or carcinoma. The patients, in whom we have at least one follow-up endoscopy study, with biopsies obtained for histologic confirmation of mucosal disease, are the subjects of this report. Esophageal disease was staged according to the Tumor-Lymph node-Metastasis (TNM) criteria \[[@B14]\]. Results ======= Since 1995, 45 patients have undergone esophagectomy for Barrett\'s dysplasia or localized adenocarcinoma with curative intent in our institution. At operation, none of these patients were found to have extension of malignant disease to paraesophageal lymph nodes and all esophageal glandular mucosa was resected with only normal squamous mucosa remaining at the proximal surgical margin. Subsequently, 36 of these patients (80%) have undergone endoscopy after surgery including 8/45 patients (18%) who were found to have recurrent Barrett\'s glandular mucosa after curative resection and are described in the table. Open transthoracic esophagectomy (Ivor Lewis procedure) with pyloroplasty was performed in most patients (39/45) including the patients diagnosed with recurrent Barrett\'s disease. Five different surgeons performed these operations. Most patients had evidence of gastric stasis (retained food) at endoscopy (30/36; 83%). Anastomotic dilation was performed at endoscopy in 16/36 patients (range of dilation procedures 1--10; median 3). It is possible that patients with anastomotic strictures may be at increased risk of recurrent Barrett\'s esophagus because of worse reflux although their swallowing symptoms may, alternatively, be related to other factors such as anastomotic ischemia or surgical sutures. Patients frequently used aspirin (42%) or COX-2 specific non-steroidal anti-inflammatory agents (25%). Twice-daily proton pump inhibitors were routinely prescribed for these patients although patient compliance is difficult to assess because of high drug costs and limited symptomatic improvement. While the small number of patients limits our analysis, these factors were found to occur in a proportional number of patients with Barrett\'s disease and no clear trends could be identified. Five of these patients have been diagnosed with Barrett\'s metaplasia or low-grade dysplasia have been followed for more than 12 months in surveillance endoscopy programs monitoring the stability of the glandular epithelium. Two of these 5 patients have been found to have short segment Barrett\'s metaplasia with lengths of 10 mm and 15 mm, after complete esophageal resection for Barrett\'s high-grade dysplasia in a 72-year-old man and Barrett\'s adenocarcinoma T~2~N~0~M~0~in a 78-year-old man. This metaplastic glandular epithelium was detected at a follow up of 90 months and 17 months, respectively. In the other 3 patients, short segment Barrett\'s low-grade dysplasia has been found in lengths between 10--25 mm after complete resection of the esophagus for Barrett\'s high-grade dysplasia (1 patient) and Barrett\'s T~3~N~0~M~0~carcinoma (2 patients). This recurrent Barrett\'s glandular dysplasia was detected at a follow up of 42--47 months. Erosive esophagitis was also noted in 4 of 5 patients indicating uncontrolled reflux disease. Subsequently all patients have been treated with high doses of proton pump inhibitors (such as esomeprazole 80 mg twice a day or 40 mg three or four times per day) in an attempt to maximally control reflux of acid and digestive juices from the stomach into the cervical esophagus. Three other patients developed recurrent Barrett\'s disease after curative resection of esophageal T~2~or T~3~N~0~M~0~adenocarcinoma. These patients varied in age from 58--80 years of age. These patients were found to have more severe erosive esophagitis suggesting worse acid reflux and mucosal injury compared to the non-carcinoma recurrent Barrett\'s patients. Recurrent Barrett\'s multi-focal high-grade dysplasia, over a 10 mm segment length was detected 88 months after esophagectomy in one patient and was successfully ablated with porfimer sodium photodynamic therapy using the methods described elsewhere \[[@B15]\]. High-grade dysplasia with features of invasive adenocarcinoma was noted in a 20 mm Barrett\'s segment diagnosed 18 months after esophageal resection in another patient. Endosonography detected focal mucosal expansion and the Barrett\'s mucosal adenocarcinoma T~1~N~0~M~0~was confirmed at repeat esophagectomy. Finally, a diminutive polypoid mass proximal to the surgical anastomosis was found in a 58-year-old woman who had 7 months previously undergone esophagectomy for Barrett\'s mucosal adenocarcinoma. Computed tomography with contrast enhancement noted esophageal wall thickening and suspicious lymphadenopathy. Repeat resection confirmed the tumor histologic grade of T~2~N~1~M~0~adenocarcinoma. Discussion ========== Over the past four decades, the incidence of esophageal adenocarcinoma has risen dramatically, particularly in older white men \[[@B16]\]. Previous studies have documented the association of esophageal cancer with the severity and duration of gastroesophageal reflux \[[@B3]\]. The most important risk factor for the development of esophageal adenocarcinoma, however, is the development of specialized intestinal metaplasia in the lower esophagus (Barrett\'s esophagus) \[[@B5]\]. It is recommended that patients with Barrett\'s mucosa undergo periodic surveillance endoscopy with systematic biopsies to detect the presence of cancerous or pre-cancerous changes (dysplasia or neoplasia) \[[@B2]\]. In these patients, especially, Barrett\'s adenocarcinoma or high-grade dysplasia is found at an early stage when it is increasingly possible to undergo esophagectomy with curative intent \[[@B17]\]. After undergoing esophageal resection, the native squamous mucosa of the cervical esophagus will be brought into contact with the acid-secreting mucosa of the gastric body. This reconstruction allows acid and duodenal juice to reflux from the gastric conduit to the remaining cervical esophagus. Reflux of gastric and duodenal content is an important factor in the pathogenesis of Barrett\'s metaplasia, dysplasia and esophageal adenocarcinoma \[[@B18]-[@B20]\]. It is well known that recurrent Barrett\'s glandular mucosa is frequently found in the cervical esophagus after esophageal resection \[[@B21]-[@B23]\]. Recently, Oberg et al performed endoscopy, esophageal manometry and ambulatory pH studies in 32 patients who had undergone esophagectomy for a variety of diagnoses including 16 patients (50%) with adenocarcinoma associated with Barrett\'s metaplasia \[[@B24]\]. These studies were performed between 3--10 years after the surgery and most of the patients (70%) were receiving at least once daily proton pump inhibitor therapy for chronic reflux symptoms. At endoscopy, Barrett\'s glandular mucosa was histologically documented in 15 patients (47%) ranging in segment lengths from 0.5--4.0 cm. Importantly, recurrent Barrett\'s glandular mucosa was significantly more likely to occur in patients with a preoperative diagnosis of Barrett\'s epithelium suggesting that the esophageal mucosa of these patients may be pathogenetically predisposed to develop metaplastic changes in response to gastroesophageal reflux. In the study of Oberg et al, despite the use of potent acid-suppressing medications, severe esophageal acid exposure was noted in most patients. Patients with recurrent Barrett\'s epithelium were found to have significantly more severe acid exposure that occurred predominantly in the supine position \[[@B25]\]. There was also a direct correlation between the length of the metaplastic segment and the severity of acid reflux. Similar findings were reported by da Rocha, et al., who studied 48 patients after esophagectomy where 4 of these patents (8%) were found to have pathological changes of Barrett\'s metaplasia in the cervical esophageal stump \[[@B26]\]. In both of these studies, the authors concluded that the finding of recurrent Barrett\'s esophagus was related to reflux esophagitis that resulted from the action of acid-peptic and biliary secretions. These studies, however, did not detect dysplasia or neoplasia and did not address the issue of metachronous cancers and the role of endoscopic surveillance for these patients. Murata and colleagues recently reported the diagnosis of metachronous squamous cell carcinomas in five of 253 patients (2%) who had undergone esophagectomy for thoracic esophageal squamous cell carcinoma more than two years previously. These superficial carcinomas (T~is~or T~1~) were detected at surveillance endoscopy and were treated with endoscopic laser ablation, mucosal resection or surgical resection. While squamous cell carcinomas are not related to gastroesophageal reflux, this paper also suggests that esophageal cancer patients (squamous or adenocarcinoma) are predisposed to the development of metachronous carcinomas in the remnant cervical esophagus. This is consistent with DeMeester\'s experimental model of Barrett\'s dysplasia and adenocarcinoma occurring after complete gastrectomy with esophago-jejunostomy and reflux of bile and digestive enzymes into the cervical esophagus \[[@B27]\]. Also, it is likely that these patients had other important risk factors, such as tobacco and alcohol use, that predispose to both squamous and adenocarcinomas \[[@B5]\]. Konishi et al reported finding an adenocarcinoma in Barrett\'s esophagus following a total resection of the gastric remnant in a 52-year-old man who had undergone distal gastrectomy for gastric cancer nearly twenty years previously \[[@B28]\]. The Barrett\'s mucosa associated with the adenocarcinoma contained high grade dysplasia supporting the acquired theory of pathogenesis for Barrett\'s esophagus that suggests that reflux esophagitis after gastrectomy may result in the dysplasia-carcinoma sequence. In addition, Streitz et al retrospectively reviewed long-term survivors after esophagectomy for adenocarcinoma \[[@B13]\]. With a follow-up as long as 14 years, they found 4 patients who subsequently were diagnosed with esophageal adenocarcinoma. However, the time period between the development of the first and second tumors was not specified making it not possible to determine if these were recurrent tumors or new, metachronous lesions. Finally, Riben et al have reported the development of a secondary Barrett\'s adenocarcinoma in a patient who had 19 years previously undergone esophagectomy for a stage IIb Barrett\'s adenocarcinoma \[[@B12]\]. These studies have demonstrated that the cervical esophagus is exposed to high amounts of acid and refluxate despite the use of proton inhibitor medications and often in the absence of severe reflux symptoms. Although our group of patients has been observed for only a median of 2 years after esophagectomy, our study confirms that the development of metaplastic columnar mucosa in the cervical esophagus is a common complication related to reflux associated injury to the squamous epithelium. Further, our findings suggest that this recurrent glandular mucosa is unstable and predisposed to the development of dysplasia and invasive carcinoma, as has already developed in most of patients. The early detection of this recurrent disease remains vitally important to preserve all possible treatment options including surveillance endoscopy follow-up, endoscopic ablation with porfimer sodium photodynamic therapy, and if necessary repeat esophagus resection surgery. Our specific recommendations include surveillance endoscopy every 6--12 months for patients who have undergone \"curative\" esophagectomy for Barrett\'s dysplasia or adenocarcinoma. In addition, we also routinely recommend indefinite use of proton pump inhibitors, regardless of symptom status, starting at twice daily dosing and increasing as necessary to control reflux symptoms and mucosal damage due to acid, bile and digestive enzymes. Whether these drug doses should be titrated based on ambulatory pH and impedance test results remains to be determined. We have generally been disappointed by prokinetic agents such as metoclopromide in improving reflux symptoms in these patients. For esophagectomy patients who develop recurrent Barrett\'s metaplasia we recommend the use of COX-2 inhibitors or aspirin chemoprevention to protect against the development of metachronous Barrett\'s carcinoma. \[[@B29],[@B30]\]. Competing interests =================== None declared. Author\'s contributions ======================= All authors participated in the study design and coordination as well as case collection and review of histopathologic and endoscopic results. All authors read and approved the final manuscript. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Recurrent Barrett\'s Disease after Esophagectomy with Curative Intent ::: **Pre-operative diagnosis** **Sex** **Age** **F/u diagnosis** **Time to F/U** **Tx** ----------------------------- --------- --------- ----------------------- ----------------- --------- Barrett\'s T3 N0 ACA F 58 Barrett\'s T2 N1 7 mos Surgery Barrett\'s T3 N0 ACA F 64 Barrett\'s LGD 42 mos PPI Barrett\'s T3 N0 ACA F 64 Barrett\'s LGD 42 mos PPI Barrett\'s HGD M 64 Barrett\'s LGD 47 mos PPI Barrett\'s HGD M 72 Barrett\'s Metaplasia 90 mos PPI Barrett\'s T3 N0 ACA M 69 Barrett\'s T1 N0 18 mos Surgery Barrett\'s T2 N0 ACA M 78 Barrett\'s metaplasia 17 mos PPI Barrett\'s T2 N0 ACA M 80 BE+HGD 88 mos PDT ACA = Adenocarcinoma HGD = High-grade dysplasia LGD = Low-grade dysplasia PDT = Photodynamic therapy PPI = Proton pump inhibitor medical therapy ::: Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-230X/4/18/prepub>

PubMed Central

2024-06-05T03:55:47.453197

2004-8-25

PMC516034

Background ========== Platinum-based combination chemotherapy is the standard first-line treatment for patients with advanced non-small-cell lung cancer (NSCLC). Although meta-analysis of clinical trials proved that combination chemotherapy for advanced NSCLC is superior to best supportive care \[[@B1]\] and new cytotoxic agents have been developed over the past decade, the 5-year survival rate for these patients remains \<1% \[[@B2]\]. The pace of progress is too slow and new therapeutic approaches are required. Based on our increasing understanding of the biology of NSCLC, targeted therapy is providing some promising new agents. Targeted therapy is usually aimed at a key protein implicated in tumour cell proliferation, survival, invasion or resistance to conventional treatments but spares the normal cells, thereby producing less toxicity than conventional therapies \[[@B3]\]. Many solid tumours express or highly express the epidermal growth factor receptor (EGFR). In lung cancer, deregulation of EGFR is seen mainly in NSCLC: EGFR is highly expressed in NSCLC at levels varying from 32--80% \[[@B4]-[@B6]\]. Abnormal signal transduction arising from the receptor is implicated in the growth and proliferation of these tumours. Thus, disruption of the EGFR signal transduction pathway is an ideal target for anticancer therapy. Gefitinib is an orally active small molecule EGFR tyrosine kinase inhibitor (EGFR-TKI). The tolerability and efficacy of gefitinib have been explored extensively in the West and Japan. In a Phase I trial, Nakagawa et al compared the safety profile, pharmacokinetic parameters and antitumour activity of gefitinib in patients with solid malignant tumours in Japan, the USA and Europe, and no significant difference was found for ethnicity \[[@B7]\]. However, severe interstitial pneumonia related to gefitinib therapy was more common in the Japanese population \[[@B8],[@B9]\]. In IDEAL (\'Iressa\' Dose Evaluation in Advanced Lung cancer) 1, which included 102 Japanese patients with advanced NSCLC, the response rate of Japanese patients taking gefitinib was significantly higher than that of non-Japanese patients but bias of baseline factors between strata, not ethnicity, was thought to account for the difference \[[@B10]\]. But Owing to the paucity of current data on gefitinib between defferent races, the possibility of differences in the toxicity and efficacy of gefitinib for ethnicity cannot be excluded. At present, there are few data regarding the tolerability and efficacy of gefitinib in Chinese patients. Patients with advanced NSCLC who had no alternative therapeutic options have been able to receive gefitinib treatment in a worldwide compassionate-use programme. Here we report the outcome of treatment with oral gefitinib 250 mg/day in Chinese patients with advanced NSCLC participating in the compassionate-use programme at Peking Union Medical College Hospital. Methods ======= Patient population ------------------ Patients aged ≥18 years with histologically or cytologically confirmed advanced or metastatic NSCLC were eligible for enrolment into the compassionate-use programme after providing written, informed consent. They were required to have failed prior chemotherapy, and had no other treatment options available. Patients who had received chemotherapy were included if the treatment ended \>28 days prior to the study. Other eligibility criteria included: adequate bone marrow function (white blood cell count \>4 × 10^9^/L, absolute neutrophil count \>1.5 × 10^9^cells/L, platelet count \>100 × 10^9^/L); liver function (total bilirubin \<34 μmol/L, aspartate aminotransferase \[AST\] \<40 IU/L, alkanine aminotransferase \[ALT\] \<41 IU/L); renal function (serum creatinine \<150 μmol/L, blood urea nitrogen \<10.7 mmol/L); PaO~2~\>60 mmHg. Exclusion criteria included a serious pre-existing medical condition (eg uncontrolled infection, interstitial pneumonia or pulmonary fibrosis, severe chronic diarrhoea), pregnancy and lactation. Drug administration ------------------- One oral gefitinib tablet (250 mg) was taken at about the same time each day. Gefitinib was administered every day without interruption unless disease progression or unacceptable toxicity occurred. Evaluation before and during treatment -------------------------------------- In this study, 28 days of treatment was defined as one treatment cycle. Baseline evaluation included medical history and physical examination, electrocardiogram, chest X-ray, thorax computed-tomography scan and ultrasonography of the upper abdomen. Laboratory investigations included complete blood counts, urinalysis, and renal and liver function tests. Performance status (PS) was evaluated according to Eastern Cooperative Oncology Group (ECOG) criteria. Brain magnetic resonance imaging and radionuclide bone scans were only performed if metastatic disease was suspected according to the clinical manifestations of each patient. Regarding interstitial lung disease, we paid attention to clinical respiratory symptoms (e.g. dyspnea, cough) and radiographic findings of patients, and monitored PaO~2~during therapy. Patients were evaluated after the first and third cycles of therapy, then every three cycles. Tumour response was evaluated according to Response Evaluation Criteria In Solid Tumors \[[@B11]\]. In line with the criteria, all patients included in the study were assessed for response to treatment. Each patient was assigned to one of the following categories: complete response (CR); partial response (PR); stable disease (SD); progressive disease (PD); early death from lung cancer; early death from toxicity; early death because of other disease; unknown (not assessable or insufficient data). All AEs were recorded and graded according to National Cancer Institute Common Toxicity Criteria version 2.0. Assessment of QoL ----------------- QoL was assessed using the Chinese version of the European Organization for Research and Treatment of Cancer (EORTC) core questionnaire, the Quality of Life Questionnaire (QLQ)-C30 (version 3.0) and the supplemental lung-cancer-specific module QLQ-LC13 \[[@B12]\]. Following the scoring procedure recommended by the EORTC, scores were converted into linear transformation ranging from 0--100. For the functional and global health status/QoL, higher scales represent better functioning. For symptoms, a higher score represents worse symptoms \[[@B13]\]. QoL and disease-related symptoms were assessed before the start of the therapy and then at the end of each cycle in the first 3 months of treatment. Statistical analysis -------------------- Logistical regression test models were used to identify baseline factors (gender, PS, histology, TNM stage and prior chemotherapy) that might independently predict tumour response. Median progression-free survival (PFS) and overall survival (OS) were calculated using the Kaplan-Meier method and a log-rank test was used to detect differences OS between strata. Changes in symptoms and QoL were assessed in two different ways. Firstly, the mean scores of QoL and disease-related symptoms at baseline were compared with those at the end of the second cycle of treatment using a paired-sample t test. Secondly, the response rates of symptoms and QoL were calculated. For the assessment of symptoms and QoL, the classification system suggested by Stephens et al was used \[[@B14]\]. The results were presented as response (ie improvement, control or prevention), no response and nonevaluable cases in the form of percentages. The classification system \[[@B15],[@B16]\] has been used successfully to evaluate QoL and symptoms in patients with lung cancer. Response rates for QoL and disease-related symptoms were compared between patients with and without objective tumour response using Pearson\'s χ^2^test or Fisher\'s exact test. Results ======= Patient characteristics ----------------------- Thirty-one eligible patients were enrolled into the EAP at Peking Union Medical College Hospital between October 2002 and October 2003. Patient characteristics are listed in Table [1](#T1){ref-type="table"}. The patient series included 18 (58%) men and 13 (42%) women between 28 and 85 years of age (median age 64). All patients had received at least one platinum-based regimen and most had received more than two regimens (different combinations including platinum, taxane, docetaxel and gemcitabine). Ten (32.3%) patients had squamous-cell carcinoma (SCC) and 20 (64.5%) had adenocarcinoma. Most patients (61.3%) had an ECOG PS of 0--1, 25.8% had a PS of 2 and 12.9% had a PS of 3. TNM stages were as follows: stage IIIa, 1 patient; stage IIIb, 4 patients; stage IV, 26 patients. Toxicity -------- All patients were assessed for toxicity. Twenty-three (74%) patients had at least one AE. Treatment-related toxicities are listed in Table [2](#T2){ref-type="table"}. Almost all AEs (but for one grade 3 acneform rash) were mild (grade 1 or 2). The most frequently reported AEs during treatment were acneform rash (67.7%) and diarrhoea (35.5%). Four (12.9%) patients had grade 1 or 2 stomatitis. Other AEs included nausea (6.5%), vomiting (3.2%), increased ALT (3.2%) and increased AST (3.2%). The majority of AEs were transient and reversible. Only one patient withdrew from treatment due to an AE (on the 50th day of treatment with grade 3 skin rash and exacerbation of dysphagia). This patient had SCC, a PS of 3 and mediastinal lymph node metastases before treatment. The patient suffered moderate haemoptysis prior to gefitinib therapy, which reduced to a mild severity during therapy. Fifteen days after withdrawal from treatment, the patient died of massive haemoptysis. Patient response ---------------- The assessment of patient response is listed in Table [3](#T3){ref-type="table"}. Of the 31 patients, 1 (3.2%) achieved CR and 10 (32.3%) achieved PR with an overall objective response rate of 35.5% (95% confidence interval \[CI\]: 18.6--52.3). SD was documented in 7 (22.6%) patients and the overall DCR (CR+PR+SD) was 58.1% (95% CI: 49.2--67.0). The response rate of adenocarcinoma was significantly higher than that of squamose carcinoma \[50% (10/20) vs. 10% (1/10)\]. Multivariate logistic analysis also showed that the odds of objective response were 10 times higher (odds ratio 10; 95% CI: 1.0 to 93.4; p = 0.028) for patients with adenocarcinoma than for patients with other tumour histologies. Of the 31 patients who received treatment, 2 died due to PD before efficacy assessment. Both these patients had SCC and a PS of 3, and no improvement in either patient had been observed during gefitinib treatment. Median PFS was 5.5 months (95% CI, 1.6 to 9.4); median OS was 11.5 months (95% CI, 5.6 to 17.3). The tumor response was associated with improved OS. Median OS of patients with response was significantly higher than that of those with no response (log-rank test p = 0.0058) (17 *vs.*4.4 months). QoL and symptom improvement --------------------------- At baseline, 25 (81% \[25/31\]) patients returned QoL questionnaires. All 25 patients returned the questionnaires at the first cycle of therapy, 23 (93% \[23/25\] of those patients returning questionnaires and remaining alive) at the second cycle of therapy, and 18 (74% \[18/24\] of those patinets returning questionnaires and remaining alive) at the third cycle. The most frequently reported general symptoms were fatigue (100%) and appetite loss (68%), while dyspnoea (100%) and coughing (84%) were the most frequently reported respiratory symptoms. The changes in the mean scores for symptom scales, functioning and global QoL scales are presented in Table [4](#T4){ref-type="table"}. A statistically significant increase in mean score was observed for physical functioning, role functioning, emotional functioning, social functioning and global QoL after 8 weeks of treatment. There was also a trend towards a higher score (p = 0.08) for cognitive functioning. Mean scores for two general symptoms (fatigue and appetite loss) and disease-related symptoms (dyspnoea, coughing, pain in chest, pain in arm or shoulder and pain in other parts) decreased statistically. Patient responses for functioning, global QoL and main symptoms are shown in figure [1](#F1){ref-type="fig"}. A \>50% response rate was observed in all five functioning criteria and global QoL. With regard to QoL, the highest response rate was observed for emotional functioning (88%), followed by cognitive (72%), physical (68%), social (64%) and role functioning (60%), and global QoL (56%). Haemoptysis had the highest response rate (84%), the rate of improvement (63%) being just lower than that of appetite loss among the five symptoms. The response rates for dyspnoea and coughing were 56% and 68%, respectively. For fatigue and appetite, the response rates were 44% and 80%, respectively. Differences in response rates for symptoms and QoL between patients with or without objective tumour response are shown in Table [5](#T5){ref-type="table"}. Response rates of physical functioning, role functioning, social functioning and global QoL were significantly higher in patients with objective tumour response than in those without. Response rates for emotional functioning and cognitive functioning were also higher in patients with tumour response, though the difference was not statistically significant. Differences were also observed in response rates for symptoms. For dyspnoea, coughing, appetite loss and fatigue, the response rates in the objective responders were higher than those in nonresponders. No significant association was found between haemoptysis and objective tumour response. Discussion ========== A daily oral 250 mg dose of gefitinib was well tolerated in this series of Chinese patients with advanced NSCLC, who received gefitinib as part of a compassionate-use programme. AEs were mild and reversible and different to those associated with conventional chemotherapy, such as neutropenia, thrombocytopenia and neuropathy. No haematological or neural toxicity was observed in the study. As with prior clinical trials evaluating the toxicity of gefitinib, the most frequently reported AEs were acneform rash and diarrhoea (grade 1 or 2). Four patients developed grade 1 or 2 stomatitis during treatment, which has previously been associated with gefitinib treatment in 7.8% of patients in IDEAL 1 but was not reported in IDEAL 2 \[[@B17]\]. Severe acute interstitial pneumonia is the most serious AE that has been linked with gefitinib therapy. Inoue et al recently reported that 4 of 18 patients receiving gefitinib in their clinic developed severe acute interstitial pneumonia \[[@B8]\]. The authors stated that 291 of 17,500 patients (1.7%) treated with gefitinib in Japan had developed suspected interstitial pneumonia or acute lung injury. In contrast to the higher incidence of interstitial pneumonia, this severe AE was lower in the rest of the world \[[@B9]\]. The worldwide frequency of interstitial lung disease to date in \~92,750 patients who have received gefitinib is \<1.0% \[[@B18]\]. The ethnicity between Chinese and Japanese people is probably similar, but we saw no evidence of interstitial pneumonia in the series of patients. Although no patients developed acute interstitial pneumonia, the possibility of drug-related interstitial lung disease could not be excluded because of the small number of patients in our series. The response rate (35.5%) was higher than that of recently presented Phase II trials (18.4% and 11.8% in patients receiving gefitinib 250 mg/day in IDEAL 1 and 2, respectively). The DCRs were 54.4% and 42% for the 250 mg/day dose in IDEAL 1 and 2, respectively, in contrast to 58.1% in our study. Unlike IDEAL 2, which was performed in the USA, IDEAL 1 was a global trial that recruited patients from 43 centres across Europe, Australia, South Africa and Japan and included a total of 102 Japanese patients with advanced NSCLC. The response rate for Japanese patients was higher than that of non-Japanese patients (27.5% versus 10.45%; odds ratio 3.72; p = 0.0023). However, the difference in response rate for ethnicity was not verified using multivariate logistical regression analysis. The difference in response rate between Japanese and non-Japanese patients was attributed to bias of baseline predictive factors of patients (gender, PS and histology) \[[@B19],[@B10]\]. QoL is an important endpoint for assessment of gefitinib treatment and has been included in some phase I and phase II trials. In these trials, the Functional Assessment of Cancer Therapy -- Lung questionnaire was used to assess QoL and symptoms. Its validation and sensitivity were verified. Significant improvements in symptoms and QoL were observed in IDEAL 1 and 2. In IDEAL 1, symptom and QoL improvement with a dose of 250 mg/day were 40.3% and 23.9%, respectively, and in IDEAL 2 they were 43.1% and 34.3%, respectively. In both trials, improvements in symptoms and QoL correlated well with tumour response \[[@B20]\]. In our patient series, the EORTC QLQ-C30 and its supplement module QLQ-LC13 were used for assessment of QoL and symptoms. Together, they are thought to be one of the best developed instruments for lung cancer assessment \[[@B21]\]. The questionnaires have been translated into \>20 languages and validation of the standard Chinese version has been verified \[[@B22]\]. In our series of patients, mean scores of four functioning scales (physical, role, emotional and social functioning) and global QoL increased significantly after 8 weeks of treatment and were associated with high response from 56--88%. Mean score of cognitive functioning also increased, from 66 at baseline to 76 at the end of second cycle with response of 72% (not statistically significant; p = 0.08). General and disease-related symptom scores increased from 44% to 84%. Importantly, improvement of symptoms and QoL correlated with objective tumour response. Patients with objective tumour response had higher rates of symptom and QoL response. The results indicate that placebo effect is unlikely to explain completely the improvement in symptoms and QoL with gefitinib, though there is no placebo control in our patient series or the EAP. Emotional functioning response was observed in all patients with objective tumour response, but this was not statistically significant when compared with nonresponders (100% versus 79%; p = 0.23). As emotional functioning contained more subjective questions, such as \"Did you feel tense?\", \"Did you worry?\", \"Did you feel irritable?\" and \"Did you feel depressed?\", the limited difference of emotional functioning response between responders and nonresponders with gefitinib cannot completely exclude the possibility of placebo effect. Thus, a placebo control group should be included in future studies of QoL and symptom improvement with gefitinib. Conclusions =========== In conclusion, the data from this patient series show that an oral dose of gefitinib 250 mg/day is well tolerated and has significant antitumour activity in Chinese patients with advanced NSCLC who had failed previous chemotherapy and for whom no other treatment options were available. Although clinical trials are needed to confirm these positive findings, the data suggest that treatment with gefitinib may be beneficial for some Chinese patients who do not respond to chemotherapy and have poor prognosis. \'Iressa\' is a trademark of the AstraZeneca group of companies Competing interests =================== None declared. Authors\' contributions ======================= XLM and LYL designed the experiments and wrote the manuscript. XTZ, SLW and MZW participated in patient\'s follow-up and tumour response evaluation. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2407/4/51/prepub> Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Patient response for QoL and main symptoms QoL, quality of life; PF, physical functioning; RF, role functioning; EF, emotional functioning; CF, cognitive functioning; SF, social functioning :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Characteristics of patient series (n = 31) ::: --------------------------- ------------- Median age, years (range) 64 (28--85) Gender, n (%)  male 18 (58)  female 13 (42) ECOG PS, n (%)  0 3 (9.7)  1 16 (51.6)  2 8 (25.8)  3 4 (12.9) Histology, n (%)  squamous 10 (32.3)  adenocarcinoma 20 (64.5)  unspecified 1 (3.2) TNM staging, n (%)  IIIa 1 (3.2)  IIIb 4 (12.9)  IV 26 (83.9) Prior chemotherapy, n (%)  1 regimen 5 (16.1)  2 regimens 10 (32.3)  ≥3 regimens 16 (51.6) --------------------------- ------------- ECOG, Eastern Cooperative Oncology Group; PS, performance status ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Treatment-related toxicity (n = 31) ::: **Adverse events** **Grade** **Total** -------------------- ----------- ----------- ---- ----- ---- ----- ---- ------ Acneform rash 17 54.8 3 9.7 1 3.2 21 67.7 Diarrhoea 10 32.3 1 3.2 \- \- 11 35.5 Nausea 1 3.2 1 3.2 \- \- 2 6.5 Vomiting \- \- 1 3.2 \- \- 1 3.2 Stomatitis 4 12.9 \- \- \- \- 4 12.9 Increased ALT 1 3.2 \- \- \- \- 1 3.2 Increased AST 1 3.2 \- \- \- \- 1 3.2 ALT, alkaline aminotransferase; AST, aspartate aminotransferase ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Tumour response (n = 31) ::: **Patients** ---------------------------------- ------------------- Complete response, n (%) 1 (3.2) Partial response, n (%) 10 (32.3) Stable disease, n (%) 7 (22.6) Progressive disease, n (%) 11 (35.5) Early death, n (%) 2 (6.5) Response rate, % (95% CI) 35.5 (18.6--52.3) Disease control rate, % (95% CI) 58.1 (49.2--67.0) CI, confidence interval ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Changes in mean score for QoL and symptoms ::: **QoL items** **Baseline mean (standard deviation)** **2nd cycle mean (standard deviation)** **p value** -------------------------- ---------------------------------------- ----------------------------------------- ------------- Global QoL 36.0 (24.7) 55.4 (16.2) 0.01 Physical functioning 47.47 (24.74) 65.22 (20.59) 0.01 Role functioning 42.67 (35.05) 56.52 (25.99) 0.03 Emotional functioning 67.67 (28.39) 84.78 (19.57) \<0.01 Cognitive functioning 66.00 (28.66) 76.09 (18.69) 0.08 Social functioning 42.67 (32.30) 60.87 (26.88) 0.01 Symptoms items  Fatigue 64.89 (25.99) 45.89 (22.30) \<0.01  Appetite loss 41.33 (36.36) 20.29 (24.08) 0.01  Dyspnoea 60.44 (28.16) 39.61 (23.88) \<0.01  Coughing 58.67 (36.36) 28.99 (30.66) \<0.01  Haemoptysis 12.00 (18.95) 7.25 (17.28) 0.21  Pain in chest 36.00 (33.22) 17.39 (19.77) 0.04  Pain in arm or shoulder 24.00 (28.09) 13.04 (19.43) 0.03  Pain in other parts 25.33 (27.69) 14.49 (22.08) 0.02 QoL, quality of life ::: ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### QoL and symptom response rates among objective tumour responders and nonresponders ::: **Items** **Responders (n = 11)** **Nonresponders (n = 14)** **p value** ---------------- ------------------------- ---------------------------- ------------- ---- ------- QoL  Global QoL 10 91 4 29 0.004  PF 10 91 7 50 0.042  RF 10 91 5 36 0.012  EF 11 100 11 79 0.23  CF 10 91 8 57 0.09  SF 10 91 6 43 0.033 Symptoms  dyspnoea 9 82 5 36 0.021  coughing 10 91 7 50 0.042  haemoptysis 11 100 10 71 0.105  appetite loss 11 100 9 64 0.046  fatigue 8 73 3 21 0.01 QoL, quality of life; PF, physical functioning; RF, role functioning; EF, emotional functioning; CF, cognitive functioning; SF, social functioning :::

PubMed Central

2024-06-05T03:55:47.454626

2004-8-19

PMC516035

Background ========== Hyperthermic isolated limb perfusion is an established treatment for multiple, locoregional intransit metastases of malignant melanoma and soft tissue sarcomas of the extremities. It was introduced 1957 by Creech and Krementz \[[@B1]\]. The procedure allows a high locoregional concentration of cytostatics with few systemic side effects \[[@B2]\]. Isolated limb perfusion achieves local tumor control in a high percentage of cases. Complication rate of perfusion is acceptably low. A surviving model for isolated perfused rat hindlimbs was reported by Wu et al. \[[@B3]\] based on the perfusion model of Nagel et al.\[[@B4]\]. Angioneogenesis has been recognized for many years to play a central role in the growth of primary tumors and the formation of metastases in general \[[@B5],[@B6]\]. Melanomas express basic fibroblast growth factor (bFGF) and fibroblast growth factor receptor-1 (FGFR-1) in their dermal nevocytes and in the stroma. bFGF promotes angiogenesis *in vivo*\[[@B7]\] and *in vitro*\[[@B8]\]. Antisense targeting of bFGF/FGFR-1 in malignant melanomas blocks intratumoral angioneogenesis \[[@B9]\]. Hyperthermia inhibits angioneogenesis. Some of the antineoplastic effects of hyperthermia are caused by ischemia due to obstruction or destruction of the tumor vessels. Eikesdal et al. described by hyperthermic temperatures disrupture of 25--50 % of the vasculature in malignant tumors. The heated tumors (44°C) had a blood flow reduction of 40--60 % after 24 hours \[[@B10]\]. *In vitro*data indicate that capillary endothelials cells of malignant neoplasms are more thermosensitive than endothelial cells of normal tissues. The vascularization of tumors can be significantly damaged at temperatures which may alter but do not damage the vasculature of normal tissue \[[@B11]\]. The extent of the proliferation of vessels is also inversely related to temperature \[[@B12]\]. The critical temperature at which a direct vascular damage occurs is between 42.7 and 43.7 degrees C \[[@B13]\]. Hyperthermia increases the effects of angioneogenesis inhibitors on tumor growth and tumor angioneogenesis \[[@B14]\]. The aim of this experimental study was to determine the impact of hyperthermia on tumor response and angioneogenesis in an isolated limb perfusion model. Methods ======= Thirty inbred male nude athymic rats (Rowett rnu/rnu) weighting 200 to 280 g were used in this study. Animals were kept separately during the experiment with 12 hours of light per day. They were fed a standard laboratory diet and tap water *ad libidum*. Maintenance and care of all experimental animals were carried out according to the guidelines of the local responsible Animal Protection Commission and carried out in compliance with national guidelines (National Institute of Health for Use of Laboratory Animals). A solid sample (2 × 2 × 2 mm) of a human melanoma xenograft (SK-MEL-3), derived from a lymph node melanoma metastasis, was implanted into the hindlimbs of 30 nude rats. The largest width and the maximum tumor diameter perpendicular to the width were measured with a micrometer. Only tumors with a diameter greater than \> 12 mm in the largest diameter were included in the study. The details of the perfusion system have been described previously \[[@B4]\]. The equipment consisted of a miniature oxygenator, a heat exchanger and a roller pump. Venous blood of the limb was oxygenated in the oxygenator (99.15% O~2~; 0.85% CO~2~) and warmed in the heat exchanger. The warmed arterialized perfusate is driven by a roller pump with two synchronously running pump-heads on a single axis for the arterial and venous lines (Masterflex^®^). The outer diameters of the tubes corresponded with the diameter of the femoral vessels of nude rats weighting about 250 g (arterial 0.7 mm; venous 1.0 mm). Rat hindlimbs were perfused with ringer\'s solution and sodium heparin (25 IU/100 μl saline) at 37 or 41.5 degrees C for 30 minutes with a flow rate of 4 ml/min. External hyperthermia was applied by an infrared lamp positioned in a fixed distance to the tumor. For temperature measurement during limb perfusion, a nickel-chrom-nickel thermocouple of 0.6 mm in diameter (Standard Integrated Thermocouple Thermocoax, Phillips, Hamburg, Germany) was placed at the macroscopic tumor margin. The thermocouple was calibrated before use in a high-precision water bath. Baseline temperature was recorded for 5 minutes before treatment. Temperature was continuously measured during application. In group I (control) and group II (sham operated) temperature measurement was not performed. The animals were randomized into a control and four study groups of six animals each: group I: control (no therapy), group II: sham group (skin incision without ILP), group III: external hyperthermia with a tissue temperature of 41.5°C for 30 minutes without ILP, group IV: normothermic ILP (Ringer\'s solution as perfusate, tissue temperature 37°C for 30 minutes, group V: hyperthermic ILP (tissue temperature 41.5°C for 30 minutes). ILP was subsequently performed with the above-mentioned parameters. Antineoplastic agents were not applied. The rats were anaesthetized with Ketamin (Ketanest^®^, 80 mg/kg i.m.; Pearl-Davis, Berlin, Germany) and Xylazin (Rompun^®^, 10 mg/kg i.m.; Bayer, Leverkusen, Germany), and fixed in a supine position. For perfusion, the animals underwent a 3-cm incision of the groin. The left femoral artery and vein were isolated using an operation microscope. A vascular clip was placed across the artery and a silicon catheter (outer diameter 0.7 mm) was introduced via a transverse arteriotomy. The femoral vein was cannulated with a silicon catheter of 1 mm outer diameter (Fig. [1](#F1){ref-type="fig"}). Tourniquets were applied around the left hindlimb to ensure isolation. Rat hindlimbs were perfused at a flow rate of 4 ml/h, for 30 min. After treatment, the silicon catheters were removed, and the vein and the artery were sutured (8/0 Prolene), to restitute normal blood flow. After the isolated limb perfusion the wound was closed in two layers. Total operation time rated about 90 min. During the first 5 days after operation the rats were weighed daily, and once a week there-after. The animals were kept under standardized conditions and were killed by an overdose of anesthetic and cervical dislocation 6 weeks after treatment. The greatest tumor width and the maximum diameter perpendicular to the width were macroscopically measured with a micrometer. Lesion volume was calculated using the formula v = 4 × π × a × b 2/3, in which a and b are the radii of the measured axes. The tumor was then completely sectioned. Each tumor was divided into two parts. One part was fixed in 4 per cent formalin solution for 5 days, embedded in paraffin and stained with hematoxylin and eosin (H&E). The other part was fixed in 0.9 % saline solution, frozen and stored at -80°C until use. Vascular density and tumor morphology were evaluated by immunohistochemistry. Tumor tissue was tested for expression of basic fibroblast growth factor (bFGF) and CD34. Tissue specimens were fixed in 10 % formaldehyde and embedded in paraffin according to routine protocols. Sections 4 μm thick were deparaffinized and rehydrated using graded ethanols following routine protocols. For antigen retrieval all sections were microwave treated for 20 minutes in target retrieval solution (for bFGF; DAKO, Glostrup, Denmark) or 10 mM citrate buffer (for CD34) at 700 W. The sections were incubated overnight with primary mouse anti-bFGF (1:50; Becton Dickinson, Heidelberg, Germany) and anti-CD34 antibody (1:100; Immunotech, Hamburg, Germany). After rinsing in Tris buffer detection was carried out with a biotinylated secondary rabbit anti-mouse antibody diluted 1:50 (DAKO, Glostrup, Denmark) for 30 minutes at room temperature. The sections were washed again as above in Tris buffer and then incubated with streptavidin-biotin-alkaline-phosphatase (10 μl each solution A+B (Sigma, Deisenhofen, Germany) in 10 ml Tris). Immunohistochemical staining was visualized with Fast Red (2 mg Naphtol-As-Mx-phosphate, 10 μl 1 M Levamisol and 10 mg Fast-Red (Sigma) in 0.2 ml N, N-dimethylformamide, (Merck, Darmstadt, Germany) and 9.8 ml 0.1 M Tris-HCl buffer at pH8.6). The sections were rinsed in H~2~0 and counterstained with hemalaun. Negative controls without primary antibody were run for each sample. Tumor response was graded according to the system described by de Wilt et al. \[[@B15]\]. Progressive disease (PD) was defined as a tumor size \> 125 % of the size at the time of treatment, in tumor sizes of 100 ± 25 % no change (NC) was stated. Response was graded as partial (PR) if tumor sizes was between 10 and 75 % and complete (CR) for tumor size \< 10 % of its original diameter. In areas of intense neovascularized spots microvessels were counted in a 100× field. A microvessel was defined as a lumen surrounded by a rim of endothelial cells highlighted by immunostaining with anti-CD34 antibodys or anti-bFGF antibodys. Five separate intense neovascularized areas were assessed, and the mean was calculated as microvessel density of each tumor. The score was assessed by two independent observers. Data were analyzed using SPSS/PC+ statistical software. The mean and range of tumor and lesion volumes were calculated for each group. For comparison of tumor volume and microvessel density between the diffent groups a non-parametric test was used (Kruskal-Wallis). A p value \< 0.05 was considered to be significant. Results ======= During intervention, the required limb tissue temperature was reached within 8--12 minutes in group IV and 10--15 minutes in group III. Stable temperatures were then maintained for a further 30 minutes with a mean of 41.4°C ± 0.5. To reach limb tissue temperature the perfusion fluid maintained 43°C ± 0.4. Tumor and lesion Volumes ------------------------ Before treatment, the mean (s.e.m) volume of the treated tumors in groups I, II, III, IV and V was 105 mm^3^(4.3), 98 mm^3^(5.7), 108 mm^3^(3.4), 95 mm^3^(5.0) and 107 (4.1) mm^3^respectively, with no significant difference between the groups. Average tumor volume of the controls (I) increased to 1388 mm^3^(101) during six weeks. In the sham operated group (II) tumor volume was significantly larger than in group I (2350 mm^3^(198.6), *P*= 0.021). Tumor volume in group IV was significantly smaller (1009 mm^3^(122.5)) than in group I and lowest in group V 405 mm^3^(103.6) (*P*= 0.036 and *P*= 0.021, respectively.). There were no significant differences in size between group I and group III (1135 mm^3^(99)) after six weeks (*P*\> 0.05)(Fig. [2](#F2){ref-type="fig"}). Body weight ----------- After the perfusion, body weight decreased in all groups during the first 5 days with a maximum of 8 %. In the following weeks, body weight mounted up to 130 % of the preoperative value. In the last week of the experiment, the weight almost reached a plateau. No significant differences between the five experimental groups were found. Histology --------- In the H&E staining the untreated melanoma (group I) presented as a subcutanous unilocular nodule of moderate differentiation. The tumor showed mainly tubular structures with vacuole like lumen formation and was infiltrated with fibrous septa. The spontaneous rate of necrosis was estimated to be 20 %. Numerous mitoses were detected. Tumor morphology did not change with ageing or tumor size. Histology was similar in group II. In group V the tumors showed signs of irreversible cell damage after treatment. Tumor cells displayed clear shrinkage and partial loss of cell contact. Thromboses of the larger adjacent vessels were found on the tumor-skin border. Macrophage infiltration was present in all groups, but was more numerous pronounced in group V. In group III and IV there were less signs of irreversible cell damage after treatment than in group V. Immunhistochemical analysis of the histological sections for bFGF demonstrated that microvessel density of group I (Fig [3](#F3){ref-type="fig"}) ranged from 18 to 23 with a mean value of 20. Angioneogenesis was more pronounced in group II (Fig [4](#F4){ref-type="fig"}) than in group I (p = 0.003) and less pronounced in group IV and V (Fig [5](#F5){ref-type="fig"}) compared with group I (p = 0.023, p = 0.001). There were no significant differences in the expression of angiogenetic markers between group I and group III (p \>0.05). Immunhistochemical analysis of the histological sections for CD34 showed that microvessel density of group I ranged from 17 to 23 with a mean value of 19. Angioneogenesis was more pronounced in group II than in group I (p = 0,001) and less pronounced in group IV and V compared with group I (p = 0.023, p = 0.017). There were no significant differences in the expression of angiogenetic markers between group I and group III (p \> 0.05). Strong expression of bFGF and CD34 was found in the cytoplasm of intimal endothelial and medial smooth muscle cells. Expression of bFGF and CD34 was stronger at the invasion front than in the centre of tumors. Summarized data of the expression intensity of bFGF and CD34 in the tumors are given in table [1](#T1){ref-type="table"}. Tumor remission --------------- In group I and II tumor progression was observed in all animals macro- and microscopically. In group III and IV 5 animals showed tumor progression and one had a partial tumor response. In group V, 2 animals showed tumor progression, 2 had a complete remission. One had a partial tumor response and one had no change in tumor response (table [2](#T2){ref-type="table"}). Discussion ========== A phenomenon that has been described by numerous authors is the rapid increase of tumor growth following surgical manipulation \[[@B16],[@B17]\]. This was also observed in our sham-operated group of animals (group II) in which tumors grew more rapidly than in the control group (group I). This implies that resection is associated with an accelerated growth in residual tumors. It is assumed that the surgical manipulation induces liberation of growth factors that in addition to their effect on healing also have a stimulating effect on tumor proliferation. This hypothesis is supported by the results of our immunohistochemical investigations, which revealed an enhanced expression of fibroblast growth factor (bFGF) in group II. In an intraperitoneal tumor model, EGGERMONT et al. showed that the surgical intervention of laparotomy induces an increase of carcinomatosis in the entire peritoneal cavity. At the same time, the immunotherapeutic effect of interleukin 2 and natural killer cells is significantly reduced. It must therefore be speculated that surgical trauma induces a temporary immunosuppressive effect that influences tumor growth \[[@B16],[@B18],[@B19]\]. In our experiments, group V (hyperthermic perfusion) showed a significant slowing of tumor growth in comparison with the control group. This was also confirmed by the immunohistochemical investigations that showed a reduction in the expression of relevant markers of the vascular endothelium. Indeed, two of the six animals occurred a complete remission of their tumor. The observation that two animals in this group experienced tumor progression can probably be explained by the duration of the intervention, since factors affecting the kinetics of cell death during hyperthermia are not only the maximum temperature induced, but also the duration of the elevated temperature. In an overview report BHUYAN noted that, as a function of various tumor cell types, a temperature of 43°C must be applied between 30 and 150 minutes to irreversibly damage the tumor cells. At a temperature of 45°C, the application time required varied between 13 and 85 minutes \[[@B20]\]. In our experiments, the application time was only 30 minutes at a temperature of 41.5°C. Nevertheless, the relatively short application time also had an appreciable damaging effect on the metastatic vessels. Although the temperature in group III (external hyperthermia) was also 41.5°C applied for 30 minutes, no significant changes in the growth pattern of the tumor could be observed. Furthermore, the histological and immunohistochemical evaluations showed a mild hyperthermic damage at best. A comparison of treatment groups III and V revealed no differences in terms of duration of treatment and induced tissue temperatures. In contrast to group III, however, hyperthermia in group V has been mediated directly by the vascular system. In normal tissues subjected to warming, blood perfusion increases reactively via dilatation of the vessels together with an increase in vessel wall permeability. A decisive factor is that the tumor cells themselves, as well as the endothelial cells of tumor vessels are more thermosensitive than normal tissue \[[@B16],[@B21],[@B22]\]. Other authors have also shown that hyperthermia is capable of destroying vascular architecture, so that perfusion of the tumor is reduced. The sensitivity of various tumors, however, varies considerably \[[@B14]\]. ILP and other vasculotoxic effects like TNFalpha in combination with melphalan or doxorubicin increase the uptake of these drugs of three to six times \[[@B23]\]. Heat transfer between the vascular system and the tissue is determined in particular by the nature of the blood vessels. Important factors are the number, length, and diameter of the vessels, and also the velocity of blood flow \[[@B24]\]. CREZEE described the relationship between tumor perfusion and heating of the tumor tissue under systemic hyperthermia. He could show that the induction of heat in the tumor depends on the density of the blood vessels within the tumor itself. The more vessels in the tumor exist, the easier it is to heat the lesion via the vascular system. Heattransfer is particularly pronounced with large-calibre blood vessels or high flow velocity, and its rate is all the greater the higher the temperature gradient between the tumor tissue and the blood \[[@B25],[@B26]\]. On the other hand, tumor vasculature carries away the externally applied heat (group III) which is then no longer available to destroy the tumor cells. In this case, the temperature gradient decreases from the tumor to the vascular system. Furthermore, in the presence of hyperthermia, microcirculation in the tumor tissue is more affected than in normal tissue \[[@B27]\]. In addition, manipulation to the vascular system, and thus oxygenation of the tumor, also appears to play an important role. Although the body\'s own oxygenated blood is added to the perfusate, the dilution effect together with the clamping time, leads to a relative reduction in perfusion. It is known that hypoxia, in conjunction with a decrease in the pH and energy status of the cell, enhances thermosensitivity \[[@B28]-[@B30]\], and this increase correlates with the size of the tumor \[[@B31]\]. Our results have also confirmed that surgical manipulation of the vascular system and perfusion have an influence on the tumor. Overall, the effect of the intravascular application of hyperthermia appears to be mediated by a combination of hypoxia and heat. Conclusions =========== Our results suggest that even a surgical manipulation such as a skin incision promotes tumor growth, probably by induction of growth factors like bFGF. External hyperthermia of 41.5°C tissue temperature for 30 minutes has no impact on tumor growth and angioneogenesis in vivo. Hyperthermic isolated limb perfusion effectively suppresses tumor angioneogenesis respectively tumor growth. Competing interests =================== None declared. Authors\' contributions ======================= JP carried out the treatment and drafted the manuscript. MM carried out the treatment. CS carried out the immunohistochemical studies. JG participated in the design of the study. AD carried out the histological studies. WH participiated in the design of the study. TM participiated in the design and coordination of the study. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2407/4/53/prepub> Acknowledgements ================ This study was supported by grants from ELAN (Erlanger Leistungsbezogene Anschubfinanzierung und Leistungsförderung). Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Diagrammatic outline of the experimental setup. 1: roller pump system; 2: bubble oxygenerator; 3: heat exchanger, T: temperature :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### The mean tumor volumes + SEM of groups I to V (n = 30) 6 weeks after treatment; Group II, III, IV and V compared to group I (§: *P*= 0.036 ; \#: *P*= 0.021 (Kruskall-Wallis) n.s.:not significant). :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Immunohistochemical staining for bFGF in group I (control): Microvessels are represented by red clusters, which stand out sharply from other tissues. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Immunohistochemical staining for bFGF in group II (sham operated): More microvessels are represented in the margin of the tumor. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Immunohistochemical staining for bFGF in group V (HILP): Angioneogenesis is less pronounced as in group I. Brown clusters represent melanin. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Summarized data of the expression intensity of bFGF and CD34 ::: CD34 P to group I bFGF P to group I ----------- ------ -------------- ------ -------------- group I 19 \-\-\-- 20 \-\-\-- group II 28 0.001 26 0.003 group III 20 \> 0.05 21 \> 0.05 group IV 27 0.023 26 0.023 group V 16 0.017 16 0.001 ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Tumor response to treatment of isolated limb perfusion. Tumor diameters were measured and expressed as a percentage of the size at the time of treatment. ::: **response** **group I** **group II** **group III** **group IV** **group V** -------------- ------------- -------------- --------------- -------------- ------------- complete 0/6 0/6 0/6 0/6 2/6 partial 0/6 0/6 1/6 1/6 1/6 no change 0/6 0/6 0/6 0/6 1/6 progressive 6/6 6/6 5/6 5/6 2/6 :::

PubMed Central

2024-06-05T03:55:47.457638

2004-8-23

PMC516036

Background ========== The investigative agent tipifarnib is a member of a new class of drugs that were designed to function as a non-peptidomimetic competitive farnesyltransferase inhibitor (FTI). The principal behind this drug class is that protein farnesylation is required for many cell-signaling processes and that dysregulation of cell signaling is thought to be instrumental in driving cell proliferation in several malignancies. The hypothesis that gave rise to this exciting class of drugs is that the inhibition of this enzyme would reduce the uncontrolled cell signaling and provide some control over cell division and malignant cell proliferation. In hematological cancers, tipifarnib has shown significant inhibition of the proliferation of a variety of human tumor cell lines both *in vitro*and *in vivo*\[[@B1]-[@B3]\]. A recent phase I clinical trial of tipifarnib demonstrated a 32% response rate in patients with refractory or relapsed acute myeloid leukemia \[[@B4]\]. Furthermore, tipifarnib activity has also been seen in early clinical trials for patients with myelodysplastic syndrome (MDS) \[[@B5],[@B6]\], multiple myeloma (MM) \[[@B7]\], and chronic myeloid leukemia (CML) \[[@B8]\]. Mechanism of action (MOA) and biomarker studies with tipifarnib have focused on the oncogenic Ras protein. However, it has since been shown that inhibition of Ras farnesylation does not account for all of the compound\'s actions. For example, FTIs do not require the presence of mutant Ras protein to produce anti-tumor effects \[[@B4]\]. Several other proteins have been implicated as downstream targets that mediate the anti-tumorigenic effects of FTIs. The regulation of RhoB, a small GTPase that acts down-stream of Ras and is involved in many cellular processes including cytoskeletal regulation and apoptosis, has been proposed as a mechanism of FTI-mediated anti-tumorogenesis \[[@B9]\]. Additional proteins involved in cytoskeletal organization are also known to be farnesylated including the centromere proteins, CENP-E and CENP-F, protein tyrosine phosphatase, and lamins A and B. Thus, one possible mode of action of FTI\'s may be due to their inhibiting effects on cellular reorganization and mitosis. In addition to possibly inhibiting cellular reorganization and mitotic pathways, it is also known that FTIs indirectly modulate several important signaling molecules including TGFβRII \[[@B10]\], MAPK/ERK \[[@B11]\], PI3K/AKT2 \[[@B12]\], Fas (CD95) and VEGF \[[@B13]\]. The regulation of these effectors can lead to the modulation of signaling pathways involving cell growth and proliferation, and apoptosis. Thus, FTIs may have complex inhibitory effects on a number of cellular events. Where there are multiple candidate pharmacologic biomarkers as is the case with tipifarnib, a comprehensive, parallel study of all candidates is required. Here we describe the application of DNA microarray technology to the measurement of the steady-state mRNA level of thousands of genes simultaneously. This comprehensive experimental approach allows for the simultaneous analysis of candidate biomarkers as well as the generation of novel hypothesis on MOA and previously uncharacterized biomarkers. Biomarkers that enable the monitoring of drug response have the potential to facilitate clinical evaluation of the compound\'s safety and efficacy in humans. In the present paper we describe the use of global gene expression monitoring to identify genes and gene pathways that are modulated in acute myeloid leukemia (AML) following treatment with tipifarnib. Several genes involved in FTI biology were identified as being modulated following treatment with tipifarnib in addition to pathways involved with cytoskeletal organization, cell signaling, immunity, and apoptosis. This genome-wide approach of gene expression analysis has provided insight into genes that can be used as surrogate biomarkers for FTI drug activity as well as identifying putative pathways that are involved in the drug\'s anti-leukemic mechanism of action. This is the first successful report of the application of genomics to this novel class of drugs. Methods ======= Cell culture ------------ The AML cell lines AML-193, HL-60, THP-1, and U-937 were obtained from the American Type Culture Collection (Manassas, VA). Cells were grown in RPMI supplemented with 20% FBS. AML-193 was also supplemented with GM-CSF (10 ng/ml; PeproTech Inc., Rocky Hill, NJ), insulin (0.005 mg/ml; Sigma-Aldrich, St. Louis, MO), and transferrin (0.005 mg/ml; Sigma-Aldrich, St. Louis, MO). Cell numbers were counted in a hemocytometer and cell viability was determined by trypan blue dye exclusion assay. Tipifarnib was dissolved in 0.1% DMSO. The IC~50~was defined as the dose at which the number of viable cells in the treated sample was 50% of that in the control. This was determined after 7 days of drug treatment. Cytotoxicity assays were performed in duplicate. Control cultures were grown in medium containing vehicle (0.1% DMSO) only. Cells were analyzed for apoptosis by treating with vehicle or tipifarnib (100 nM and 1 μM) over a 5-day time course. Cells were stained with Annexin V and propidium iodide daily according to the manufacturers protocol (Roche Applied Science, Indianapolis, IN) and analyzed by FACS. Bone marrow processing ---------------------- Bone marrow samples were collected from consenting patients both before and during treatment with tipifarnib \[[@B4]\], diluted with PBS and centrifuged with Ficoll-diatrizoate (1.077 g/ml). White blood cells were washed twice with PBS, resuspended in FBS with 10% DMSO and immediately frozen at -80°C. Some characteristics of the two patient samples used in the present study are shown in Table [1](#T1){ref-type="table"}. Ras mutational status --------------------- Analysis of activating mutations in N-*ras*, K-*ras*, *and*H-*ras*codons was determined by PCR and RFLP analysis as previously described \[[@B1]\]. Microarray analysis ------------------- Total RNA was isolated using the Qiagen RNeasy kit (Qiagen, Valencia, CA) and treated with DNase1 (DNase1 kit, Qiagen, Valencia, CA) to remove any residual genomic DNA. Probe preparation was performed as previously described \[[@B14]\]. Linear amplification was performed on total RNA to obtain at least 15 μg of amplified RNA. Cell line mRNA and patient sample mRNA underwent one and two rounds of linear amplification respectively. Microarrays were generated and probes hybridized as described \[[@B15]\]. Samples were hybridized to arrays that contained 7452 cDNAs from the IMAGE consortium (Integrated Molecular Analysis of Genome and their Expression: ResGen™, Invitrogen Life Technologies, Carlsbad, CA) and Incyte libraries (Incyte, Palo Alto, CA). The intensity level of each microarray was scaled so that the 75^th^percentile of the expression levels was equal across micro-arrays. To control for chip errors, replicate clones on each chip that displayed a coefficient of variance (CV) greater than 50% of the mean were excluded from the analysis. Since background intensity was a maximum of 30 relative fluorescent units (RFU) for all experiments, a threshold of 30 RFU was assigned to all clones exhibiting an expression level lower than this. The microarray data were then normalized by quantile normalization and logarithmically transformed before further analysis. Statistical analysis -------------------- Analysis of variance (ANOVA) and t-tests were used to investigate the effect of drug treatment and time and their interactions for each gene. Multiple hypotheses testing was controlled by applying the false discovery rate (FDR) algorithm \[[@B16]\]. All statistical analyses were performed in S-Plus 6.1 (Insightful Corporation). Ratio matrices were generated based on pair-wise analysis of treated versus control samples. Hierarchical clustering was performed using a correlation metric and complete linkage (OmniViz Pro™, OmniViz, Maynard, MA). Pathway analysis ---------------- A total of 1198 genes that had a false discovery rate (FDR) \< 0.1 (p \< 0.05) in at least one cell line were used for the pathway analysis. Gene refseq accession numbers were imported into the Ingenuity Pathway Analysis software (Ingenuity Systems). 898 of these genes were mapped to the Ingenuity database. Seventy-two of these genes were also affected in patient samples (p \< 0.05, FDR \< 0.3) and were, therefore considered to be significantly regulated by tipifarnib. The identified genes were mapped to genetic networks available in the Ingenuity database and were then ranked by score. The score is the probability that a collection of genes equal to or greater than the number in a network could be achieved by chance alone. A score of 3 indicates that there is a 1/1000 chance that the focus genes are in a network due to random chance. Therefore, scores of 3 or higher have a 99.9% confidence of not being generated by random chance alone. This score was used as the cut-off for identifying gene networks significantly affected by tipifarnib. Real Time RT-PCR ---------------- The genes and primers used for RT-PCR are listed in Table [2](#T2){ref-type="table"}. Due to the limited amount total RNA from the patient samples, RNA that had been through one round of linear amplification was used. The Roche Molecular LightCycler (Roche Applied Science, Indianapolis, IN) with Syber Green I system detection was used for real time PCR. PCR thermocycling consisted of denaturation at 95°C for 45 seconds, followed by 30 cycles at 62°C for 10 seconds, and 72°C for 12 seconds. Samples were run in triplicate with both test primer sets and the control gene eukaryotic elongation factor 1 alpha (EEF1A1). This gene was used to control for differences in the amount of target material since initial microarray experiments found that expression of the EEF1A1 gene did not vary significantly between drug-treated and control cells. A standard curve was also run in each PCR reaction. Fold changes were calculated by normalizing the test crossing thresholds (Ct) with the EEF1A1 amplified control Ct. Results and Discussion ====================== Response of AML-like cell lines to tipifarnib --------------------------------------------- Tipifarnib inhibited the growth of 4 human AML cell lines in a dose-dependent manner. The IC~50~of these cell lines when treated with tipifarnib ranged from 19 to 134 nM (Table [3](#T3){ref-type="table"}). The mutation status of the *ras*oncogenes in the AML cell lines are also shown. These data indicate that the four AML-like cell lines are sensitive to tipifarnib treatment at concentrations well below the micromolar concentrations that is achievable in the bone marrow of leukemia patients \[[@B4]\]. However, there was no correlation between the type of *ras*mutation and sensitivity to the drug. These data are consistent with the activity of tipifarnib *in vivo*and allowed for further characterization of gene expression changes in these cells after treatment with pharmacologically relevant drug concentrations. Identification of genes differentially expressed in tipifarnib-treated AML cells -------------------------------------------------------------------------------- We next asked what genes are modulated following treatment of AML cells with tipifarnib and if there are differences between the affected gene networks in cell lines compared to primary cells from patients. To this end we first selected the three most sensitive cell lines and treated them with tipifarnib or vehicle alone over a 6-day time course. A standard concentration of 100 nM tipifarnib was chosen to ensure exposure within the pharmacologically active range of the compound (Fig [1](#F1){ref-type="fig"}). Samples for RNA analysis were harvested daily from duplicate cell cultures. Message RNA was isolated, amplified and hybridized to the cDNA microarrays containing approximately 7000 genes. Based on scatter plot analysis the microarray data was found to be highly reproducible between duplicate samples (Fig [2](#F2){ref-type="fig"}). A one-way ANOVA was employed to identify genes that were significantly changed over the 6-day time course compared to time-matched controls. A total of 1198 genes were significantly regulated (p \< 0.05 with a false discovery rate of less than 10%) in at least one of the cell lines over the time course (Supplementary Figure A \[see [Additional file 1](#S1){ref-type="supplementary-material"}\]). We also had access to bone marrow samples from two newly diagnosed AML patients enrolled in a phase I trial for tipifarnib \[[@B4]\]. The gene expression profiles in pre-treated leukemia cells were compared to those during drug treatment at days 8, 15 and 22. 1016 genes were significantly changed (p \< 0.05, FDR \< 0.3) during farnesyltransferase inhibition in vivo (Supplementary Figure B \[see [Additional file 1](#S1){ref-type="supplementary-material"}\]). A total of 180 genes were common between the cell line and patient data sets, 141 of these had known functions (see [Additional file 2](#S2){ref-type="supplementary-material"}). Real time RT-PCR showed good agreement with the microarray data (r^2^= 0.87; Fig [3](#F3){ref-type="fig"}). There are several known targets of FTIs including ras, RhoB, centromere proteins, lamins, PI3K/AKT, and TGFβRII \[[@B3],[@B10]\]. While the majority of these genes were present on our expression array (except the lamins) we only found k-ras to be significantly regulated. However, while not significant, up-regulation of TGFβRII was confirmed by RT-PCR (Fig [3](#F3){ref-type="fig"}). The absence of strong regulation of TGFβRII in the current data set may be due to the different FTI and/or the different culture conditions that were employed compared to previous reports \[[@B10]\]. Interestingly, k-ras was significantly down-regulated in our system. While k-ras is a target of FTIs it has been shown to undergo alternative geranylgeranylation when farnesylation is inhibited and may therefore not be an important anti-tumorgenic target post-translationally; however, it maybe a relevant target at the transcriptional level \[[@B17]\]. Repression of k-ras transcription has also been shown recently in a mouse model designed to identify genes that are related to the transformation-selective apoptotic program triggered by FTIs \[[@B18]\]. K-ras may therefore warrant further investigation as a candidate transcriptional target of FTIs. Identification of genetic networks affected by tipifarnib --------------------------------------------------------- To further refine the list of FTI-affected genes we next investigated which of these genes are known to interact biologically. To this end we carried out pathway analysis on the above 180 genes using the Ingenuity Pathway Analysis (IPA) tool. Seventy-nine (72 unique) of these 180 genes mapped to genetic networks as defined by the IPA tool. These networks describe functional relationships between gene products based on known interactions in the literature. The tool then associates these networks with known biological pathways. Five networks were found to be highly significant in that they had more of the identified genes present than would be expected by chance (Table [3](#T3){ref-type="table"}). These networks were associated with the cell cycle, apoptosis, proliferation, chemotaxis, and immunity pathways. The study by Kamasani et al also found cell cycle pathways were repressed and immunity and cell adhesion pathways were activated by FTI treatment \[[@B18]\]. The 79 genes were then analyzed by two-way hierarchical clustering to compare the expression profiles of the AML samples (Fig [4](#F4){ref-type="fig"}). A number of observations could be made using this visual approach. First, although there were some outliers, the majority of duplicate samples clustered close together again demonstrating the reproducibility of the results. Similarly, a number of replicate clones of the same gene clustered next to each other thereby improving the confidence of the microarray data. As expected, samples from the same cell line or patient clustered together. However, samples from late in the time courses have very different expression profiles possibly reflecting greater differences in the transcriptional activity between control and treated cells at this late stage of drug treatment. Interestingly, the cluster analysis showed that the HL-60 profile was most similar to the patient samples indicating it has a more similar response to tipifarnib compared to the patient cells than THP-1 and U-937. This similarity cannot be associated with FAB sub-type since HL-60 was isolated from a patient with M2 AML and the patients examined in this study were M4 and M5 sub-types. Therefore, it is suggested that the different expression profiles seen are due to other genetic differences that impact the specific down-stream effects of FTI inhibition. This may be important when considering appropriate models for FTI investigations. While the cell lines portrayed higher heterogeneity in expression changes compared with the patient samples, the hierarchical clustering did reveal a common set of up- and down-regulated genes. A set of 23 genes was found to be down-regulated in the cell line and patient samples (Fig. [4](#F4){ref-type="fig"}). The major network associated with these genes contained several involved in proliferation including CSK, FGFR3, KRAS2, PPARG, RET, and USF1. Alternatively, 29 genes were commonly up-regulated and network analysis of these revealed activation of apoptotic- and immune-related genes, including CASP6, CD48, FGR, IGF2R, PECAM1, and TNFRSF5. It will be of interest to investigate these genes further to see if they are transcriptional targets of FTIs and if their regulation is additive or synergistic to FTI efficacy. Due to the stringency of our gene selection process it is likely that many genes that are indeed regulated by FTIs, were not identified. For instance, as noted above, of the targets known to be affected by FTIs we identified only k-ras at the transcriptional level. However, the use of pathway analysis tools allows for the identification of networks of genes that are known to interact with each other. This procedure therefore provides additional confidence in the selected genes as well as clues to other genes that may also be regulated but not identified as being significant by the microarray analysis. For example, the network of up-regulated genes (Fig [5A](#F5){ref-type="fig"}) includes the lamin B gene, which is indeed a direct target of FTIs. Also, the PIK3R2 gene, which regulates AKT and is a known target of FTIs \[[@B3]\], can be found in the down-regulated network of genes (Fig. [5B](#F5){ref-type="fig"}). This illustrates that the pathway analyses correctly identifies genes that have previously been demonstrated to be either direct or indirect targets of farnesyltransferase inhibition and provides a greater context for screening candidate genes modulated by FTIs. Investigation of apoptosis -------------------------- Since a number of apoptotic genes were identified as being affected by tipifarnib we performed experiments in THP-1 and HL-60 cell lines to verify if they were indeed undergoing apoptosis. Previous reports have shown that two other FTIs can induce apoptosis in myeloid leukemia cell lines \[[@B11]\] and that tipifarnib causes apoptosis in other malignancies including multiple myeloma \[[@B19]\], and melanoma \[[@B1]\]. Annexin V staining demonstrated a significant increase in FTI-mediated apoptosis in THP-1 for both 100 nM (p = 0.027) and 1 uM (p = 0.032) concentrations of tipifarnib (Fig [6](#F6){ref-type="fig"}). A maximum of 23% apoptotic cells were demonstrated at day 5 (Fig. [6](#F6){ref-type="fig"}). No difference in the level of apoptosis was seen between 100 nM and 1 μM of tipifarnib. While apoptosis was activated in the HL-60 cell line this was found to be non-specific since control cells also exhibited this phenomenon during cell culture (data not shown). The lack of FTI-specific apoptosis in HL-60 is consistent with a recent report that also failed to demonstrate tipifarnib-mediated apoptosis in primary AML blasts \[[@B20]\]. However, in that report apoptosis was measured only two days after treatment where here we found a marked increase in apoptosis at days 3--5. Therefore, our data indicate that tipifarnib can cause apoptosis in AML but may not be detectable at early time points or in AML with certain genetic backgrounds. Conclusions =========== Tipifarnib is one of three FTIs that are currently in clinical trials for treating a variety of cancers \[[@B21]\] and it is showing promise in hematological malignancies \[[@B3]-[@B8]\]. While FTIs were originally designed to inhibit the function of the *ras*oncogene it has been recently demonstrated that there is no correlation between patient response and *ras*mutational status \[[@B4]\]. Additionally, it is clear that other targets of FTIs exist that provide equally important anti-cancer properties. We have reported the use of microarray analysis of both primary human AML cells and AML cell lines following treatment with tipifarnib in order to identify genes and gene pathways that are modulated by this FTI. In particular, genes involved in signaling pathways, down-stream cytoskeletal pathways, and apoptotic events were described. Pharmacodynamic markers that are currently used in the clinic, such as lamin A and HDJ2 \[[@B22]\], are direct markers of farnesyltransferase inhibition while the majority of genes identified in this work are likely downstream transcriptional targets. Both of these current candidate markers were not present on our microarrays so we did not report on their expression changes. Further analysis will be required to elucidate whether the expression changes seen in our work are due to direct or indirect effects of FTIs. Also, while the currently used clinical biomarkers do not correlate with patient response to FTIs the genes identified here may be candidates for patient stratification \[[@B3]\]. We are therefore in the process of examining bone marrow specimens from larger phase 2 clinical trials with the aim of validating the panel of pharmacodynamic gene expression markers we have identified here. Such pharmacogenomic analysis will be very important in further elucidating the action of FTIs while providing a platform for identifying patients who could potentially respond to tipifarnib therapy. Competing interests =================== MR, RTB, DA, and YW are employees of Veridex LLC., a Johnson and Johnson company. JEK and JEL have received fees from Johnson & Johnson Pharmaceutical Research and Development in the last 5 years. Author\'s contributions ======================= MR wrote the manuscript and performed experiments. RTB performed Ras analysis. DA helped conceive the experiments. JEK and JEL were principal investigators in the clinical trials from which samples were received. YW helped conceive the experiments. All authors read and approved the final manuscript. List of abbreviations ===================== AML: acute myeloid leukemia; FDR: false discovery rate; FTI: farnesyltransferase inhibitor; MDS: myelodysplastic syndrome; MOA: mechanism of action; MM: multiple myeloma; CML: chronic myeloid leukemia Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2407/4/56/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional File 1 A. Two-way hierarchical clustering of 1198 genes regulated in three AML cell lines after tipifarnib treatment. A fold-change ratio was calculated using the treated sample and its matched untreated sample. B. Gene expression changes in patient AML cells. Two-way hierarchical clustering of 1016 genes regulated in two AML patients over a three-week time course. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 A list of 141 genes with known function that were regulated in both the cell line and patient data sets. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We gratefully acknowledge Christine Lloyd for performing real time RT-PCR assays, and Anton Bittner and the Johnson and Johnson Pharmaceutical Research and Development microarray department for assisting with microarray procedures. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Growth profiles of AML cell lines treated with tipifarnib. Duplicate cultures were inoculated into 6-well plates at an initial concentration of 1 × 10^5^cells/ml. Tipifarnib was supplemented at a concentration of 100 nM in 0.1% DMSO. Duplicate control cultures were grown in medium containing 0.1% DMSO only. Duplicate cultures were harvested daily for a total of six days. Error bars are standard deviations. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Scatter plot analysis of microarray data. Duplicate THP-1 and HL-60 cell line cultures were harvested, processed independently, and hybridized to the cDNA array. Duplicate samples from patients A and B were also analyzed for reproducibility. The lines on the scatter plots indicate the 1.5-fold and 1.7-fold boundaries used for selecting genes with differential expression in cell line samples and patient samples, respectively. Less than 5% of genes were outside these fold-change thresholds. Linear regression was performed and correlation coefficients are shown. Axes show the fluorescence intensity associated with each gene (log~10~). :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Real time RT-PCR validation of microarray data. Nine genes were randomly selected for real time RT-PCR. Two of these genes (adipsin and vimentin) were identified as being significantly regulated in both the cell lines and in *de novo*AML patients. The \"fold-change\" (Log~2~) of the RNA transcripts was calculated in the patient who responded to tipifarnib (patient A) by using the treated (day 15) versus the matched baseline sample for both PCR and microarray data. Linear regression analysis was performed and the coefficient of variation was calculated. Italicized genes were identified as being significantly regulated by tipifarnib in both AML cell lines and patient samples. Error bars are standard deviations. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Hierarchical clustering of genes regulated after tipifarnib treatment. A fold-change ratio was calculated using the treated sample and its matched untreated sample. Duplicate samples are indicated with suffices \"a\" and \"b\". Number suffices indicate day of treatment. The color bar indicates the fold-change (log~2~). Red is up-regulated, blue is down-regulated. White indicates no change. Cluster A and B were associated with genes that were largely down- and up-regulated across all samples, respectively. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### Networks of genes commonly regulated after tipifarnib treatment. **(A)**Twenty-three genes that were down-regulated in patient leukemic cells and AML cell lines were analyzed by the Ingenuity Pathway Analysis tool. The shown major network that was found to be significantly down-regulated by tipifarnib was associated with proliferation (p = 10^-10^). **(B)**Twenty-nine genes that were up-regulated were also analyzed for associated networks that were significantly affected by tipifarnib. The shown network was significantly associated with apoptosis (p = 10^-10^) and immunity (p = 10^-7^). Shaded genes are the genes identified by microarray analysis and others are those associated with the regulated genes based on the pathway analysis. The meaning of the node shapes is also indicated. Asterisks indicate genes that were identified multiple times. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### Detection of tipifarnib-mediated apoptosis in AML cells. **(A)**Annexin V staining shows that a decrease in cellular proliferation correlates with an increase in apoptosis in the THP1 cell line following treatment with tipifarnib. **(B)**Apoptosis assay of control and treated cells at day 5. Annexin V stains both apoptotic and necrotic cells, propidium iodide stains necrotic cells only. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Characteristics of patient AML samples. ::: Patient Dose (mg) Age/sex Diagnosis Stage Antigen expression *ras*status --------- ----------- --------- ----------------- --------- -------------------- ------------- B 100 59 M M5, de novo AML Relapse CD34+, CD33+ WT A 300 75 M M4, de novo AML New CD34+, CD33+ WT WT = wild-type ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Primer sequences used for RT-PCR. ::: **Gene** **Forward Sequence** **Reverse Sequence** ---------- ----------------------- ----------------------- RAC1 CACGATCGAGAAACTGAAGGA AGCAGGCATTTTCTCTTCCTC TIMP1 TACTTCCACAGGTCCCACAAC GTTTGCAGGGGATGGATAAAC TGFβR II CAGCGTTTCAAAAAGTGAAGC CTAGACTGGGGTCCAGGTAGG βGLOGIN GCAACCTCAGACAGACACCAT ACCTTAGGGTTGCCCATAACA PI3K TGAGCAAGAGGCTTTGGAGTA TTCCTATGCAATCGGTCTTTG ERK3 GAGCCAGTAGAGGATGGGAAG GATGAGGAATTTGAGGGGAAG VIMENTIN ATCGATGTGGATGTTTCCAAG TGTCTCCGGTACTCAGTGGAC FTP ATCCCTTAGCATCAGCTCCTC CGTTCTTTTGGCATTAGTTGG ADIPSIN CCTGCATCTGGTTGGTCTTTA AGCCTCCTGAGTAGCTGGAAC EEF1A1 GATGCATTGTTATCATTAACC CATGCAAGTTTGCTGAGCTG ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Anti-proliferative effects of tipifarnib for AML cell lines. ::: Cell line IC~50~(nM)† *ras*status ----------- ------------- ----------------------------- AML-193 134 H-*ras*(12), N-*ras*(13) HL-60 24 H-*ras*(12) THP-1 19 K-*ras*(13) N-*ras*(12, 61) U-937 44 Wild-type † The IC~50~was calculated from two independent experiments. The mean value is shown. ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Genetic networks affected by tipifarnib ::: **Network** **Genes in Ingenuity network\*** **Associated pathways** **Score†** ------------- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ----------------------------------------------------- ------------ 1 CD226, CD24, CD36, CD63, **CEBPD**, **CRKL**, CSF2RB, **CSK**, **DF**, DOK1, **FACL2**, FCAR, **FCER1G**, **FGFR1**, GAB2, GP6, **GRIN1**, ICAM1, ICAM2, ICAM3, INPP5D, **INSR**, **ITGAL, ITGB2**, LAT, **LYN**, **MAD2L1**, NEDD9, **PECAM1**, **PPARG**, **PTK2B**, SORBS1, STAT5B, SYK, **UCP2** Immunity Inflammation Apoptosis Cell death Adhesion 24 2 **ABCB1**, **ADORA3**, **APOC3**, AR, **BRCA1**, **C20ORF14**, CAV1, CCNA1, CCNB1-RS1, CCND1, **CDK2**, CDKN2A, **CDKN2D**, CKS1B, CREM, E2F1, E2F4, **ELA2**, **ESRRA**, **FGFR3**, GNAQ, GNB1, **GNG5**, HDAC1, ICAM1, JUN, KLKB1, LAMP1, **MKI67**, **PLCB2**, PTPN2, **PTPN9**, RB1, **RBBP2**, TTK Cell cycle Transcription Apoptosis 20 3 ABCB1B, ARL6IP, BAX, BCL2, **CAPN1**, CDC25C, CDKN1A, CEBPE, CREM, CSF2RA, **CSF3R**, CTCF, **EMP1**, F2, **FACL3**, GPI, **LTB4R**, MAPK14, **MARK3**, **MEF2A**, MYC, MYCN, **PIM1**, PPP1R15A, PTGS2, RPL21, RPL6, RPS12, RPS16, **RPS5**, **SERPIND1**, SLC19A1, SP1, TP53, **USF1** Cell cycle Cell death Proliferation 13 4 ANXA1, **ANXA2**, APOH, BCR (BREAKPOINT CLUSTER REGION), CD22, **CD48**, DOK2, EIF3S2, EIF3S6, **EIF3S7**, EPHA4, **ERBB3**, **FGR**, FN1, FYN, GRB2, HCLS1, **IGF2R**, **IL6R**, IL6ST, **KRAS2**, LCK, LIF, LY6, PLAU, PLAUR, PLG, **RET**, S100A10, SHC1, **SLC11A1**, **ST14**, TNC, TRPV4, ZFP106 Chemotaxis Proliferation Apoptosis 13 5 ACT1, **APLP2**, BIRC4, CAMKL, CASP3, **CASP6**, CASP8, CASP9, **CCL20**, DRPLA, EGR1, EIF4B, **FKBP5**, HNF4A, **IFI30**, LMNB1, **LYZ**, NR0B2, NR3C1, NUMA1, PCYT1A, PLEC1, PRKCZ, RELA, RELB, RIPK2, **RPS6KA1**, TAX1BP1, **TNFRSF25**, **TNFRSF5**, TRAF1, TRAF3, TRAF5, **VIM**, WEE1 Apoptosis Cell death Proliferation 11 \* Bold genes are those identified by the microarray analysis. Other genes were either not on the expression array or not significantly regulated. † A score of \> 3 was considered significant (p \< 0.001). :::

PubMed Central

2024-06-05T03:55:47.459870

2004-8-25

PMC516037

Background ========== Upper limb pain and dysfunction are frequent complaints associated with computer work. However, the responsible pathology and the pathophysiological mechanisms are insufficiently understood. In addition, there is no consensus with regard to physical findings that may reflect symptoms. The involvement of the nerves in \"non-specific\" upper limb disorder, e.g. in computer operators, is suggested by various observations: The demonstration of an elevated threshold to vibratory stimulation \[[@B1]-[@B3]\]; abnormal upper limb tension tests \[[@B4],[@B5]\]; reduced nerve mobility \[[@B6],[@B7]\]; abnormal nerve tenderness (mechanical allodynia) \[[@B8]\]; changed axonal flare reaction \[[@B9]\]; allodynic response to supra-threshold vibration \[[@B2]\]; reduced muscle strength \[[@B10],[@B11]\] and sympathetic reflexes \[[@B12]\]; and thermographic changes \[[@B13]\]. Still, clinical practice and epidemiological studies tend to attribute upper limb symptoms in computer operators to a disorder in muscle, tendon, or insertion \[[@B14]\]. Focal neuropathy including carpal tunnel syndrome is infrequently reported \[[@B15],[@B16]\]. Upper limb pain in computer operators shares the features of a neuropathic pain: Common analgesics tend to be ineffective. Pain may be evoked spontaneously or may appear to constitute an abnormal response to stimuli with frequent occurrence of allodynia. In addition, there are often non-painful abnormal spontaneous or evoked sensory phenomena such as numbness/tingling. The common experience of weakness which may further deteriorate on use would also be compatible with an upper limb nerve affliction. A precise and accurate diagnosis is crucial for effective management and rehabilitation, and also for epidemiological studies concerning causation. In order to get a better understanding of the pathophysiological mechanisms, the injured tissue should be precisely located. This might not necessarily be where symptoms predominate. I have aimed at studying a clinical series of computer operators with upper limb complaints and dysfunction in terms of • exposure characteristics; • symptoms and past treatment; • physical findings which may reflect an affliction of the peripheral nerves; • prognosis with regard to symptoms and work-status. Methods ======= Patients -------- This study comprises a consecutive series of 21 computer-aided designers with pain and functional limitations in the dominant upper limb. All patients were referred to a department of occupational medicine for diagnostic and aetiological assessment and management. Three patients were males of median age 27 years (range 25--41) and 19 were females of median age 35 years (range 25--55). Clinical examination and interpretation --------------------------------------- ### Interview Patients were interviewed about the character, distribution, initial presentation and development of their symptoms. Special attention was given to the presence of upper limb pain, subjective weakness and numbness/tingling, and to other symptoms included in a standard protocol for work-related upper limb disorders \[[@B17]\]. ### Physical examination A subsequent physical examination included extracts of diagnostic criteria for selected clinical disorders (tension neck syndrome, cervical syndrome, supra-and infraspinous tendinitis, bicipital tendinitis, frozen shoulder, acromioclavicular arthrosis, epicondylitis, tenosynovitis, and wrist and forearm peritendinitis) \[[@B17]\]. Upper limb nerve afflictions were defined from an additional neurological examination consisting of the following components: • Manual assessment of the isometric strength in a selection of ten upper limb muscles (Figure [1](#F1){ref-type="fig"}). Any reduction of strength was registered as weakness \[[@B18],[@B19]\]. Patients were encouraged to provide maximal muscle effort on both sides for each muscle tested, despite any potential discomfort. • The sensibility (algesia by pinprick, aestesia by moving touch \[[@B19]\]) was assessed in five homonymous innervation territories (Figure [2](#F2){ref-type="fig"}): [• The axillary nerve (the deltoid area);]{.underline} [• The musculocutaneous nerve (the dorsal forearm);]{.underline} [• The radial nerve (the first dorsal web);]{.underline} [• The median nerve (the tip of the second finger);]{.underline} [• The ulnar nerves (the tip of the fifth finger).]{.underline} *[• The perception of vibration (tuning fork 256 Hz]{.underline}*\[[@B20]\]) *[was additionally estimated in the ulnar and median territories (tips of second and fifth fingers).]{.underline}* Any sensory deviation from normal was registered as abnormal. • Assessment of tenderness with slight pressure at 14 locations along the course of nerves \[[@B8]\]. Any mechanical allodynia was registered as abnormal: [• The brachial plexus (scalene triangle, passage behind the pectoralis minor muscle);]{.underline} [• The suprascapular nerve (suprascapular notch);]{.underline} [• The axillary nerve (quadrilateral space);]{.underline} [• The musculocutaneous nerve (passage through the coracobrachial muscle);]{.underline} [• The median nerve (just proximal to the elbow, at the passage between the two heads of the pronator teres muscle, at the passage below the arcade of the common superficial flexor muscle, and at the carpal tunnel);]{.underline} [• The radial nerve (triceps and brachioradial arcades);]{.underline} [• The posterior interosseous nerve at the arcade of Frohse (supinator tunnel);]{.underline} [• The ulnar nerve (sulcus of the ulnar nerve and Guyon\'s canal in the hypothenar).]{.underline} Assessments in patients with unilateral disorder were based on comparison to contra-lateral findings defined as normal. In patients with bilateral disorder, test-results were related to other findings in the same limb assumed to be normal, e.g., strength in adjacent muscles or sensibility in adjacent innervation territories \[[@B21]\]. The definition and location of a nerve affliction (\"neuropathy\") was based on a traditional approach with a focus on the topography and innervation patterns of the upper limb nerves. Special consideration was given to the presence of normal strength in certain muscles and of reduced strength in others \[[@B18],[@B19]\], and to localized mechanical allodynia at the appropriate location(s) along the nerve trunk \[[@B8]\]. I have operated with two sets of criteria for the definition of focal neuropathy assuming the second criterion to be more convincing: • *Criterion 1:*The presence of a pattern of muscle-weakness suggesting a focal neuropathy at a defined location, at which mechanical allodynia with slight pressure at the nerve is present. • *Criterion 2:*Criterion 1 plus sensory deviations from normal in one or several sensory territories located peripherally to focal neuropathy. In addition, double crush \[[@B22]\] at the appropriate location(s) was arbitrarily defined when strength reductions and/or mechanical allodynia were either equivalent or more prominent distally. The double crush theory refers to the phenomenon that a focal neuropathy increases the vulnerability of the nerve as a whole, resulting in a tendency of focal neuropathy to occur at several locations along the course of a nerve. Location of neuropathy ---------------------- *Brachial plexus neuropathy at chord level*was defined with reduced strength in the deltoid, biceps, and radial flexor of the wrist muscles, when weaknesses were accompanied by brachial plexus tenderness at its passage behind the pectoral muscle. Depending on the extent of brachial plexus involvement, additional muscles may be weak and mechanical allodynia may extend in the proximal or medial direction. *Median neuropathy at elbow level*was defined with reduced strength in the radial flexor of the wrist muscle along with mechanical allodynia involving the median nerve at elbow level (at the passage proximal to the elbow, between the two heads of the pronator teres muscle, and/or at the arcade of the superficial flexor of digits muscle). With an isolated median neuropathy, the deltoid, biceps, and ulnar extensor of the wrist muscles must be intact. Double crush involving the brachial plexus [and]{.underline} the median nerve was defined in the following situations: • Strength in the radial flexor of the wrist muscle was reduced as much as / more than it was in the deltoid or biceps muscles. • Mechanical allodynia was either the same or more conspicuous at the median nerve at elbow level than it was at plexus level. *Posterior interosseous neuropathy*was defined with reduced strength in the ulnar extensor of the wrist muscle along with tenderness at the nerve-passage below the arcade of Frohse in the dorsal proximal forearm. With an isolated posterior interosseous neuropathy, the deltoid, biceps, short radial extensor of the wrist, and radial flexor of the wrist muscles must be intact. Double crush involving the brachial plexus [and]{.underline} the posterior interosseous nerve was defined in the following situations: • Strength in the ulnar extensor of the wrist muscle was reduced as much as / more than in the deltoid, biceps, or radial flexor of wrist muscles. • Mechanical allodynia was either the same or more conspicuous at the arcade of Frohse than it was at plexus level. *Other potential focal neuropathy*was defined according to similar criteria, e.g., an isolated carpal tunnel syndrome would require reduced strength in the short abductor of the wrist muscle but preserved strength in the radial flexor of the wrist muscle. An isolated ulnar neuropathy at elbow or wrist level would require reduced strength in the abductor of the fifth digit and intact proximal muscles. In addition, mechanical allodynia should be present at the appropriate locations along nerve trunks. Management ---------- Patients were recommended to freely move and use the symptomatic upper limb within the limits of immediate and subsequent pain aggravation. All patients were offered physiotherapy based on the concept of adverse neural tension \[[@B23],[@B24]\] and encouraged to return to computer work after optimizing the work station ergonomics and work organization. Patients unable to return to work were advised concerning rehabilitation: Maximizing variation during future work; keeping the upper limbs close to the body; and avoiding repetition and static postures. Questionnaire ------------- 1/2 -- 1 1/2 years after the initial examination the patients responded to a questionnaire: The exposure characteristics; symptoms (pain, weakness/fatiguability, numbness/tingling); past treatment; pain intensity at the first encounter and at follow-up to be quantified on a VAS-scale from 0 (no pain) to 10 (extreme pain); and the present status with regard to functional limitations and work. Statistics ---------- The change of the level of reported pain between the first consultation at the department and at follow-up was assessed by Friedman\'s test. Results ======= Exposure characteristics ------------------------ All 21 patients returned the questionnaire. The mean duration of work with computer-aided design was 95 months (16--260 months). The self-reported daily mean of time spent with computer work constituted 81% (50--100%) of the total working time. 86% of the respondents reported aggravating factors during the months prior to the onset of symptoms, including high work intensity, overwork or other work conditions causing an unusual strain. Symptoms and past treatment --------------------------- Pain in the dominant upper limb was common to all patients. It had a mean duration of 24 months (1--60 months) and was the main symptom in 13 patients. All but one patient had a subjective feeling of weakness/fatiguability. Five patients reported this to be the most disturbing symptom. 19 patients experienced numbness/tingling which constituted the main symptom in three of them. Five patients had bilateral symptoms. All patients had received treatment prior to admission: A limited and transitory effect of past physiotherapy was reported in four out of 17, of pain killers in one out of 10, and of local steroid injections in two out of three patients. For the remaining patients the past treatment had no effect. Physical examination -------------------- According to the defined criteria for work-related upper limb disorders \[[@B17]\], non-neuropathic disorders were not identified. In all 21 patients reduced strength was demonstrated in the following muscles: Deltoid, biceps, triceps, and the radial flexor, short radial extensor and ulnar extensor of the wrist. In a smaller number of patients there were additional strength-reductions in the pectoral, infraspinatus, latissimus and abductor of the fifth digit muscles (Figure [1](#F1){ref-type="fig"}). Sensory abnormalities were identified in 19 out of 21 patients. The median nerve territory was most frequently involved. However, most patients had additional sensory deviations in territories innervated by the radial, musculocutaneous, axillary, or ulnar nerves (Figure [2](#F2){ref-type="fig"}). In all patients, mechanical allodynia was present at the brachial plexus at chord level, i.e., on its passage behind the pectoral muscle. In two patients it was also present at trunk level, i.e., at the scalene triangle. Furthermore, mechanical allodynia was present in all patients at the posterior interosseous nerve at the arcade of Frohse, and at the median nerve at one or several locations around the elbow. No mechanical allodynia was observed at the suprascapular, axillary, musculocutaneous, ulnar or radial nerves, nor at the median nerve at the volar wrist (Figure [3](#F3){ref-type="fig"}). Contra-lateral mechanical allodynia was present at the brachial plexus in five patients, and additionally at the posterior interosseous and median nerves in four and three of these patients, respectively (Figure [3](#F3){ref-type="fig"}). The five patients with contra-lateral mechanical allodynia had bilateral complaints. According to the defined criteria, the patterns of physical findings in all 21 dominant limbs suggested the presence of a brachial plexus neuropathy in combination with a median and posterior interosseous neuropathy at elbow level (Figure [4](#F4){ref-type="fig"}). No other nerve entrapments were defined in this sample. Prognosis --------- At follow-up after 1/2 -- 1 1/2 years, only two out of the 21 patients remained in computer work. Three were employed in other jobs while the majority was training for other jobs (eight patients) or unemployed (eight patients). Eleven patients reported a beneficial effect of the proposed physiotherapy, five experienced no effect, and five did not pursue the suggested treatment. On a group basis, the mean of pain when worst was reduced from 8.0 at enrolment to 6.1 at follow-up. The corresponding figures when pain was least were 4.3 and 2.7, respectably. This reduction was significant (χ^2^= 8.0 and 9.0, p \< 0.005 and 0.003, respectively). However, the pain persisted on a disturbing level in the majority of patients (Figure [5](#F5){ref-type="fig"}). The severity of pain at follow-up was unrelated to the present occupational status and to the severity of pain at enrolment. Neither of the two parameters was related to sex, age, or the duration of exposure. Discussion ========== Working in computer-aided design involves an almost continuous operation of the pointing device for extended periods of time. Consequently, the dominant upper limb strain is more pronounced than in other computer work. Following this exposure, all the patients were severely handicapped in this very limb. In most patients, the unilateral disorder enabled the examiner to conveniently compare the outcomes of the physical examination in the dominant limb with the contra-lateral findings. There is a general consensus that reduced muscle strength, sensory deviations from normal, and localized mechanical allodynia are related to afflicted peripheral nerves \[[@B19]\]. The relation to an underlying neurological change is indicated by the occurrence of these abnormalities in patterns, in accordance with anatomical facts. The complaints of pain in all dominant upper limbs, and subjective complaints of weakness and numbness/tingling in most of them, were reflected by a rather uniform pattern of strength-reductions, mechanical allodynia, and sensory deviations from normal suggesting the involvement of the brachial plexus at chord level and of the posterior interosseous and median nerves at elbow level. In this sample, I found no indication of carpal tunnel syndrome, ulnar neuropathy at elbow or wrist level, radial neuropathy before division of the posterior interosseous nerve, nerve root compression, or any other neuropathic or non-neuropathic upper limb disorder. Abnormalities were detected by tests which are included in the classical neurological examination. To enable the examiner to assess a single muscle at a time, the limb position during testing should aim at limiting disturbing interference from other muscles \[[@B21]\]. The mostly minor strength reductions demanded the quantification of strength-reductions to include Grade 4+ (Contraction against gravity and strong resistance \[[@B18]\]). The identification of such minor weakness demands simultaneous testing on the right and the left side. The absence of facial expressions, withdrawal, or complaints from the patients suggested that the voluntary contraction during strength testing was not influenced by simultaneous pain. The physical examination in this clinical case study was not blinded concerning patient-related information, such as the presence and location of symptoms. However, in a recent validation of the physical tests applied we have found that blinded examiners could reliably assess the individual items (individual muscle strength \[[@B21]\], sensory qualities, mechanical allodynia), as well as the occurrence of findings in patterns in accordance with the course of nerves and the innervated tissue. Findings were also reflected by the presence of symptoms. Clinical experiences have led to hypotheses suggesting upper limb symptoms in computer operators to be related to prolonged non-neutral and predominantly static positions including a flexion of the shoulder and a submaximally pronated forearm. This may result in a muscular imbalance from some muscles being successively shortened and their antagonists passively stretched and weakened \[[@B25]\]. Pain and functional limitations may result from limited available space for the nerves especially at locations close to joints or adjacent to bony prominences, fibrous bands or tunnels. This may cause tension, friction, and compression \[[@B24],[@B25]\]. Reduced axoplasmatic flow at a proximal site may lessen the ability of nerves to withstand adverse forces at a more distal site (or the reverse), as described in the double crush phenomenon \[[@B22]\], and the mobility of the entire nerve may be impaired by such external affliction \[[@B6],[@B24],[@B26]\]. From the topography of the brachial plexus, the lateral chords would appear to be most at risk behind the pectoralis minor muscle. The muscles supplied from this part of the plexus (deltoid, biceps, radial flexor of wrist, triceps, short radial extensor and ulnar extensor of the wrist) were invariably involved. In a few limbs there was an additional involvement of the pectoral, small abductor of the fifth digit, latissimus dorsi, and infraspinatus muscles (Figure [1](#F1){ref-type="fig"}). This is concurrent with a medial or proximal extension of a brachial plexus affliction. Caution should be exercised when drawing a comparison between the outcome of this study of patients, referred with a serious disorder, and studies of \"healthy\" computer operators in occupation. Still, it would be relevant to compare with upper limb findings in computer operators described by others. A study of 533 visual display terminal workers has suggested an array of upper limb disorders in 22%, dominated by tendon related conditions in 15% and probable nerve entrapment in 4% \[[@B15]\]. In a study of 632 newly hired computer operators, the one year incidence of neck and shoulder symptoms was 58% and of hand/arm symptoms 39%. Symptoms were explained by physical findings in the neck/shoulder region in 35% of participants (\"somatic shoulder/neck syndrome\" in 33%) and in the hands/arms in 21% (de Quervain\'s syndrome in 15%) \[[@B27]\]. In a recent major cross-sectional study of almost 7000 computer operators, 20% complained of moderate to severe pain. The physical examination, however, was only able to disclose a limited number of upper limb disorders, similar to what would be expected in the general population \[[@B28]\]. Self-reported numbness/tingling in 10.9% of the computer operators was attributed to carpal tunnel syndrome in a minority (numbness/tingling in the median nerve territory in 4.8%, and symptoms at night in 1.4%) and unexplained in the remaining subjects \[[@B16]\]. Based on the same material, nerve entrapment was only diagnosed in 12 subjects (supinator syndrome and pronator syndrome defined by localized palpation tenderness with withdrawal, and pain with provocative maneuvers). No new cases of nerve entrapment occurred during a one year follow-up \[[@B29]\]. However, the diagnoses depend on the choice and validity of the clinical tests employed and on the diagnostic criteria applied. The \"somatic shoulder/neck syndrome\" \[[@B27]\] is characterized by nonspecific signs and may well be a neuropathic condition. Discomfort with the Finkelstein maneuver \[[@B15],[@B27]\] is not specific for de Quervain\'s syndrome \[[@B30]\]. If not associated with first dorsal compartment tenderness and swelling, this diagnosis would seem to be unjustified. The common occurrence of de Quervain\'s syndrome in computer operators who hardly move their thumb would seem unlikely. My findings are more in accordance with those of Pascarelli, who studied 485 upper limb patients out of which 70% were computer operators. A detailed and comprehensive physical examination demonstrated protracted shoulders in 78% and head forward position in 71%. This was also frequent in my study-patients but not systematically registered. A neurogenic thoracic outlet syndrome in 70% was suggested by tests stressing the brachial plexus and by the demonstration of mechanical allodynia \[[@B31]\]. In a former study of 53 computer operators with severe upper limb disorders Pascarelli has found a high prevalence of reduced muscle strength and impaired passive wrist deviation associated with an increase in forearm pain. These findings were attributed to myofascial shortening and found to be useful clinical indicators of injury \[[@B32]\]. Conclusions =========== The limited success of the prevention and management of computer-related upper limb disorders demands new approaches to practice and research in the field. The inclusion in future studies of the presented systematic examination of the upper limb nerves may provide additional diagnostic information. This may lead to future improvement of the prevention and management of computer-related upper limb disorders. Competing interests =================== None declared. Authors\' contributions ======================= JRJ designed the study and conducted all the clinical examinations, and wrote the manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2474/5/26/prepub> Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Strength reductions in 21 dominant limbs. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Sensory abnormalities (pain, touch, and/or vibratory sense) in 21 dominant limbs. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### Mechanical allodynia with mild pressure at three locations along nerves in 21 dominant limbs compared to the contra-lateral limb. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Definition according to the applied criteria of focal neuropathy in 21 dominant limbs. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Change of pain from enrolment until follow-up in 21 dominant limbs.**The mean pain scores when pain was worst and when pain was least is indicated for each patient at enrolment and at follow-up. Red dots represent one case. Green dots represent two cases. :::  :::

PubMed Central

2024-06-05T03:55:47.462653

2004-8-13

PMC516038

Background ========== A work-related musculoskeletal disorder (WRMD) is defined as a musculoskeletal injury that results from a work-related event. This may result in lost work time, work restriction, or transfer to another job \[[@B1]-[@B4]\]. These types of injuries are common among physiotherapists \[[@B1],[@B2],[@B5]-[@B9]\]. This group has a moderately high prevalence of occupational low-back pain \[[@B2],[@B9]-[@B12]\]. Physical therapists routinely perform manual therapy, such as soft-tissue mobilization, which means that the upper limb is also exposed to risk factors associated with musculoskeletal and neurovascular disorders \[[@B1],[@B2],[@B6]-[@B8],[@B13],[@B14]\]. In addition, these professionals routinely perform activities that involve transferring a patient (from exercise mat to chair, to parallel bar etc), assisting with activities on the exercise mat, and lifting and using cumbersome equipment \[[@B1],[@B2],[@B6]-[@B15]\]. These work tasks put therapists at risk for both acute and cumulative musculoskeletal pain. Understanding the issues related to musculoskeletal disorders in physical therapists requires some awareness of the context in which these professionals work. In Turkey, physiotherapists cannot treat patients directly. According to law, direct accesses to physiotherapists are not possible \[[@B16]\]. Further, physiotherapists are not allowed to plan patients\' physiotherapy programs; this is done by a physiatrist. Currently, efforts are underway to enact legal amendments that will permit a doctor to direct a patient to a physiotherapist, with the physiotherapist creating the physiotherapy program him- or herself. The purpose of this study was to investigate the prevalence and features of WRMDs in a group of Turkish physiotherapists working in various capacities in a large Turkish city, and to compare the findings to reports from other countries. Methods ======= The study was approved by the Ethics Committee of Dokuz Eylül University. The research population was selected physiotherapists who were employed in a broad spectrum of practice settings in İzmir, Turkey. All data were collected by a mailed questionnaire (Appendix 1 \[see [Additional file 1](#S1){ref-type="supplementary-material"}\]) that consisted of 17 close-ended questions. Questionnaires were distributed to all physiotherapists (205 total, 157 females and 48 males) who were registered members of the Izmir Branch of the Turkish Physiotherapy Association. Each member was asked to complete the self-administered questionnaire if they had more than 2 years of experience in practice. One hundred and twenty questionnaires were returned and analyzed (total response rate 59%, for women 59%, for men 58%). The questionnaire was based on another published survey \[[@B1],[@B6]\] and simply adapted and translated for the Turkish context. The questionnaire was composed of two parts, personal and occupational. The personal portion asked about general characteristics, including sex, age, weight, and height. The occupational portion inquired about years of experience, work setting, and number of hours of contact with patients per week. This section also asked whether the subject had experienced any WRMDs. If the answer was yes, the person was asked to state the type of injury, the body part affected, specific activities caused on occupational injury, the work setting in which the injury occurred, whether the injury was reported or a physician was consulted, and what sort of treatment was applied. They were also asked whether they lost work time as a result of the injury, what activities caused symptoms to recur, and whether the injury had caused the respondent to alter his or her work habits, reduce hours with patients, or change employment settings. Data were analyzed using SPSS 10.0 for Windows. Results for the general information items were expressed as mean ± standard deviation, and results for items in the occupational portion were expressed as percentages. χ^2^were used to analyze influence personal characteristics (sex, age, number of years in physiotherapy practice, number of hours per week in direct patient care) to WRMDs. Results ======= The 120 respondents included 92 females and 28 males of mean age 30.4 ± 6.9 years (range, 22--55 years). Total response rate was 59%, for women 59%, for men 58%. General information about the group is given in Table [1](#T1){ref-type="table"}. The questionnaire answers indicated that the respondents spent an average of 38.8 ± 9.2 hours per week in direct patient care, and had 8.0 ± 6.0 years of work experience on average. Eighty-six of the physiotherapists worked in general hospitals, 24 worked at hydrotherapy centers, and 10 worked at pediatric rehabilitation centers. Frequency of WRMDs was not correlated with sex (χ^2^=.234, P=.629), age (χ^2^=.383, P=.536), number of years in physiotherapy practice (χ^2^=.067, P=.795), or number of hours per week in direct patient care (χ^2^=.151, P=.698) (Table [2](#T2){ref-type="table"}). One hundred and two respondents (85%) reported that they had experienced work-related musculoskeletal pain or discomfort at some time in their occupational lives. Seventy-eight (65%) reported that they had sustained more than one WRMD. The lower back was the body part with the highest frequency of occupational injury (26%), and wrist-hand (18%), shoulder (14%), and neck (12%) were other sites frequently affected (Table [3](#T3){ref-type="table"}). The main types of injury reported were tendinitis (21%), vertebral disk problems (16%), muscle strain (16%), ligament sprain (16%), degeneration (15%), synovitis (6%), tear (2%), dislocation (1%), fracture (1%) and other (6%). The factors that most frequently led to WRMD were transferring patients (15%), performing repetitive tasks (14%), and lifting (14%) (Table [4](#T4){ref-type="table"}). The respondents who had experienced a WRMD indicated that lifting (18%), maintaining a position for prolonged period of time (17%), performing repetitive tasks (16%), and transferring patients (16%) were the activities that most often exacerbated their symptoms during clinical practice (Table [5](#T5){ref-type="table"}). Improvements in body mechanics (21%), avoiding lifting (16%), and changing working positions frequently (14%) were the top three changes that the injured respondents said they made in their working habits (Table [6](#T6){ref-type="table"}). Sixty-nine percent of the respondents with WRMDs had visited a physician for the problem, and 46% of whom stated that they had officially reported the injury to their employer. The respondents who had suffered WRMDs reported they used their own occupational knowledge, rest, medications, and exercise to treat the problem (Table [7](#T7){ref-type="table"}). Sixty-seven percent of the respondents who had had WRMDs indicated that they had not permanently reduced their patient-contact time as a result of their injury, and 82% said they did not limit their areas of practice after sustaining the injury. Most of the physiotherapists (63%) who had had WRMDs indicated that they would not consider a job change because of their injury or due to the risk of sustaining another injury. Discussion ========== Musculoskeletal system problems connected to occupational conditions are common among health care workers. The costs of these are substantial, both in terms of money and in terms of work time lost \[[@B1],[@B7],[@B9],[@B17],[@B18]\]. Research has shown that musculoskeletal problems are particularly common in health care workers who are in direct contact with patients \[[@B1],[@B7],[@B8],[@B17],[@B18]\]. Physiotherapists have a high prevalence of WRMDs \[[@B1],[@B7],[@B8],[@B15]\]. The results from studies on WRMDs in physiotherapists have generally been similar, though some have differed according to country. Such variations are linked to level of development, the status of the profession of physiotherapy in a given country, psychosocial, and epidemiological factors \[[@B2],[@B5],[@B8],[@B11],[@B13],[@B14]\]. In Turkey the law bars physiotherapists from providing primary care; each patient must be referred. Considering the differences in physiotherapy practice among countries and regions, we felt it would be valuable to investigate the prevalence and features of WRMDs in a group of Turkish physiotherapists working in various capacities, and to compare to findings in other countries. In this study, we collected demographic and WRMD data from 120 physiotherapists in Izmir, Turkey, and analyzed rates of injury, risk factors, injury types and sites, and post-injury management. We asked to complete the self-administered questionnaire if they had more than 2 years of experience in practice. Thus, response rate of questionnaire was low (58.5 %) in our study. The survey answers revealed that 85% of the respondents had experienced WRMDs. Cromie \[[@B8]\] reported that younger physiotherapists have a higher prevalence of musculoskeletal problems related to occupational conditions. Rugelj \[[@B9]\] investigated low-back pain in physiotherapists, and found an incidence of 66% in subjects between the ages of 20 and 40 years \[[@B1]\]. In line with this, a study of Australian physiotherapy students by Nyland \[[@B2]\] revealed that the 20- to 21-year age group had the highest frequency of low-back pain. The average age of the physiotherapists in our study was 30.4 years. This mean age corresponds with other findings in the literature, and confirms that physiotherapists tend to experience WRMDs at young age. Such injuries in younger physiotherapists may be associated with lack of professional experience, and the lower knowledge and skill levels people tend to have in the early years of this career. Concerning sites of musculoskeletal injury during professional activities, the highest incidence is in the low-back region. Biomechanical studies have shown that physical loading factors, such as body flexion, rotation and weight loading, play a role in this \[[@B18]\]. In a study that covered 25% of all physiotherapists working in Australia (824 total studied), Cromie \[[@B8]\] found that the rate of work-related low-back pain was 48%. Other authors have revealed various rates of this problem in physiotherapists: Bork \[[@B7]\] 45%, Holder \[[@B1]\] 62%, Molumphy \[[@B10]\] 29%, Scholey and Hair \[[@B12]\] 38%, Mierzejewski \[[@B11]\] 49.2%, Rugelj \[[@B9]\] 73.7%, and Nyland \[[@B2]\] 69%. Our survey of Turkish physiotherapists revealed a 25.5% incidence of low-back pain. Interestingly, this figure is lower than most other rates reported in the literature. In our study, when we categorized physiotherapists according to practice setting, low-back pain was the most common WRMD in all subgroups. Studies of WRMDs in health care of professionals have identified the lower back as the most commonly involved area of the body, followed by neck and upper extremities \[[@B3],[@B13],[@B19]-[@B21]\]. Investigations of physiotherapists have revealed similar results. Bork \[[@B7]\] and Holder and et al. \[[@B1]\] listed the regions most commonly involved musculoskeletal injuries as the lower back, hand-wrist, and neck, respectively among physiotherapists. On analyzing the different body parts injured in our subjects, we found the highest frequency of injuries in the lower back, followed by the hand-wrist (18.2%), shoulder (14.4%) and neck (11.8%). According to the literature the work-related activities that most commonly lead to injury in health professionals are lifting heavy equipment and patients, transferring patient, maintaining the same posture for a long period, manual therapy practices, responding to patients\' sudden movements, and repeated movements \[[@B1],[@B8],[@B17],[@B18],[@B22]-[@B24]\]. Bork \[[@B7]\] identified the main causes of WRMDs in physiotherapists as staying in the same position for along time and continuing to work when tired; Molumphy \[[@B10]\] emphasized lifting and leaning downwards with sudden maximal effort; and West \[[@B13]\] highlighted maintaining the same posture for along time, manual therapy, repeated movements, and increased work load. In our survey, the main causes of WRMDs were patient transfer, repeated movements, lifting heavy equipment, patients and working when too physically tired. Holder et al. \[[@B1]\] identified three activities that aggravate the symptoms of former WRMDs in physiotherapists and assistant physiotherapists as staying in the same position for a long period, lifting and patient transfer. In our study, the respondents to our questionnaire noted lifting, staying in the same position for a long time, patient transfers, and repeated movements. In Turkey, most physiotherapists work in general hospitals. In these facilities, the daily treated patient number is too high, and this number far exceeds the number of physiotherapists on staff. In addition, the majority of patients in these hospitals are seriously ill \[[@B25]\]. People who suffer injuries on the job may be treated with medication, rest and exercise. Physiotherapists have fundamental knowledge about ergonomics and biomechanics, and using this knowledge may vary depending on professional knowledge and skills. In Turkey, physiotherapists are trained in ergonomic working principles at the undergraduate level. However, in many working environments the equipment used (treatment table, chair, *etc*.) is not ergonomic, and consequently the physiotherapist cannot follow ergonomic principles while doing his or her job. A study conducted by Cromie \[[@B6]\] in 2002 examined whether physiotherapists use their own knowledge to prevent WRMDs. The author found that this was true for most of the physiotherapists investigated. In our survey, of the physiotherapists who had suffered WRMDs, 27.6% said they used their professional knowledge and 26% said they used rest to manage the injury. Previous research has shown that physiotherapists who have suffered a WRMD tend to change their professional attitudes to avoid additional injuries. Holder *et al.*\[[@B1]\] found that 79% of physiotherapists and 81% of assistant physiotherapists who had suffered injuries on the job changed their professional attitudes to avoid other WRMDs. It has been reported that the most common strategies used by physiotherapists to avoid WRMDs are correction of body mechanics, and frequent postural changes. The respondents to our questionnaire said that, after suffering a WRMD, they paid more attention to correcting body mechanics, avoided lifting heavy equipment or patients, changed position frequently, and got other personnel to help them with tasks that involved lifting. Mierzejewski \[[@B11]\] reported that 13.7% of physiotherapists in Edmonton, gave up their career after suffering work-related low-back pain, and that 35.3% continued to work after this problem. Cromie \[[@B8]\] found that 17.7% of physiotherapists changed their field of practice in connection with WRMDs. Of the respondents in our Turkish survey who had suffered WRMDs, 67% did not change their field of practice after the injury, 82% did not restrict their time with patients, and 64% said they would not consider changing their field or department of practice due to their WRMD. This means that 33% did change field of practice after a WRMD, which is 1.9 times the rate of those who changed field in Cromie\'s study. As noted above, most physiotherapists in Turkey work in general hospitals, and it is not usually possible for such physiotherapists to change their field of work. Some professions have very high WRMD prevalence, and this has led to more intensive research in recent years. These types of injury have major consequences for society, workers, employers, and the insurance sector due to loss of labor force, long-term disability and delay in returning to work, decreased productivity, and psychological effects on employees. Therefore, minimizing and preventing WRMDs has significant potential social and economic benefits. It is important to invest in studies aimed at reducing these types of injuries in physiotherapists. Our study has three main limitations. A physiotherapist\'s working time in a day is not long, but he or she treats a very large number of patients during this time. Our survey would have been more informative if we collected data for the number of daily treated patients. We also did not inquire about the physiotherapists\' activity levels. This would have been valuable information, as sports and recreational activity may affect WRMD frequency. Another limitation of our study was comparing various areas of practice and techniques used by physiotherapists and investigating their relationship to WRMDs. Conclusions =========== Our survey reveals that the WRMDs in physiotherapists in Turkey are similar to rates reported in other countries. Physiotherapists in our country suffer similar work-related injuries as their counterparts elsewhere, despite different legal working conditions and cultural differences. This study provides data related to WRMDs in physiotherapists in Turkey. Further studies can be very useful if it research prevalence of WRMDs in physiotherapists who have employed different working conditions. Competing interests =================== None declared. Authors\' Contributions ======================= YS participated in the design, managed the data collection and performed statistical analysis. AO participated in designed the study protocol, managed the coordination. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2474/5/27/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional File 1 Table 1, Table 2. Questionnaire on occupational injuries in physical therapists. Turkish version of questionnaire on occupational injuries in physical therapists ::: ::: {.caption} ###### Click here for file ::: Figures and Tables ================== ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Descriptive information of physical therapist (PT) respondents to questionnaire ::: **Female (n = 92)** **Male (n = 28)** **All** --------------------------------------- ----------- --------------------- ------------------- -------------- Age (yrs) Mean (SD) 30.3 (7.5) 30.8 (4.4) 30.4 (6.9) Range 22 -- 55 25 -- 38 22 -- 55 Height (m) Mean (SD) 1.63 (6.21) 1.76 (7.3) 1.67 (8.41) Range 1.32 -- 1.78 1.65 -- 1.90 1.32 -- 1.90 Weight (kg) Mean (SD) 57.9 (6.4) 78.1 (13.1) 62.6 (12.0) Range 45 -- 82 53 -- 100 45 -- 100 Years as a PT Mean (SD) 8.2 (6.7) 7.9 (3.6) 8.0 (6.0) Range 2 -- 29 2 -- 14 2 -- 29 Hours per week in direct patient care Mean (SD) 38.5 (9.0) 40.0 (9.8) 38.8 (9.2) Range 8 -- 60 5 -- 60 5 -- 60 ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Distributions of the physical therapists (PT) with WRMDs according to sex, age, years as a pt, and hours per week in direct patient care ::: WRMD (n) \% --------------------------------------- -------- ---------- ------ Sex Female 79 85.9 *Male* 23 82.1 Age (yrs) \<*30* 59 86.8 ≥30 43 82.7 Years as a PT \<*8* 60 85.7 ≥8 42 84.0 Hours per week in direct patient care \<*38* 24 82.8 ≥*38* 78 85.7 ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### The frequencies of WRMDs at body part affected in different practice settings. ::: Orthopedic Rehabilitation % General Physical Therapy % Neurological Rehabilitation % Cardiopulmonary Rehabilitation % ------------------------- ----------------------------- ---------------------------- ------------------------------- ---------------------------------- **Neck**n:30 12 % 9.1 10.9 15.1 6.7 **Shoulder**n:39 14 % 11.4 16.9 15.1 16.7 **Upper Back**n:15 9 % 6.8 5.0 5.8 6.6 **Elbow**n:22 8 % 13.6 7.9 8.1 3.3 **Lower Back**n:72 26 % 29.5 25.7 28.0 30.0 **Wrist/Hand**n:48 18 % 18.2 22.8 12.8 16.7 **Hip/Thigh**n:6 2 % 2.3 3.0 2.3 0.0 **Knee**n:20 8 % 6.8 5.9 9.3 10.0 **Ankle/Foot**n:9 3 % 2.3 2.0 3.5 10.0 ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Proportions of respondents who reported that specific activities caused on occupational injury ::: **Activity that caused injury** **PT (%)** ---------------------------------------------------------------- ------------ Transferring a patient 14.5 Performing repetitive tasks 13.9 Lifting heavy equipment or patients 13.6 Working when physically fatigued 12.0 Bending/twisting 11.1 Performing manual therapy 9.0 Maintaining a position for a prolonged period of the time 9.2 Responding to an unanticipated or sudden movement a by patient 5.9 Working in an awkward or cramped position 5.0 Slipping. tripping. falling 2.7 Applying modalities 1.5 ::: ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Proportions of respondents who reported that job activities caused their symptoms to recur ::: **Exacerbating activity** **PT (%)** ------------------------------------------------------- ------------ Lifting heavy equipment or patients 18.1 Maintaining a position for a prolonged period of time 16.6 Transferring a patient 15.5 Performing repetitive tasks 15.8 Performing manual therapy 10.8 Squatting 5.0 Working in an awkward or cramped position 4.3 Climbing stairs 3.9 Reaching/working or cramped positions 3.5 Walking 3.5 Performing overhead activities 1.9 ::: ::: {#T6 .table-wrap} Table 6 ::: {.caption} ###### Proportions of respondents who reported altering work habits as a result of experiencing an occupational musculoskeletal injury ::: **Altered work habits** **PT (%)** ------------------------------------------------------------- ------------ Use improved body mechanics 20.5 Avoid lifting 16.4 Change working position frequently 13.7 Increase use of other personnel 10.5 Decrease manual therapy 6.9 Increase use of mechanical aids 7.3 Increase administrative time. decrease patient care time 7.3 Encourage patient responsibility for carrying out treatment 6.4 Stop working when hurt or when symptoms occur 5.5 Change work schedule 3.2 Take more rest breaks or pauses during the workday 2.3 ::: ::: {#T7 .table-wrap} Table 7 ::: {.caption} ###### Types of treatments and response to musculoskeletal injury of respondents ::: **n** **%** ------------------------ ------------------------------- ------- ------- **Treatment** Use of occupational knowledge 70 27.5 Rest 66 26.0 Medication 65 25.6 Exercise 50 19.7 Surgery 3 1.2 **Response to injury** Visiting a physician 70 68.6 Reporting officially 45 44.1 :::

PubMed Central

2024-06-05T03:55:47.465412

2004-8-18

PMC516039

Background ========== A systematic review of topical NSAIDs reported that they were effective for relieving pain in both acute and chronic conditions \[[@B1]\]. Number-needed-to-treat (NNT), the number of patients that need to be treated for one to benefit from a particular drug, who would not have benefited from placebo, was used to estimate efficacy. In chronic conditions, NNT for topical NSAIDs at two weeks was 3.1 (2.7 to 3.8). There are three reasons why an updated review of topical NSAIDs in chronic pain is needed. First, we have a better appreciation of factors that can introduce bias \[[@B2]-[@B4]\], and would not now accept trials that were not double blind, or were very small. Second, topical salicylate and benzydamine are no longer classed as topical NSAIDs \[[@B5]\]. Thirdly, there are now more trials. We believed that updating the review would improve efficacy estimates for topical NSAIDs, with a prior intent to determine efficacy for individual drugs. Methods ======= Searching --------- Relevant studies were sought regardless of publication language, type, date or status. Studies included in the previous review were considered for inclusion, and the Cochrane Library, MEDLINE and PreMedline, EMBASE and PubMed, were searched for relevant studies published since the last review, for the years 1996 to April 2003. The search strategy included \"application: topical\" together with \"cream\", \"gel\" etc, together with generic names of NSAIDs, and proprietary preparations of topical treatment in which the principal active ingredient was an NSAID \[[@B6],[@B7]\] ([additional file 1](#S1){ref-type="supplementary-material"}:search strategy). Reference lists of retrieved articles were also searched. We wrote to 20 pharmaceutical companies in the UK, 66 in continental Europe, and two in North America, known to manufacture topical NSAIDs, asking if they could supply papers. Selection --------- We identified reports of randomised, double blind, active or placebo-controlled trials in which treatments were given to adult patients with moderate to severe chronic pain resulting from musculoskeletal or other painful disorders. We excluded treatments for mouth or eye diseases. At least ten patients had to be randomised to a treatment group and application of treatment had to be at least once daily. Outcomes closest to two weeks (but at least seven days) were extracted. Longer outcomes were also accepted when available. Quality and validity assessment ------------------------------- Trial quality was assessed using a validated three-item scale with a maximum quality score of five \[[@B8]\]. Included studies had to score at least two points, one for randomisation and one for blinding. A sixteen-point scale was used to assess trial validity \[[@B9]\]. Quality and validity assessments were made independently by at least two reviewers and verified by one other reviewer. Disputes were settled by discussion between all reviewers. Outcomes -------- We defined our own outcome of clinical success, representing approximately a 50% reduction in pain \[[@B1]\]. This was either the number of patients with a \"good\" or \"excellent\" global assessment of treatment, or \"none\" or \"slight\" pain on rest or movement (or comparable wording) measured on a categorical scale. A hierarchy of outcomes was used to extract efficacy information \[[@B1]\], shown below in order of preference: 1\) number of patients with a 50% or more reduction in pain 2\) patient reported global assessment of treatment 3\) pain on movement 4\) pain on rest or spontaneous pain 5\) physician or investigator global assessment of treatment In addition, the number of patients showing undefined \"improvement\" was also accepted. All of these outcomes were grouped together as a \"success\", and categories 1--4 were used as preferred outcomes in the sensitivity analysis. Secondary outcomes were extracted from included papers that reported them. These were the number of patients (i) reporting one or more local adverse event (itching, stinging, rash), (ii) reporting one or more systemic adverse event (iii) withdrawing from trials due to adverse events. Quantitative data synthesis --------------------------- The number of patients randomised into each treatment group (intention to treat) was used in the efficacy analysis. Information was pooled for the number of patients in each trial approximating at least 50% pain relief, or similar measure, for both topical NSAID and control. These were used to calculate NNT with a 95% confidence interval (CI) \[[@B10]\]. Relative benefit and relative risk estimates with 95% CIs were calculated using the fixed effects model \[[@B11]\]. A statistically significant benefit of topical NSAID over control was assumed when the lower limit of the 95% confidence interval (CI) of the relative benefit was greater than one. A statistically significant benefit of control over active treatment was assumed when the upper limit of the 95% CI was less than one. Homogeneity of trials was assessed visually \[[@B12]-[@B14]\]. Number-needed-to-harm (NNH) and relative risk were calculated in the same way as for NNTs and relative benefit. All calculations were performed using Microsoft Excel X for the Macintosh and RevMan 4.2. In sensitivity analyses the z test was used \[[@B15]\]. QUOROM guidelines were followed \[[@B16]\]. Sensitivity analysis -------------------- Our prior intention was to perform sensitivity analyses on pooled outcomes using the z test \[[@B15]\] for quality score (2 versus 3 or more), validity score (8 or less versus 9 or more), trial size (less than 40 patients per group versus more than 40 patients per group), reported outcome (higher versus lower preference), drug, and condition treated (knee osteoarthritis versus other musculoskeletal). At least three studies had to be available in each category before information was pooled. Results ======= Study characteristics --------------------- Ten out of the 20 UK companies, and two out of the 66 continental European companies replied to our request for studies. Only three companies supplied useful material, either published studies or bibliographies. None provided unpublished material. Searches identified 60 target papers, but 35 were excluded; 23 studies failed to meet the inclusion criteria and 12 had no useable data. Twenty-four of these 60 target papers were included in the previous review. We included 13 of those in this review, and excluded 11; seven were not double blind, two compared a salicylate with placebo or oral analgesics, one did not have daily application, and one had insufficient data ([additional file 2](#S2){ref-type="supplementary-material"}: excluded studies, [additional file 3](#S3){ref-type="supplementary-material"}: QUOROM flow diagram). Twenty-five trials therefore met the selection criteria, 12 of which were additional trials. Fifteen trials had only placebo controls \[[@B17]-[@B31]\], seven only active controls \[[@B32]-[@B38]\], and three had both placebo and active controls \[[@B39]-[@B41]\]. Of the 10 active controlled trials, four compared a topical NSAID with a different topical NSAID, three compared a topical NSAID with a different oral NSAID, and one each compared a topical NSAID with a homeopathic gel, a topical rubefacient, and topical trinitroglycerin (GTN). Details of all included studies with outcomes and quality and validity scores are in [additional files 4](#S4){ref-type="supplementary-material"} (Outcome details of placebo-controlled trials) and [additional files 5](#S5){ref-type="supplementary-material"} (Outcome details of active-controlled trials). Patients were generally over 40 years old, predominantly with musculoskeletal disorders, and with baseline pain of moderate to severe intensity. Fourteen studies examined general musculoskeletal conditions, and eleven examined osteoarthritis (9 studies of the knee, one of finger joints, and one of mixed sites). Five studies in osteoarthritis specified use of a standard scale (ACR, Kellgren and Lawrence, ISK) to assess the severity of disease, four specified that the disease was radiologically confirmed, one specified that patients had \"well documented mild osteoarthritis\", and one made no statement. Quality scores were high, with 16/18 placebo controlled and 9/10 active controlled trials scoring 3 or more points out of a maximum of 5. Validity scores were also high, with 14/18 placebo controlled and 8/10 active controlled trials scoring 9 or more out of a maximum of 16 ([additional files 4](#S4){ref-type="supplementary-material"} and [5](#S5){ref-type="supplementary-material"}). Placebo controlled trials ------------------------- Dichotomous information was available to pool from 14 placebo controlled trials for efficacy, from 16 placebo controlled trials for local adverse events, 17 placebo controlled trials for systemic adverse events, and from 11 placebo controlled trials for adverse event related withdrawals. ### Efficacy Fourteen trials (1,502 patients) provided data on efficacy. Topical NSAIDs were significantly better than placebo (Table [1](#T1){ref-type="table"}). The mean placebo response rate was 26% ranging from 7% to 78%. The mean treatment response rate was 48% ranging from 2% to 90% (Figure [1](#F1){ref-type="fig"}). The NNT was 4.6 (95% CI 3.8 to 5.9) for one patient to experience improvement in chronic musculoskeletal pain at two weeks with topical NSAIDs, compared with placebo. Sensitivity analyses (Table [1](#T1){ref-type="table"}) showed no significantly greater effect for low quality trials (quality score 2/5) compared with higher quality trials (quality score 3--5/5) (z = 1.69, p = 0.091). There was no significant difference for smaller versus larger trials using 50 patients per group (median group size for topical NSAID was 49) as a cut off (z = 0.40, p = 0.69), for preferred outcomes versus lower preference outcomes (physician determined or general improvement) (z = 1.56, p = 0.12), or for patients with knee osteoarthritis compared with other musculoskeletal conditions (z = 0.99, p = 0.32) (Figure [2](#F2){ref-type="fig"}). The 10 trials with both a quality score of 3/5 or greater and a validity score of 9/16 or greater had an NNT of 4.4 (95% CI 3.6 to 5.6). There were insufficient data to allow comparisons of efficacy between different NSAIDs. ### Harm All 18 placebo controlled trials (2,032 patients) provided some information on adverse events (Table [2](#T2){ref-type="table"}). There was no statistically significant difference between topical NSAID and topical placebo for the number of patients experiencing local adverse events (6%), systemic adverse events (3%), or the number withdrawing due to an adverse event (1%). With topical NSAID or topical placebo, local adverse events were usually described as rash, itching or stinging, and were predominantly mild. Active controlled trials ------------------------ ### Efficacy There was sufficient information to pool results only from the three trials comparing a topical NSAID with an oral NSAID in patients with osteoarthritis of the knee or finger joints. One trial \[[@B34]\] compared piroxicam 0.5% gel with oral ibuprofen 1200 mg daily, another \[[@B38]\] compared diclofenac 1% gel with oral ibuprofen 1200 mg daily, and the third \[[@B41]\] compared eltenac 1% gel with oral diclofenac 100 mg daily. In these trials, with 764 patients, 37% had a successful outcome both with topical NSAID and oral NSAID. There was no statistically significant difference (relative risk 1.1; 95% CI 0.9 to 1.3). The other seven studies used different topical preparations and different comparators in small trials ([additional file 5](#S5){ref-type="supplementary-material"}: Outcome details of active-controlled trials). ### Harm Eight of the active controlled trials (1,461 patients) provided some information on adverse events (Table [2](#T2){ref-type="table"}). In two active controlled trials comparing topical with oral NSAID, local adverse events occurred more frequently (8%) with topical than with oral NSAID (3%). Systemic adverse events and adverse event withdrawals did not differ between topical and oral NSAID. No study documented specific instances of upper gastrointestinal bleeding or symptomatic ulcers. Discussion ========== Patients in these trials all had moderate to severe baseline pain, and for those with osteoarthritis, disease severity was generally mild to moderate. Patients with most severe disease were specifically excluded in several trials because authors regarded topical NSAID to be inappropriate for their treatment. Both the original and this updated review concluded that topical NSAIDs were effective in chronic conditions. However, removing trials of lower quality, and topical agents that are not now regarded as topical NSAIDs, increased (worsened) the NNT from 3.1 (95% CI 2.7 to 3.8) to 4.7 (95% CI 3.8 to 5.9) for the outcome of at least half pain relief at two weeks for all topical NSAIDs compared to placebo. For every four or five patients with chronic pain treated with topical NSAID, one would benefit who would not have done with placebo. Three trials comparing topical with oral NSAID found no difference in efficacy. There are a number of aspects of this review that might question this demonstration of efficacy. The trials spanned several decades and retrospective examination finds fault with them in several respects. Many trials were small, and small size can allow chance effects to influence treatment and placebo event rates \[[@B4]\]. Different preparations were used, with different formulations, concentrations of active ingredient, and application schedules. Reported outcomes were not consistent, and a hierarchy of outcomes had to be constructed. It was inevitable that there would be some clinical heterogeneity, even when similar patients were treated, and when trials were both randomised and double blind, and of appropriate duration. We addressed these limitations with pre-planned sensitivity analyses. Using studies with higher quality and validity scores, larger size, or higher rather than lower preference outcomes made no difference. Patients treated for knee osteoarthritis derived the same degree of pain relief as those treated for general musculoskeletal conditions. The evidence was that topical NSAIDs were effective whatever strategy was used for sensitivity analysis, improving the robustness of the overall result. A possible criticism might be that there has been selective publication of trials showing topical NSAIDs to be effective, and suppression of trials where there was no difference between topical NSAID and placebo. Funnel plots do not reliably detect publication bias \[[@B13],[@B14]\], so we did not use them or make any adjustment for possible publication bias \[[@B42]\]. We did approach every company in the world that we could identify as being involved with topical NSAID manufacture or sale for any additional unpublished trials, but no more unpublished material was identified. When unpublished material is found, it often does not change the relevance of a result \[[@B43]-[@B45]\]. It is important to emphasise that both active and placebo treatments were rubbed on, making any effect of rubbing equal in both groups. The mean placebo response in the included trials was 26%, compared with the mean response of 48% with topical NSAID. The response with placebo is consistent with that found in acute and chronic pain with a variety of conditions and endpoints \[[@B46]\]. Local adverse events were reported with equal frequency for topical NSAID and topical placebo in placebo-controlled trials, but more often for topical NSAID than oral NSAID in active controlled trials. There were no differences between topical NSAID and topical placebo, or topical NSAID and oral NSAID, for systemic adverse events, or withdrawals due to adverse events. Studies of short duration will not capture important long-term safety information, and this may be important for ongoing applications of gels, creams or sprays in chronic conditions. There is, however, information that indicates that topical NSAIDs do not cause the gastrointestinal harm found with oral NSAIDs \[[@B47]\], nor are they associated with increased renal failure \[[@B48]\]. Clearly there is a body of evidence to support the efficacy of topical NSAIDs in chronic painful musculoskeletal conditions. Despite removing smaller studies that were not double blind, and substituting newer, larger trials of higher quality, topical NSAIDs remained effective, though the NNT was higher (worse) than originally estimated \[[@B1]\]. More information of high quality is required, to compare the relative efficacy of topical and oral NSAIDs, and between different topical NSAIDs. We are able to compare the evidence for different topical analgesics in chronic musculoskeletal pain (Table [3](#T3){ref-type="table"}). Systematic reviews of topical salicylate \[[@B49]\] and capsaicin \[[@B50]\], tell us what is known about those treatments. As Table [3](#T3){ref-type="table"} shows, topical NSAIDs have been tested in many more studies, and in four times as many patients as these other topical analgesics, and have the lowest (best) NNT. The limitation of this comparison is essentially the same limitation as with all these reviews, that the included trials were too short and too small to be sure of the result. Topical NSAIDs have the best evidence for chronic musculoskeletal pain, supporting the excellent evidence available in acute painful conditions \[[@B51]\]. Authors\' contributions ======================= LM was involved with planning the study, searching, data extraction, analysis, and preparing a manuscript; RAM with planning, data extraction, analysis and writing the manuscript; JE with searching, data extraction, and writing; SD with data extraction, analysis, and writing; HJM with planning, analysis and writing. All authors read and approved the final manuscript. Competing interests =================== RAM & HJM have consulted for various pharmaceutical companies. RAM, HJM & JE have received lecture fees from pharmaceutical companies related to analgesics and other healthcare interventions. All authors have received research support from charities, government and industry sources at various times, but no such support was received for this work. No author has any direct stock holding in any pharmaceutical company. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2474/5/28/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional File 1 Search strategy for RCTs of topical NSAIDs in chronic pain ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 Excluded studies ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 3 QUOROM flow diagram ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 4 Outcome details of placebo-controlled trials ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 5 Outcome details of active-controlled trials ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ The study was supported with Pain Research funds and by the Oxford Pain Relief Trust. Figures and Tables ================== ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Topical NSAIDs in chronic musculoskeletal pain Randomised double-blind studies of topical NSAID compared to topical placebo for two-week outcome of successful treatment. Inset scale shows size of individual trials. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Analysis of trials of topical NSAIDs in chronic musculoskeletal pain by condition. This Forrest plot was created using RevMan 4.2. Details of the statistical tests used can be found in the Cochrane Handbook. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Summary data and sensitivity analyses for placebo controlled trials ::: **Number of** **Success/total** ------------------------------------------ --------------- ------------------- --------- --------- ------------------ ------------------ **All trials** 14 1502 371/771 193/731 1.9 (1.7 to 2.2) 4.6 (3.8 to 5.9) **Quality and validity** Quality score 3 to 5 11 1312 294/678 144/634 2.0 (1.7 to 2.4) 4.8 (3.9 to 6.4) Quality score 2 3 190 77/93 49/97 1.6 (1.3 to 2.0) 3.1 (2.2 to 5.1) Validity score ≥ 9 and quality score ≥ 3 10 1197 247/622 98/575 2.4 (2.0 to 3.0) 4.4 (3.6 to 5.6) **Trial group size** ≤ 50 patients 6 343 95/172 54/171 1.8 (1.4 to 2.3) 4.2 (3.0 to 7.4) \> 50 patients 8 1159 276/599 139/560 2.0 (1.7 to 2.3) 4.7 (3.8 to 6.3) **Outcome type** preferred outcomes 8 920 198/462 85/458 2.3 (1.9 to 2.9) 4.1 (3.3 to 5.4) lower preference outcomes 6 582 173/309 108/273 1.6 (1.4 to 1.9) 6.1 (4.1 to 12) **Condition** knee osteoarthritis 5 567 127/307 58/260 2.0 (1.6 to 2.6) 5.3 (3.8 to 8.6) other musculoskeletal 9 935 244/464 135/471 1.9 (1.6 to 2.2) 4.2 (3.3 to 5.6) ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Placebo contolled trials ::: **Number of** **Events/total** ----------------------------------------------- --------------- ------------------ --------------- ------------- -------------------- Local adverse events 15 1734 53/949 48/785 1.0 (0.7 to 1.4) Systemic adverse events 16 1838 33/1002 14/836 1.7 (0.96 to 2.85) Withdrawals beacuse of adverse events 10 1225 10/697 7/528 0.9 (0.4 to 2.1) **Active controlled trials: topical vs oral** **Number of** **Events/total** **Type of adverse event** **Trials** **Patients** **Treatment** **Placebo** **RR (95% CI)** Local adverse events 2 443 19/243 4/118 3.0 (1.1 to 8.5) Systemic adverse events 3 764 82/408 87/356 0.83 (0.6 to 1.1) Withdrawals because of adverse events 3 764 19/408 24/356 0.7 (0.4 to 1.3) ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Comparison of topical analgesics in chronic musculoskeletal pain ::: **Number of** **Percent success with** ------------- --------------- -------------------------- ---- ---- ------------------ NSAID 14 1502 48 26 4.6 (3.8 to 5.9) Salicylates 6 429 54 36 5.3 (3.6 to 10) Capsaicin 3 368 38 25 8.1 (4.6 to 34) :::

PubMed Central

2024-06-05T03:55:47.467963

2004-8-19