Neurotoxicology and Teratology 80 (2020) 106890

Contents lists available at ScienceDirect

Neurotoxicology and Teratology

journal homepage: www.elsevier.com/locate/neutera

Regions of the basal ganglia and primary olfactory system are most sensitive to neurodegeneration after extended sevoﬂurane anesthesia in the perinatal rat

☆

T

Susan M. Burks, John F. Bowyer, Jennifer L. Walters, John C. Talpos

⁎

National Center for Toxicological Research, 3900 NCTR Rd, Jeﬀerson, AR 72079, United States of America

A R T I C L E I N F O

A B S T R A C T

Keywords: Neurotoxicity Development Fluoro-Jade Hypoxia Indusium griseum

Extended general anesthesia early in life is neurotoxic in multiple species. However, little is known about the temporal progression of neurodegeneration after general anesthesia. It is also unknown if a reduction in natural cell death, or an increase in cell creation, occurs as a form of compensation after perinatal anesthesia exposure. The goal of this study was to evaluate markers of neurodegeneration and cellular division at 2, 24, or 72 h after sevoﬂurane (Sevo) exposure (6 h) in fully oxygenated postnatal day (PND) 7 rats. Neurodegeneration was ob- served in areas throughout the forebrain, while the largest changes (fold increase above vehicle) were observed in areas associated with either the primary olfactory learning pathways or the basal ganglia. These regions included the indusium griseum (IG, 25-fold), the posterior dorso medial hippocampal CA1 (17-fold), bed nucleus of the stria terminalis (Bed Nuclei STM, 5-fold), the shell of the nucleus accumbens (Acb, 5-fold), caudate/ putamen (CPu, 5-fold), globus pallidus (GP, 9-fold) and associated thalamic (11-fold) and cortical regions (5- fold). Sevo neurodegeneration was minimal or undetectable in the ventral tegmentum, substantia nigra, and most of the hypothalamus and frontal cortex. In most brain regions where neurodegeneration was increased 2 h post Sevo exposure, the levels returned to < 4-fold above control levels by 24 h. However, in the IG, CA1, GP, anterior thalamus, medial preoptic nucleus of the hypothalamus (MPO), anterior hypothalamic area (AHP), and the amygdaloid nuclei, neurodegeneration at 24 h was double or more than that at 2 h post exposure. Anesthesia exposure causes either a prolonged period of neurodegeneration in certain brain regions, or a distinct secondary degenerative event occurs after the initial insult. Moreover, regions most sensitive to Sevo neurodegeneration did not necessarily coincide with areas of new cell birth, and new cell birth was not consistently aﬀected by Sevo. The proﬁle of anesthesia related neurotoxicity changes with time, and multiple mechanisms of toxicity may exist in a time-dependent fashion.

1. Introduction

Early in life, the brain of mammalian species undergoes a period of rapid growth known as the brain growth spurt (BGS) (Dobbing and Sands, 1979; Workman et al., 2013). During this time, the brain is thought to be more vulnerable to toxic insults such as general an- esthesia exposure (Eriksson, 1997; Ikonomidou, 2009; Rice and Barone, 2000). Anesthesia related neurotoxicity in animal models of human neonatal brain development was ﬁrst established by Olney and collea- gues after ketamine or nitrous oxide exposure in the rat (Jevtovic- Todorovic et al., 2001). Since then, markers of neurotoxicity have been anesthetic. This reported with every FDA approved general

phenomenon has been observed in multiple species including rats (Jevtovic-Todorovic et al., 2000; Scallet et al., 2004), mice (Istaphanous et al., 2011; Zheng et al., 2013b), nonhuman primates (Brambrink et al., 2010; Paule et al., 2011; Slikker Jr. et al., 2007; Talpos et al., 2019), nematodes (Gentry et al., 2013; Na et al., 2017), and zebraﬁsh (Guo et al., 2015; Kanungo et al., 2013). The clinical relevance of these ﬁndings has been debated (Vutskits and Culley, 2019). However, sev- eral large scale retrospective studies have shown that multiple ex- posures to anesthesia within the ﬁrst two years of life is associated with an increased incidence of learning disabilities and attention-deﬁcit hyperactivity disorder (Flick et al., 2011; Sprung et al., 2012; Wilder et al., 2009), as well as an increased use of psychotropic drugs to treat a

☆

The information in these materials is not a formal dissemination of information by the Food and Drug Administration (FDA) and does not represent agency

position or policy.

⁎

Corresponding author. E-mail address: John.Talpos@fda.hhs.gov (J.C. Talpos).

https://doi.org/10.1016/j.ntt.2020.106890 Received 13 December 2019; Received in revised form 10 April 2020; Accepted 29 April 2020

Available online 12 May 2020 0892-0362/ Published by Elsevier Inc.

S.M. Burks, et al.

variety of conditions (Ing et al., 2020). It seems that early life exposure to general anesthesia increases the likelihood of individuals having cognitive diﬀerences during development (Ing and Brambrink, 2019). Perinatal anesthesia related neurotoxicity is well established in animal models of human use (Brambrink et al., 2012; Brambrink et al., 2010; Ikonomidou, 2009; Jevtovic-Todorovic et al., 2001; Paule et al., 2011; Slikker Jr. et al., 2007; Talpos et al., 2019; Walters and Paule, 2017). However, the temporal progression of this neurotoxicity has not been described. For example, the pattern of neurodegeneration (post- exposure) that develops over time and across brain regions has not been determined. Also, the temporal aspects of any “compensatory” eﬀects that involve increased birth of new cells in the days after insult are unknown. Addressing these knowledge gaps will help in understanding the mechanisms behind anesthesia related neurotoxicity and in de- termining the clinical relevance of animal models to human perinatal anesthesia exposure.

More than 100 laboratory animal studies have described the eﬀects of prolonged anesthetic exposure to neonates in various species. While several studies evaluate toxicity in multiple brain structures (Deng et al., 2014; Lee et al., 2017; Perez-Zoghbi et al., 2017; Rizzi et al., 2008), most focus on one or two areas of interest. This makes it diﬃcult to determine which areas of the brain are most vulnerable to anesthesia related neurotoxicity and to resolve discrepancies within the literature (Brambrink et al., 2012; Brambrink et al., 2010; Zhang et al., 2016; Zou et al., 2009). Moreover, by focusing on speciﬁc brain regions, we cannot determine if neurotoxicity is caused by anesthetic drugs acting on in- dividual cells or is the result of disrupted activity at a network level.

Another obstacle in understanding the nature of anesthesia-related neurotoxicity is the focus on a single time point to assess neurodegen- eration. Most studies quantify neurodegeneration several hours after the ending of exposure. In some ways, this approach is logical as many markers of neurotoxicity, such as Fluoro-jade C (FJC) or caspase-based stains, are ephemeral in nature. However, assessing markers of toxicity at a single timepoint assumes neurodegeneration happens at the same pace and via a single mechanism throughout the brain. A lack of ap- propriate neuronal stimulation can increase apoptosis (Kilb et al., 2011; Lewin and Barde, 1996). Accordingly, a “second wave” of neurode- generation in brain regions heavily innervated by areas eﬀected in the initial insult might be expected at later time points. Conversely, internal mechanisms to ameliorate early increases in neurodegeneration within a region may only be observable at later time points. For example, a decrease in baseline apoptosis or neurogenesis may be observed in animals exposed to anesthesia to compensate for earlier damage (Jiang et al., 2016). There may be a decrease in the number of newborn cells (neuronal or glial) in regions with high levels of neurodegeneration due to loss of neurotrophic factor(s) that would have been released by dying neurons (Lewin and Barde, 1996). It is unlikely that moments where the highest level of neurodegeneration are observed are optimal to detect endogenous compensation.

Sevo is currently the most frequently used general anesthetic in humans, and its neurotoxic potential has been well described (Amrock et al., 2015; Brioni et al., 2017; Delgado-Herrera et al., 2001; Fang et al., 2012; Lerman and Johr, 2009; Pellegrini et al., 2014; Walters and Paule, 2017; Zheng et al., 2013a). Accordingly, the primary goal of this study was to determine regional and temporal patterns of acute neu- rodegeneration that occurs in the rat forebrain following perinatal an- esthesia exposure in fully oxygenated animals. To do this, PND 7 rats were exposed to Sevo (2.5% in 75% oxygen/25% nitrogen carrier gas for 6 h). Animals were sacriﬁced at 2, 24, or 72 h after cessation of Sevo. This enabled the determination of how markers of neurodegen- eration (FJC) diﬀered by region and changed with time. Throughout this study our primary endpoint was on the incidence of FJC positive cells. We selected FJC because it is one of the most commonly used methods to study degenerating neurons, making it an excellent mark of neurotoxic insult. It is eﬀective at highlighting areas of the brain that have been impacted by general anesthesia and serves to eﬀectively

2

Neurotoxicology and Teratology 80 (2020) 106890

demonstrate areas of potential interest. However, an increased in- cidence in neurodegeneration is only one of the many changes that have been observed in the brains of animals exposed to general anesthesia early in life. Early life exposure has been shown to decrease dendritic spine densities (Briner et al., 2011), alter development of GABAeric networks (Osterop et al., 2015), inﬂuence neurogenesis (Stratmann et al., 2010), alter neurotransmitter receptor densities (Zurek et al., 2014), and induce neuroinﬂammation (Zheng et al., 2013a). Any one of these changes may cause the observed changes in cognition and beha- vior that have been reported.

2. Methods

2.1. Animals

Pregnant Sprague-Dawley rats (Charles Rivers, USA) arrived on gestational day 5. Litters were culled to four males and four females on PND 4. A within-litter treatment design was used to evaluate the eﬀect of Sevo. Four animals, 2 males and 2 females, were selected from each litter. One animal of each sex was randomly assigned to the Sevo group with the other being assigned to the vehicle condition. Three animals per sex were assigned to each treatment / timepoint combination (N = 72). Multiple stains were used on tissue from the same animal. Two animals were removed from the 72 h Sevo group. One animal was removed for failing to meet inclusion criteria for oxygenation (no hy- poxic animals were included in the study), and a second animal died between exposure and sacriﬁce. Rats were housed in a light (12 h/12 h light/dark cycle) and temperature (22 ± 2 °C) controlled vivarium and given free access to food and water (NIH41 laboratory animal diet, Envigo, Madison, WI). All animal procedures were carried out in ac- cordance with the Guide for the Care and Use of Laboratory Animals. Animal use and procedures were approved by the NCTR Institutional Animal Care and Use Committee (IACUC), which has full NIH-OLAW accreditation. Animals were housed in the NCTR facility in isolator top boxes with wooden chip bedding and ad libitum food and water.

2.2. Study design

On PND 7, rats were exposed to vehicle gas alone (75% oxygen/25% nitrogen) or 2.5% Sevo (in vehicle gas) for 6 h. During anesthesia ex- posure, each pup was placed in an individual airtight acrylic chamber and the selected gas mixture was delivered at a ﬂow rate of 0.75–1 L/ min (Walters et al., 2020). The concentration of Sevo was set using a commercial gas analyzer (Riken, USA). Surface body temperature was collected prior to and every 2 h following the start of Sevo exposure using an infrared thermometer (Micro-Epsilon, Ortenburg, Germany). Heating plates located beneath each chamber were used to maintain body temperatures at baseline levels. In addition, arterial oxygen sa- turation (SpO2), breath rate, heart rate, and pulse distention were monitored in each pup continuously using a pulse oximeter (Starr Life Sciences Corp, USA). The average SPO2 value was calculated every 30 min (Supplemental Table 1); if an individual rat's SPO2 fell below 85% during any of the 30 min intervals, it was excluded from the study. Upon recovery, the pups were removed from the chambers, rubbed with bedding material from their home cage, and returned to their dams. Control animals were treated the same as the experimental group ex- cept they were not exposed to anesthesia and there SPO2 was not monitored.

Pups were sacriﬁced 2 h (PND 7), 24 h (PND 8), and 72 h (PND 10) after the cessation of Sevo or vehicle gas exposure. Brieﬂy, the rats were deeply anesthetized with pentobarbital and transcardially perfused with 0.9% heparinized saline followed by 10% neutral buﬀered for- malin. Brains were removed and post-ﬁxed in 10% neutral buﬀered formalin for 24 h, cryoprotected in 20% sucrose until they sank, and subsequently frozen on dry ice and stored at −80 °C. Tissue was cut into 30 μm thick coronal sections using a cryostat, stored in 0.08%

S.M. Burks, et al.

sodium azide in PBS for up to two weeks and then transferred to freezing solution (0.02 M phosphate buﬀer (pH 7.4) containing 25% (v/ v) glycerol and 30% (v/v) ethylene glycol) until processed for im- munohistochemistry or histology.

2.3. FJC immunolabeling

For FJC labeling, a modiﬁed method (Bowyer et al., 2018b; Schmued et al., 2005) was used. Brieﬂy, sections of interest were re- moved from freezing solution and rinsed three times in 0.1 M phosphate buﬀer (PB, pH 7.4) for 1 min. Sections were then mounted on gelatin coated slides in 0.005 M PB (pH 7.4) and dried at 50 °C for 2 h. Sub- sequently, slides were immersed for: 3 min in basic alcohol, 2 min in 70% ETOH, 2 min in Millipore water, 11 min in 0.06% potassium permanganate, 2 min in Millipore water, 10 min in FJC (0.00001% in 0.1% glacial acetic acid), and three 2 min washes in Millipore water. Slides were then dried at 50 °C for 5–10 min, cleared with xylene for 1 min, and cover-slipped with DPX mounting media.

2.4. Mki67 immunolabeling

A buﬀer of 0.1 M PB (pH 7.4) containing 0.4% Triton X-100 was used in all the steps involving free ﬂoating sections agitated on an or- bital shaker. Sections containing regions of interest were initially wa- shed in buﬀer three times (15 min each) to remove excess freezing solution. After a 30 min pre-incubation in 4% normal goat serum, the sections were incubated in 4% serum and chicken polyclonal antibody to Mki67 (1:2000, EnCor Biotechnology, USA) for 1 to 2 h at room temperature followed by 18 to 24 h at 5 °C. Sections were then washed three times for 15 min and incubated in a biotinylated goat anti-chicken antibody (1:350, Invitrogen, USA) for 1 h at room temperature. The sections were then washed three times (15 min per wash) and incubated in Streptavidin TRITC (1:200, Jackson ImmunoResearch, USA) for 1 h. The sections were then washed three times (15 min per wash) and mounted on Superfrost Plus slides (Thermo Fisher Scientiﬁc, USA) and dried at room temperature for ≥12 h in the dark. Finally, the slides were cleared in xylene and cover-slipped with DPX mounting medium.

2.5. NeuN immunolabeling

Sections containing the four regions (IG, anterior CPu (CPua), anterior thalamus, and CA1) with the highest per mm2 levels of FJC labeled cells were immunolabeled with an antibody to NeuN in con- junction with DAB visualization. Sections were washed in 0.1 M PB (pH 7.4) for 15 min and then incubated in 0.1 M PB containing 0.05% H2O2 for 10 min to suppress the endogenous peroxidases. From this point on, except for the last step of 3,3′-diaminobenzidine (DAB) pro- cessing, incubation and washing solutions consisted of 0.1 M PB con- taining 0.25% Triton X-100. Sections were then washed three times for 5 min. Following a 20 min pre-incubation in 5% normal goat serum, the sections were incubated in rabbit anti-NeuN (1:1000, Abcam, USA) antibody for 18 to 24 h at room temperature. Sections were then wa- shed three times for 5 min and incubated in a biotinylated goat anti- rabbit antibody (1:300, Thermo Fisher Scientiﬁc, USA) for 2 h. The signal was then ampliﬁed using the avidin and biotinylated horseradish peroxidase macromolecular complex (Vector Laboratories, USA) and visualized with 0.5 mg/mL of DAB in Tris-HCl buﬀer. Sections were washed twice for 5 min in Tris-HCl, mounted, and dried on a slide warmer for ≥12 h. Finally, the slides were cleared in xylene and cover- slipped with DPX mounting medium.

2.6. Thionine staining

Thionine staining was performed to verify brain regions. Sections from regions where the highest levels of FJC staining were observed from PND 7, 8 and 10 were mounted from 0.1 M PB (pH 7.4) on

3

Neurotoxicology and Teratology 80 (2020) 106890

Superfrost Plus slides (Thermo Fisher Scientiﬁc, USA) and dried at 55 °C for 15 min. They were then immersed in double distilled water for 4 min. Subsequently, the sections were immersed in a solution of 0.1% thionine acetate (Sigma-Aldrich, USA) in double distilled water for 8 min. The sections were then transferred through two washes of water (2 min each) followed by 70% ethanol in water (2 min), 95% ethanol (2 min) and 100% ethanol (2 min). The sections were then transferred to xylene for ≥2 min and cover-slipped as described above.

2.7. Image capturing and analysis

Imaging of brain tissue was conducted using a Nikon Eclipse Ni microscope equipped with digital cameras (Photometrics, USA; Nikon, USA). FJC, Mki67, NeuN, and thionine labeling were quantiﬁed in the somatosensory cortex, motor cortex, CPu, thalamus, CA1 region of the hippocampus, septum and amygdala at 10× magniﬁcation using NIS Elements AR automated software (Nikon, USA). Brain regions were deﬁned in accordance with brain atlases for adult and neonatal rats (Paxinos and Watson, 2014; Ramachandra and Subramanian, 2011).

2.8. Stereological analysis

The brain regions in which the highest levels of FJC positive neu- rons were identiﬁed with the aid of adult and neonatal atlases as guides. Given the absence of a complete neonatal atlas, these locations were veriﬁed using thionine stained sections from the same regions that the FJC sections were taken to determine neurodegeneration from the pups sacriﬁced at PND 7, 8 and 10 [see Fig. 1]. Subsequent to identifying these regions, unbiased stereological estimates of positively im- munolabeled cells/structures were performed from images that were captured with Photometrics (ﬂuorescent) or Nikon (brightﬁeld) digital cameras using NIS elements AR software for analysis.

Unless otherwise stated, six animals were used for each treatment condition; however, only 4 animals reached inclusion criteria under the 72 h Sevo condition. For each animal included, one instance of each region was utilized. When more than one instance of a region was available for an animal, the average count for that region, in that an- imal, was utilized for statistics. Therefore, N is reﬂective of the number of animals included in each treatment group. All FJC, NeuN and Mki67 cell counts represented are calculated from both hemispheres, and thus represent the region as a whole. Data were normalized by the total area of the counted region of interest and expressed as number of label positive cells per mm2. All animals were given a unique ID that was not indicative of treatment prior to analysis. The investigator who took images and conducted immunohistochemical analysis was unaware of treatment conditions and post-exposure intervals at the time of analysis. FJC images were taken at 10× using a FITC ﬁlter and Photometrics camera. FJC positive cells were counted in NIS Elements AR by re- stricting the area to ≥4μm2 and ≤ 75 μm2, MinFret ≥1.93 μm, and SumIntensity ≥102 and ≤ 2,876,455. Threshold settings were as fol- lows: Smooth 5×, Clean 2×, Fill holes ON, Separate 1×. Binning was set at 8. Mki67 images were taken at 20× using TRITC ﬁlter, ND4 ﬁlter, and Photometrics camera. Mki67 positive cells were counted in NIS Elements AR by restricting Area ≥ 24 μm2 and ≤ 300 μm2, Width ≥ 2.29 μm, and MeanIntensity ≥229 and ≤ 65,535. Threshold settings were as follows: Smooth OFF, Clean OFF, Fill holes OFF, Separate 4×. Binning was set at 8. NeuN images were taken at 10× using brightﬁeld imaging and a Nikon camera. Images of NeuN positive cells in the CPua and anterior thalamus were ﬁrst processed by a medium equalization accuracy strength of 30, followed by Fourier transform noise reduction at 0.880 and detail enhancement at 0.045 with average intensity maintained. NeuN positive cells were then counted in NIS Elements AR by restricting area ≥ 0.42 μm2 and ≤ 2276.45μm2, threshold set to intensity with the following set- tings: Smooth OFF, Clean 1×, Fill holes OFF, Separate 4×. Binning was set at 8. For the densely populated regions of CA1 and IG, to get

S.M. Burks, et al.

4

Neurotoxicology and Teratology 80 (2020) 106890

(caption on next page)

S.M. Burks, et al.

Neurotoxicology and Teratology 80 (2020) 106890

Fig. 1. Coronal brain sections showing regions with highest levels of increased neurodegeneration after Sevo anesthesia. The left most column (A1, +2.76 from bregma through G1, −4.80 from bregma) shows coronal sections in adult rats (Paxinos and Watson, 2014) that correspond to coronal sections in the PND 7 (A2 to G2), PND 8 (A3 to G3) and PND 10 (A4 to G4). The gold color highlights superimposed on the adult sections correspond to the regions in the neonates where neurodegeneration was highest. The regions in neonates do not correspond perfectly with the adults, as seen in the septal, posterior cortical and midbrain regions. The lateral septum is not highlighted in gold because it occurs rostral to C1, at 0.24 mm from bregma, while the retromammillary decussation and ventral tegmental area (rostral), and substantia nigra, reticular part and dorsal tier are not shown because they are caudal of bregma −4.80 mm. Image G1 is a composite of images from the adult diencephalon (bregma −3.96 mm) and cortex / hippocampus (bregma −4.80). This was necessary to provide an adequate representation of the more caudal aspects of the PND 7–10 rat brain. Abbreviations for the gold highlighted brain regions of interest in A1 through G1 are: A29c-1 = more anterior region of the retrosplenial cortex A29c-2 = more posterior region of the retrosplenial cortex Acb = accumbens shell AHP = anterior hypothalamic area, posterior part AIV & LO = agranular insular cortex ventral + lateral orbital cortex Anterior Thalamus = VA, VL and intralaminar nuclei of the thalamus BMA & MeA = basomedial amygdaloid nucleus + medial amygdaloid nuclei ST = Bed Nuclei STM; lateral + medial division of bed nucleus of the stria terminalis CA1 = ﬁeld CA1 of hippocampus CG Ctx = cingulate cortex CPua = caudate/putamen, ~ 2.16 mm bregma CPub = caudate/putamen, ~ −0.12 to −0.24 mm bregma DLG = dorsal lateral geniculate nucleus GP = lateral globus pallidus IG = indusium griseum, hippocampal rudiment M1 & M2 = primary + secondary motor cortex, layer I & II M2 = secondary motor cortex, layer I & II MPO = medial preoptic nucleus of the hypothalamus PMCo = posteromedial cortical amygdaloid nucleus VTA = retromammillary decussation + ventral tegmental area, rostral S1BF = primary somatosensory cortex, barrel ﬁeld SNR & SNCD = substantia nigra, reticular part + dorsal tier.

Fig. 2. Regions with the highest levels of FJC labeling at 2 h after Sevo. The average ± SEM of FJC positive cells after 2 h Sevo, normalized to 1 mm2 area, are displayed in de- creasing order. N = 6, however for the Bed Nuclei STM, AHP, and SNR & SNCD regions, N = 5 due to proces- sing/regional complications. As the sections used were 30 μm thick, the total number of degenerating neurons per 1 mm3 would be 40× that shown for each region. Abbreviations present in Fig. 2 are the same as used in Fig. 1. The blue background indicates that the region is associated with the primary olfactory learning system while the tan background indicates that the region is part of the basal ganglia and motor movement system. The arrow points to the values related to the SNR and SNCD. **P ≤ 0.002, ***P ≤ 0.0002, ****P ≤ 0.00002.

accurate cellular discrimination of NeuN, images were ﬁrst processed by Fourier transform noise reduction at 0.543 and detail enhancement at 0.034 with average intensity maintained. Subsequently, a green component contrast was applied with the following settings: Low = 0, High = 177, Gamma =5. Then the positive cells were counted using the same settings as for CPua and anterior thalamus.

2.9. Statistical analysis

Results are presented as mean ± SEM. The 2, 24, and 72 h groups were analyzed using separate two-way ANOVAs (region and treatment). Post-hoc analysis was performed using a Sidak's post-hoc test (alpha = 0.02), or unpaired t-tests, (eﬀect of sacriﬁce interval; alpha = 0.05). All analyses were performed in GraphPad Prism version 6 (GraphPad Software, Inc.; USA). Due to processing/regional compli- cations, in certain circumstances N of 6 was not available. The Bed

5

S.M. Burks, et al.

6

Neurotoxicology and Teratology 80 (2020) 106890

(caption on next page)

S.M. Burks, et al.

Neurotoxicology and Teratology 80 (2020) 106890

Fig. 3. Visual presentation of FJC labeling in three regions with high levels of FJC labeling at 2 h after Sevo exposure. Representative micrographs of FJC labeling in the control and Sevo animals for the CPua, Acb, and M1 & M2 are shown 2 h after Sevo. Both high and low magniﬁcations are shown; the red arrows point to the same location on the two diﬀerent magniﬁcations (indicated by the red magniﬁcation bars). The larger FJC labeled structures in the three regions ranged from 2 to 8 μm2.

Fig. 4. Regions with highest levels of FJC labeling at 24 h after Sevo. The regions with the highest increases in (> 4 and ≤ 75 μm2) FJC structures present at 24 h after Sevo are shown on the x- axis. The average of means along with the SEMs shown on the y-axis are nor- malized to 1 mm2 area for both control and Sevo groups. Abbreviations present are the same as used in Fig. 1. The blue background indicates that the region is associated with the primary olfactory learning system while the tan back- ground indicates that the region is part of the basal ganglia and motor move- ment system. The arrow points to the bars related to the SNR and SNCD. ****P ≤ 0.00002.

the

numbers

of

large

Nuclei STM, AHP, and SNR & SNCD regions of the 2 h post Sevo ani- mals, N = 5 (Fig. 2). For the 24 h post Sevo animals in the Anterior Thalamus, Bed Nuclei STM, A29c-1, S1BF, BMA & MeA, VTA, MPO, and SNR & SNCD regions, N = 5 (Figs. 4 and 9). For the 24 h post Sevo animals in the GP and AHP, N = 4 (Fig. 4). For NeuN, 24 h post Sevo controls were N = 4 in the IG (Fig. 7). Controls for Mki67 at the anterior thalamus where N = 5 (Fig. 9). In the CPua, 24 h post Sevo and control animals were N = 4 for Mki67 (Fig. 9).

3. Results

Preliminary data identiﬁed over ten speciﬁc brain regions in which the density of FJC labeling within the region was at least three-fold increased over other brain regions. FJC labeling within these regions, as well as some regions which have signiﬁcant synaptic connections to the identiﬁed regions, was then conducted to determine the statistical dif- ferences between regions over the three timepoints (PND 7 at 2 h post Sevo, PND 8 at 24 h post Sevo, and PND 10 at 72 h post Sevo).

Regions highlighted in gold are superimposed over coronal sections (+2.76 to-4.8 mm from bregma) of an adult rat brain in Fig. 1, in- dicating increased FJC labeling due to Sevo. Corresponding thionine- stained brain sections from neonatal pups at PND 7, 8 and 10 are shown in the remaining three columns. The anatomy of the brain regions in the adult versus neonate coronal sections approximately correspond. However, more posteriorly, the corpus collosum appears to extend further back with respect to the midbrain. Thus, there is an apparent 1 mm discrepancy caudally at −5.0 mm from bregma on PND 7–10. At this age, the neocortex and the hippocampal morphology agrees with a position at −4.8 to −5.0 mm from the bregma in adults while the midbrain shown corresponds to about −4.0 mm in adults. This dis- crepancy is less pronounced at PND 10. In the sections more rostral, the correspondence seems to be uniform up to +2.78 mm from bregma. Note that the lateral septum is not highlighted in Fig. 1 because it

occurs rostral to C1, at 0.24 mm from bregma. As can be seen in Fig. 1, the size of the coronal sections in the neonatal brain increase about 30% from PND 7 to 10.

Increased FJC labeling was determined by evaluating the number of FJC-labeled neurons per mm2 in brain regions of the Sevo or control group. Two hours after Sevo, neurodegeneration was most prominent in brain regions associated with the primary olfactory learning system (Fig. 2, blue), as well as the basal ganglia and thalamic and cortical regions related to motor movement (Fig. 2, tan). The regions of the primary olfactory learning system included: the indusium griseum (IG), lateral septum, accumbens shell (Acb), posterior hippocampal CA1 (lighter color indicates looser association), bed nucleus of the stria terminalis (Bed Nuclei STM), posteromedial cortical nuclei of the amygdala (PMCo), and the amygdaloid nucleus (BMa & MeA). In the basal ganglia associated brain regions, layer I and II of the anterior motor cortex (M1 & M2), anterior thalamus, lateral globus pallidus (GP), and caudate/putamen (CPua,b, both anterior medial and more posterior ventral) were aﬀected. The retrosplenial cortex (A29c-1 and A29c-2), barrel ﬁelds of cortical primary somatosensory (S1BF), sec- ondary motor cortex (M2), and cingulate cortex (CG Ctx) were the other regions most aﬀected by Sevo. The total number of FJC-labeled neurons per region, irrespective of its total area, are found in Supplemental Fig. 1. (Because the areas of regions analyzed varied greatly, the ab- solute numbers per region are greater in the regions of the anterior thalamus and CPu, which encompass larger total areas. From that standpoint, the regions of the anterior thalamic nuclei (−1.2 to 3.0) and CPub had the greatest number of FJC labeled neurons followed by CA1, PMCo, Bed Nuclei STM, A29c-1, A29c-2, and Acb.) FJC labeling in the substantia nigra (SNR & SNCD) and ventral tegmental area (VTA) was minimal (Fig. 2 and Supplemental Fig. 2). Also, there was very little labeling in the frontal cortex at +4.2 mm from bregma (Supplemental Fig. 1). Representative micrographs show Sevo exposure increases FJC labeling for the CPua, Acb, and M2 (Fig. 3).

7

S.M. Burks, et al.

Twenty-four h after Sevo, the FJC labeling was within 3-fold of control in half the brain regions (Fig. 4). There was still a 3 to 6-fold increase in the numbers of FJC labeling in the Bed Nuclei STM, Acb and more anterior medial CPu (CPua) relative to control at 24 h post Sevo.

8

Neurotoxicology and Teratology 80 (2020) 106890

Fig. 5. Visual presentation of FJC labeling in three regions with high levels of neurodegeneration 24 h after Sevo exposure. Representative micrographs for neurodegeneration/FJC labeling in the control and Sevo groups for the IG, Bed Nuclei STM and CA1 are shown at 24 h after Sevo. Both high and low magniﬁcations are shown; the red arrows point to the same location on the two diﬀerent magniﬁcations as indicated by the red magniﬁcation bars. FJC labeling in the CA1 is present in objects approaching 9 μm in diameter with a pronounced layer of FJC labeled puncta above in the region of ﬁbers of passage. FJC-labeled objects of the same size are seen in the IG with a light interspersion of puncta. FJC-labeled objects present in the Bed Nuclei STM are ≤8 μm with very few puncta being present.

However, after 24 h, in the IG, posterior CA1, anterior thalamic nuclei, and GP, there was still 25.7, 17, 11.7, and 9.8-fold, respectively, higher levels of FJC labeling compared to control. There was very little eﬀect of Sevo on FJC labeling in most of the hippocampus except for the striking increases in CA1, caudally just before the appearance of the subiculum where the morphology of the cortex and hippocampus cor- respond to the CA1 region present from about −4.5 to −5.0 mm from bregma (Fig. 5). High levels of FJC labeling can be seen in the lateral dorsal (both AD and LDVL) central medial (CM) and ventromedial (VM) nuclei of the thalamus (Fig. 6). The levels of FJC labeling in all brain regions were within 3.8-fold or less at 72 h after Sevo compared to control except for the M1 & M2 motor cortex at −0.6 mm from bregma (Supplemental Fig. 3). Interestingly, high levels of FJC labeled struc- tures of the size of degenerating neurons were seen at 72 h after Sevo in two of the four rats evaluated (Supplemental Fig. 4). These are most likely images of neurons dying within the last 24 h.

The total number of surviving neurons per brain region (at PND 7 and 8) with the high increases in FJC labeling were determined by using NeuN immunolabeling, which will detect most but not all neurons (Mullen et al., 1992). Micrographs of these NeuN labeled regions can be found in Supplemental Fig. 5. The number of NeuN labeled cells within a region in one section and the number of FJC cells in an adjacent section for the IG, CPua, anterior thalamus and CA1 (Fig. 5) were de- termined. A single section was not as feasible for use for dual staining due to the fragility of the PND 7 and 8 sections and the densely packed NeuN neurons labeled with DAB obscured the FJC labeled cells. The percentage of FJC labeling relative to NeuN labeled cells within a re- gion after Sevo or vehicle gas at 2 h (Fig. 7) was calculated via the following method [(total FJC positive cells per 1mm2 / total NeuN positive cells per 1mm2)*100]. Unfortunately, enough tissue did not remain at the 24 h condition to determine the total number of NeuN positive cells. However, the number of NeuN positive cells were little changed between the 2 and 72 h conditions (15% diﬀerence for the IG). Accordingly, the number of NeuN positive cells at the 2 h condition was also used for the 24 h condition.

The newly-born cells within regions were determined at 2 h and 24 h using antibodies to Mki67 to determine if the regions with more intense Sevo-related neurodegeneration (FJC labeling) had any obvious connection with regions of new cell birth. There were appreciable numbers of Mki67 labeled nuclei in the CPua, Acb, S1BF, and LDVL (Fig. 8). The regions of mitotic activity are identiﬁed at 10×; enlarged Mki67 nuclei about to separate are also shown at 20× for clarity. The numbers of Mki67 labeling nuclei per mm2 were determined in both control and Sevo groups at 2 and 24 h in the CPua, anterior thalamus, and VM nuclei of the thalamus (Fig. 9). It was expected that if the newborn cells, which would not be neuronal, in these regions were dying, that there would be fewer Mki67 labeled nuclei in the Sevo group at 2 h. At 24 h they might be either: 1) decreased due to con- tinued degeneration or, 2) increased due to compensation for loss at 2 h. However, eﬀects of Sevo were variable between the thalamus and CPua, with no clear evidence that Mki67 labeling was altered by Sevo. Mki67 labeled cells were present in high numbers in all the regions with high FJC labeling.

S.M. Burks, et al.

4. Discussion

PND 7 rats were exposed to Sevo general anesthesia after which, regions of the forebrain were quantiﬁed with FJC, and how the regional pattern of neurodegeneration varied with time was determined. Pathways of the basal ganglia and portions of the olfactory learning system were most aﬀected by Sevo. Remarkably, in some brain regions as many as 10% of the total neurons appeared to be dying 24 h after Sevo exposure (IG; Fig. 7). We also observed signiﬁcant neurodegen- eration in areas of the brain, such as the IG (Figs. 2 and 4), and in other areas that failed to reach statistical signiﬁcance, such as the lateral septum and the bed nucleus STM that have been previously overlooked. In contrast to previous studies (Lee et al., 2017; Perez-Zoghbi et al., 2017; Zhou et al., 2016), we observed clear, but spatially restricted, damage to the CA1 region of the hippocampus. By 72 h signs of neu- rodegeneration were greatly reduced. The reduced proﬁle of hippo- campal damage observed here may be caused by the exclusion of ani- mals that did not maintained adequate oxygenation throughout the course of exposure. Transient levels of low oxygen during prolonged Sevo exposure may result in a more pervasive pattern of neurodegen- eration within the hippocampus.

9

Neurotoxicology and Teratology 80 (2020) 106890

Fig. 6. Visual presentation of FJC labeling in four thalamic nuclei with high levels of in- creased FJC labeling at 24 h after Sevo exposure. The large top panel shows the entire thalamus in one hemisphere. The third ventricle and the ventral aspects of the hippocampus are present at the very top of the panel. The panel is a composite of twenty-four micrographs taken at 10× magniﬁcation. FJC labeling was prominent in the lateral dorsal thalamic nucleus, ven- trolateral portion (LDVL), anterodorsal portion (AD), ventromedial (VM) and central medial (CM) nuclei of the thalamus as shown in the bottom four panels. Magniﬁcation is the same for all four and represented in the far-left panel. White boxes highlight the areas in the lower panels. FJC labeling is present in objects up to ⩰ 9 μm in diameter with a light interspersion of puncta.

Several regions of the forebrain showed elevated levels of FJC staining 2 h after Sevo exposure (Fig. 2). The brain regions aﬀected could be roughly split into two groups: those related to the basal ganglia and motor movement (anterior thalamus, CPua,b, GP, M1& M2, GP, M2 and possibly S1BF (Gerfen and Surmeier, 2011; Ikemoto et al., 2015) and regions that have been associated with the primary olfactory learning pathway (IG, CA1, Bed Nuclei STM, Acb, lateral septum, BMa & MeA and PMCo (Shipley and Adamek, 1984)) as well as “spatial cognition” (CA1, (Gilbert et al., 2001; Kesner et al., 2004) (Gallagher and Chiba, 1996), fear learning (BMa & MeA, and PMco (Gallagher and Chiba, 1996)), reward and addiction (Acb (Di Chiara, 2002)), response to stress (Bed Nuclei STM (Choi et al., 2008)) and many other aspects of behavior. Elevated levels of FJC, although not statistically signiﬁcant, were also detected in other areas known to be important to learning and memory such as A29c-1 and A29c-2 (Nelson et al., 2018; Todd et al., 2019). By 24 h after Sevo (Fig. 4), the same general regional pattern was seen in the basal ganglia pathway and the primary olfactory learning pathways in the top twelve brain regions. Moreover, neuro- degeneration was again detected in A29c-1 and A29c-2 as well as the AHP.

Most studies investigating anesthesia related neurotoxicity evaluate

S.M. Burks, et al.

Fig. 7. Percentage of munoreactive cells at 2 and 24 h after Sevo. The percentage of FJC labeled cellular structures within a region in Sevo and control groups was indirectly calculated by determining the numbered of NeuN labeled cells within a region in one section and the number of FJC cells in an adjacent section for the IG, anterior CPu anterior thalamus and CA1. The per- centage of large (> 4 and ≤ 75 μm2 in area) FJC structures relative to NeuN labeled cells present at 2 and 24 h after Sevo or in control are shown. The mean percentages along with SEMs are displayed. ** indicates P ≤ 0.0001.

large FJC structures present relative to NeuN im-

animals at a single time point. In contrast, we evaluated Sevo related neurodegeneration at 2, 24, and 72 h after the cessation of exposure. Neurodegeneration was detected subsequent to what is seen at 2 h and regions experiencing prolonged neurodegeneration and/or those un- dergoing a “second wave” of neurodegeneration were identiﬁed. Some of the regions associated with the basal ganglia pathway and the pri- mary olfactory learning pathway showed signs of either a prolonged neurodegeneration (CPua) or a second neurodegenerative event (IG, CA1, and thalamus). These regions had multifold increases in FJC staining at 24 h compared to 2 h. In contrast, the M1 & M2, PMCo, CPub, and cingulate more rapidly returned to control, or near control levels. For example, the PMCo and CPub have levels of FJC staining approximately 5-fold higher than control at 2 h, but levels were es- sentially normal at 24 h post exposure. These data indicate there are regional diﬀerences in sensitivity to anesthesia related neurotoxicity, and that timing of neurotoxicity diﬀers in a region dependent manner. One explanation for this phenomenon could be delayed scavenging

10

Neurotoxicology and Teratology 80 (2020) 106890

of the dead neurons in brain regions with high neurodegeneration; that is, the brain could be rate-limited in the clearance of dead neurons. Dead neurons in excess of the threshold could be misinterpreted as dying at a later point. However, “Delayed scavenging” was not ob- served in the Acb and CPub. These regions had high levels of neuro- degeneration at 2 h, but 5-fold drops in the number of degenerating neurons labeled with FJC at 24 h. Moreover, delayed scavenging cannot explain the actual increase in degenerating neurons in the IG, anterior thalamus, and CA1 at 24 h compared to 2 h. These data indicate the initial loss of neurons in some regions associated with the basal ganglia and the primary olfactory learning pathway is followed by a “second wave” of cell death. Endogenous neurodegeneration might increase because of the loss of cellular inputs triggered by the “ﬁrst wave” of neurodegeneration. Alternatively, Sevo exposure may cause a second wave of neurodegeneration at 24 h via a fundamentally diﬀerent me- chanism and with a diﬀerent temporal proﬁle from that observed at 2 h. Regardless of the mechanism, the prospect of a distinct second wave of neurodegeneration increases the diﬃculty in ameliorating anesthesia related neurodegeneration and highlights the importance of con- sidering multiple time points when studying anesthesia related neuro- degeneration.

It is unclear why Sevo causes an increase in FJC positive cells in some areas, while leaving others unaﬀected. It is not as simple as an abundance of newly divided cells. Many areas of the brain where ele- vated numbers of FJC positive cells were observed were past their periods of “peak” neurogenesis. For example, neurogenesis is thought to peak at post-conception day (PCD) 17 in the Acb (Clancy et al., 2009), PCD 16 in the CA1 (Wyss and Sripanidkulchai, 1985) (Workman et al., 2013), and PCD 14 in the CPu (Clancy et al., 2009; Workman et al., 2013). In the CA1 we observed about 5% of all cells being FJC positive (Fig. 7) even though neurogenesis in the region is greatly reduced by PCD 20 (Wyss and Sripanidkulchai, 1985). These data therefore suggest that a cell being recently born is not enough to make it vulnerable to anesthesia related neurotoxicity.

Another possible explanation for certain areas being more sensitive to Sevo is the high prevalence of GABAA receptors. While the IG does have high levels of GABAA receptor mRNA (PND 5 rats (Poulter et al., 1992)), so does the dentate gyrus (Poulter et al., 1992) where little to no evidence of neurodegeneration was detected. Some of the neuro- degeneration may be the consequence of a lack of organized stimula- tory input caused by anesthesia exposure. For example, previous work has demonstrated an increase in degenerating cells in the substantia nigra (bilateral) 1–4 days after an excitatory striatal lesion (unilateral) in the PND 7 rat (Macaya et al., 1994). Similarly, stimulatory activity is required for normal cortical development (Kilb et al., 2011; Lewin and Barde, 1996). However, it is unclear if a lack of organized stimulation could so rapidly impact neurodegeneration and cause an increase in FJC staining 2 h post Sevo cessation. If that is indeed the case, it would ultimately translate into lasting behavioral changes when normal sti- mulation has been restored. Clearly, additional studies are required.

Another possibility is that some neuron types, more abundant in certain regions, are more vulnerable to insult in adults and neonates. For example, the basal ganglia associated regions of the thalamus, which are shown here to be sensitive to Sevo, are also sensitive to cell death as a result of thiamine deﬁciency and methamphetamine in adult rodents (Bowyer et al., 2008; Bowyer et al., 2018b). Except for the retrosplenial (A29c) and motor cortex, all brain regions where Sevo resulted in high levels of neurodegeneration in the perinatal rat also display pronounced neurotoxicity in adult rodents exposed to me- thamphetamine. Most of the potential mechanisms behind the hy- perthermic and excitatory neurotoxicity produced by amphetamines and seizures would not be thought to occur during anesthesia. How- ever, blood ﬂow disruption within regions where neurodegeneration occurs has been observed with methamphetamine and amphetamine exposure, and likely triggers the seizures and neurodegeneration pro- duced (Bowyer et al., 2018a). Sevoﬂurane can induce vasodilation

S.M. Burks, et al.

Neurotoxicology and Teratology 80 (2020) 106890

Fig. 8. a,b. Mki67 immunolabeling labeling in the CPua, Acb, S1BF and LDVL at 2 h after Sevo anesthesia (8a) or control (8B). Cells undergoing the process of mitosis in the CPua, Acb, S1BF and LDVL at 2 h after Sevo exposure were labeled using Mki67. The white boxes indicate where mitosis has just occurred with two new nuclei appearing in the CPua, Acb, and S1BF (a). In the LDVL, it appears some time has passed since the original nucleus has divided (a). Magniﬁcation is the same for all four regions and is shown in the two CPua panels.

(Matta et al., 1999; Sakata et al., 2019). It is therefore possible that, in the perinatal rats exposed to Sevo, a deﬁciency in brain blood-ﬂow occurs in regions with pronounced neurodegeneration. Young animals may lack the physiological robustness needed to overcome periods of abnormal regional blood ﬂow.

FJC is the most recent and most selective ﬂuorescent ligand/label developed to detect neurodegeneration, it can label the soma, den- drites, axons and terminals of degenerating neurons in adult animals (Schmued et al., 2005). However, in perinatal animals, the Fluoro Jade (versions b and c) labeling is primarily observed as circular structures ranging from 2 to ≤10 μm in diameter with very little or no labeling of the dendrites and axons as seen in the present study and previously (Scallet et al., 2004). The location, morphology and size of the larger FJC structures seen in perinatal animals after Sevo exposure coincides well with the TUNEL data, which detects apoptotic degeneration by labeling the degraded fragmenting DNA (Scallet et al., 2004; Schmued et al., 2005). The ligand(s) generated during neurodegeneration that covalently bind to FJC is unknown, but it is likely highly positively charged. It is not known why dendrites and soma (other than the nu- cleus) were not labeled with FJC in perinates.

In normal perinates, FJC labeling is present among the degrading DNA and histones found in or surrounding the collapsing nucleus during the apoptotic process. From this study and previous research,

most of the cells labeled with FJC are likely neurons (Scallet et al., 2004; Schmued et al., 2005). Sevo exposure did not decrease the Mki67 labeling, which indicates that most of the cells observed dividing were not neuronal in nature. However, this assumes that Mki67 antibodies cannot label FJC labeled neurons owing to degradation of the Mki67 protein. In the present study, the FJC labeling in the hippocampal CA1, IG and some thalamic nuclei showed morphological similarities with adult FJC labeling. FJC labeled the entire soma and a few proximal dendrites. In the CA1 region, FJC also labeled ﬁbers of passage residing about the degenerating neurons. It is not clear why this occurred only in these regions, but it could be due to a diﬀerent type of neurodegen- erative process other than classic apoptosis, or that the neurons aﬀected in this area were further along in their diﬀerentiation to the “adult” state.

During the BGS, new cells are continually created while others are degenerating via normal apoptotic processes. This dynamic state has led some to question the clinical relevance of anesthesia related death during the BGS. Hypothetically, endogenous amelioration of anesthesia related cell death could occur by either decreasing the rate of apoptosis, or by increasing the creation of new cells. Several areas of the brain did have a transient reduction in FJC levels at either 24 or 72 h post ex- posure, suggesting potential compensation. Yet, it is unlikely that the decline in endogenous neurodegeneration is enough to compensate for

11

S.M. Burks, et al.

Fig. 9. Eﬀects of Sevo on Mki67 labeling in the CPu and thalamus. The numbers of Mki67 labeled cells in the CPua, anterior thalamus and VM nuclei of the thalamus are shown in the control and Sevo groups at the 2 h and 24 h time points. No statistical signiﬁcance was observed.

the multiple fold increase in neurodegeneration that occurred due to Sevo exposure. Similarly, Mki67, a non-speciﬁc nuclear marker of the birth/mitosis of all cells (Brown and Gatter, 1990; von Bohlen und Halbach, 2011) was used to determine if more cells were born in the areas of greatest neurodegeneration. There was no convincing evidence of an increase in new cell birth (Mki67 labeling) in regions where Sevo exposure increased neurodegeneration the most (anterior thalamus, Figs. 4 and 9). Although not statistically signiﬁcant, some brain regions trended toward elevated levels of Mki67 binding under the control condition (e.g., CPu and thalamus, Fig. 9), as well as increased neuro- degeneration after Sevo exposure. The brain does not appear to meaningfully increase the supply of viable neurons within the ﬁrst 72 h

12

Neurotoxicology and Teratology 80 (2020) 106890

of Sevo exposure. In most of these regions, neuronal cell birth is thought to have ended by PND 7–8 (Feliciano and Bordey, 2013). In future re- search, it will be important to consider the developmental fate of cells born in the brain in the hours after Sevo insult.

Many of the brain areas where neurodegeneration was observed are crucial for cognition, and prolonged general anesthesia can cause lasting behavioral changes. Surprisingly, with an independent cohort of rats treated under near identical conditions, we saw no signiﬁcant ef- fects of Sevo exposure when tested during adulthood on a battery of operant tests (Walters et al., 2020). Behavioral eﬀects of these ex- posures were limited to an altered locomotor activity response after an amphetamine challenge. It is possible that we may have detected more robust eﬀects of Sevo if we evaluated Morris Water Maze (MWM) or a radial arm maze (RAM) performance. Many reports have previously shown animals treated with anesthesia early in life to have altered performance on those tasks in adulthood (Walters and Paule, 2017). However, the MWM and RAMs are both dependent upon a functioning hippocampus (Morris et al., 1982; Olton et al., 1978), and we observed less damage to the hippocampus than reported in other studies. We included only those animals that maintained adequate oxygenation throughout the entirety of the exposure, and we observed less extensive neurodegeneration in the hippocampus, potentially because of this. The diﬀuse nature of the insult, with generally fewer than 1% of neurons dying in any area (notable exceptions) may not have been suﬃcient to impair performance on operant based tasks of cognition (Walters et al., 2020).

While still controversial, an ever-growing body of evidence in- dicates that extended exposure to general anesthesia, including Sevo, early in life has the potential to be neurotoxic. Here, we conﬁrmed that in adequately oxygenated animals, Sevo exposure caused neurodegen- eration but to a lesser extent in some brain regions (e.g., hippocampus) than previously seen. Moreover, our comprehensive assessment of the brain highlights additional regions that may be vulnerable to anesthesia related neuronal degeneration. Most striking was the damage seen in the IG; over 10% of the total cells stained positive for FJC at 24 h post- exposure. By including a time course, we also established that neuro- degeneration does not progress at a uniform pace after Sevo exposure. The temporal pattern of FJC staining also suggests a second wave of degeneration at 24 h, likely driven by a diﬀerent mechanism than that which caused degeneration at 2 h. The realization that anesthesia re- lated neurodegeneration does not occur at a uniform pace, and that damage may be more spread than initially assumed, is crucial for the future study of mechanisms of and treatments for anesthesia related neurotoxicity.

Supplementary data to this article can be found online at https://

doi.org/10.1016/j.ntt.2020.106890.

Funding

This work was funded by NCTR Protocol E07601.01.

Declaration of competing interest

The authors declare that they have no known competing ﬁnancial interests or personal relationships that could have appeared to inﬂu- ence the work reported in this paper.

Acknowledgements

The authors would like to thank James Raymick for his assistance

with tissue processing.

References

Amrock, L.G., Starner, M.L., Murphy, K.L., Baxter, M.G., 2015. Long-term eﬀects of single or multiple neonatal sevoﬂurane exposures on rat hippocampal ultrastructure.

S.M. Burks, et al.

Anesthesiology 122 (1), 87–95.

Bowyer, J.F., Thomas, M.T., Schmued, L.C., Ali, S.F., 2008. Brain region-speciﬁc neuro- degenerative proﬁles showing the relative importance of amphetamine dose, hy- perthermia, seizures and the blood-brain barrier. Ann. N. Y. Acad. Sci. 1139, 127–139.

Bowyer, J.F., Tranter, K.M., Robinson, B.L., Hanig, J.P., Faubion, M.G., Sarkar, S., 2018a. The time course of blood brain barrier leakage and its implications on the progression of methamphetamine-induced seizures. Neurotoxicology 69, 130–140.

Bowyer, J.F., Tranter, K.M., Sarkar, S., Hanig, J.P., 2018b. Microglial activation and vascular responses that are associated with early thalamic neurodegeneration re- sulting from thiamine deﬁciency. Neurotoxicology 65, 98–110.

Brambrink, A.M., Evers, A.S., Avidan, M.S., Farber, N.B., Smith, D.J., Zhang, X., Dissen, G.A., Creeley, C.E., Olney, J.W., 2010. Isoﬂurane-induced neuroapoptosis in the neonatal rhesus macaque brain. Anesthesiology 112 (4), 834–841.

Brambrink, A.M., Evers, A.S., Avidan, M.S., Farber, N.B., Smith, D.J., Martin, L.D., Dissen, G.A., Creeley, C.E., Olney, J.W., 2012. Ketamine-induced neuroapoptosis in the fetal and neonatal rhesus macaque brain. Anesthesiology 116 (2), 372–384. Briner, A., Nikonenko, I., De Roo, M., Dayer, A., Muller, D., Vutskits, L., 2011.

Developmental stage-dependent persistent impact of propofol anesthesia on dendritic spines in the rat medial prefrontal cortex. Anesthesiology 115 (2), 282–293. Brioni, J.D., Varughese, S., Ahmed, R., Bein, B., 2017. A clinical review of inhalation anesthesia with sevoﬂurane: from early research to emerging topics. J. Anesth. 31 (5), 764–778.

Brown, D.C., Gatter, K.C., 1990. Monoclonal antibody Ki-67: its use in histopathology.

Histopathology 17 (6), 489–503.

Choi, D.C., Evanson, N.K., Furay, A.R., Ulrich-Lai, Y.M., Ostrander, M.M., Herman, J.P., 2008. The anteroventral bed nucleus of the stria terminalis diﬀerentially regulates hypothalamic-pituitary-adrenocortical axis responses to acute and chronic stress. Endocrinology 149 (2), 818–826.

Clancy, B., Teague-Ross, T.J., Nagarajan, R., 2009. Cross-species analyses of the cortical

GABAergic and subplate neural populations. Front. Neuroanat. 3, 20.

Delgado-Herrera, L., Ostroﬀ, R.D., Rogers, S.A., 2001. Sevoﬂurance: approaching the

ideal inhalational anesthetic. A pharmacologic, pharmacoeconomic, and clinical re- view. CNS Drug Rev 7 (1), 48–120.

Deng, M., Hofacer, R.D., Jiang, C., Joseph, B., Hughes, E.A., Jia, B., Danzer, S.C., Loepke, A.W., 2014. Brain regional vulnerability to anaesthesia-induced neuroapoptosis shifts with age at exposure and extends into adulthood for some regions. Br. J. Anaesth. 113 (3), 443–451.

Di Chiara, G., 2002. Nucleus accumbens shell and core dopamine: diﬀerential role in

behavior and addiction. Behav. Brain Res. 137 (1–2), 75–114.

Dobbing, J., Sands, J., 1979. Comparative aspects of the brain growth spurt. Early Hum.

Dev. 3 (1), 79–83.

Eriksson, P., 1997. Developmental neurotoxicity of environmental agents in the neonate.

Neurotoxicology 18 (3), 719–726.

Fang, F., Xue, Z., Cang, J., 2012. Sevoﬂurane exposure in 7-day-old rats aﬀects neuro- genesis, neurodegeneration and neurocognitive function. Neurosci. Bull. 28 (5), 499–508.

Feliciano, D.M., Bordey, A., 2013. Newborn cortical neurons: only for neonates? Trends

Neurosci. 36 (1), 51–61.

Flick, R.P., Katusic, S.K., Colligan, R.C., Wilder, R.T., Voigt, R.G., Olson, M.D., Sprung, J., Weaver, A.L., Schroeder, D.R., Warner, D.O., 2011. Cognitive and behavioral out- comes after early exposure to anesthesia and surgery. Pediatrics 128 (5), e1053–e1061.

Gallagher, M., Chiba, A.A., 1996. The amygdala and emotion. Curr. Opin. Neurobiol. 6

(2), 221–227.

Gentry, K.R., Steele, L.M., Sedensky, M.M., Morgan, P.G., 2013. Early developmental

exposure to volatile anesthetics causes behavioral defects in Caenorhabditis elegans. Anesth. Analg. 116 (1), 185–189.

Gerfen, C.R., Surmeier, D.J., 2011. Modulation of striatal projection systems by dopa-

mine. Annu. Rev. Neurosci. 34, 441–466.

Gilbert, P.E., Kesner, R.P., Lee, I., 2001. Dissociating hippocampal subregions: double dissociation between dentate gyrus and CA1. Hippocampus 11 (6), 626–636.

Guo, P., Huang, Z., Tao, T., Chen, X., Zhang, W., Zhang, Y., Lin, C., 2015. Zebraﬁsh as a model for studying the developmental neurotoxicity of propofol. J. Appl. Toxicol. 35 (12), 1511–1519.

Ikemoto, S., Yang, C., Tan, A., 2015. Basal ganglia circuit loops, dopamine and motiva-

tion: a review and enquiry. Behav. Brain Res. 290, 17–31.

Ikonomidou, C., 2009. Triggers of apoptosis in the immature brain. Brain and

Development 31 (7), 488–492.

Ing, C., Brambrink, A.M., 2019. Mayo Anesthesia Safety in Kids continued: two new studies and a potential redirection of the ﬁeld. Br. J. Anaesth. 122 (6), 716–719. Ing, C., Ma, X., Sun, M., Lu, Y., Wall, M.M., Olfson, M., Li, G., 2020. Exposure to surgery and anesthesia in early childhood and subsequent use of attention deﬁcit hyper- activity disorder medications. Anesth. Analg. https://doi.org/10.1213/ANE. 0000000000004619. 31923004.

Istaphanous, G.K., Howard, J., Nan, X., Hughes, E.A., McCann, J.C., McAuliﬀe, J.J.,

Danzer, S.C., Loepke, A.W., 2011. Comparison of the neuroapoptotic properties of equipotent anesthetic concentrations of desﬂurane, isoﬂurane, or sevoﬂurane in neonatal mice. Anesthesiology 114 (3), 578–587.

Jevtovic-Todorovic, V., Benshoﬀ, N., Olney, J.W., 2000. Ketamine potentiates cere-

brocortical damage induced by the common anaesthetic agent nitrous oxide in adult rats. Br. J. Pharmacol. 130 (7), 1692–1698.

Jevtovic-Todorovic, V., Wozniak, D.F., Benshoﬀ, N.D., Olney, J.W., 2001. A comparative evaluation of the neurotoxic properties of ketamine and nitrous oxide. Brain Res. 895 (1–2), 264–267.

Jiang, Y., Tong, D., Hofacer, R.D., Loepke, A.W., Lian, Q., Danzer, S.C., 2016. Long-term

13

Neurotoxicology and Teratology 80 (2020) 106890

fate mapping to assess the impact of postnatal isoﬂurane exposure on hippocampal progenitor cell productivity. Anesthesiology 125 (6), 1159–1170.

Kanungo, J., Cuevas, E., Ali, S.F., Paule, M.G., 2013. Ketamine induces motor neuron toxicity and alters neurogenic and proneural gene expression in zebraﬁsh. J. Appl. Toxicol. 33 (6), 410–417.

Kesner, R.P., Lee, I., Gilbert, P., 2004. A behavioral assessment of hippocampal function

based on a subregional analysis. Rev. Neurosci. 15 (5), 333–351.

Kilb, W., Kirischuk, S., Luhmann, H.J., 2011. Electrical activity patterns and the func-

tional maturation of the neocortex. Eur. J. Neurosci. 34 (10), 1677–1686.

Lee, J.R., Lin, E.P., Hofacer, R.D., Upton, B., Lee, S.Y., Ewing, L., Joseph, B., Loepke, A.W., 2017. Alternative technique or mitigating strategy for sevoﬂurane-induced neuro- degeneration: a randomized controlled dose-escalation study of dexmedetomidine in neonatal rats. Br. J. Anaesth. 119 (3), 492–505.

Lerman, J., Johr, M., 2009. Inhalational anesthesia vs total intravenous anesthesia (TIVA)

for pediatric anesthesia. Paediatr. Anaesth. 19 (5), 521–534.

Lewin, G.R., Barde, Y.A., 1996. Physiology of the neurotrophins. Annu. Rev. Neurosci. 19,

289–317.

Macaya, A., Munell, F., Gubits, R.M., Burke, R.E., 1994. Apoptosis in substantia nigra

following developmental striatal excitotoxic injury. Proc. Natl. Acad. Sci. U. S. A. 91 (17), 8117–8121.

Matta, B.F., Heath, K.J., Tipping, K., Summors, A.C., 1999. Direct cerebral vasodilatory

eﬀects of sevoﬂurane and isoﬂurane. Anesthesiology 91 (3), 677–680.

Morris, R.G., Garrud, P., Rawlins, J.N., O'Keefe, J., 1982. Place navigation impaired in

rats with hippocampal lesions. Nature 297 (5868), 681–683.

Mullen, R.J., Buck, C.R., Smith, A.M., 1992. NeuN, a neuronal speciﬁc nuclear protein in

vertebrates. Development 116 (1), 201–211.

Na, H.S., Brockway, N.L., Gentry, K.R., Opheim, E., Sedensky, M.M., Morgan, P.G., 2017. The genetics of isoﬂurane-induced developmental neurotoxicity. Neurotoxicol. Teratol. 60, 40–49.

Nelson, A.J.D., Hindley, E.L., Vann, S.D., Aggleton, J.P., 2018. When is the rat retro-

splenial cortex required for stimulus integration? Behav. Neurosci. 132 (5), 366–377.

Olton, D.S., Walker, J.A., Gage, F.H., 1978. Hippocampal connections and spatial dis-

crimination. Brain Res. 139 (2), 295–308.

Osterop, S.F., Virtanen, M.A., Loepke, J.R., Joseph, B., Loepke, A.W., Vutskits, L., 2015.

Developmental stage-dependent impact of midazolam on calbindin, calretinin and parvalbumin expression in the immature rat medial prefrontal cortex during the brain growth spurt. Int. J. Dev. Neurosci. 45, 19–28.

Paule, M.G., Li, M., Allen, R.R., Liu, F., Zou, X., Hotchkiss, C., Hanig, J.P., Patterson, T.A., Slikker Jr., W., Wang, C., 2011. Ketamine anesthesia during the ﬁrst week of life can cause long-lasting cognitive deﬁcits in rhesus monkeys. Neurotoxicol. Teratol. 33 (2), 220–230.

Paxinos, G., Watson, C., 2014. The Rat Brain in Stereotaxic Coordinates, 7th ed. Elsevier. Pellegrini, L., Bennis, Y., Velly, L., Grandvuillemin, I., Pisano, P., Bruder, N., Guillet, B., 2014. Erythropoietin protects newborn rat against sevoﬂurane-induced neurotoxi- city. Paediatr. Anaesth. 24 (7), 749–759.

Perez-Zoghbi, J.F., Zhu, W., Grafe, M.R., Brambrink, A.M., 2017. Dexmedetomidine- mediated neuroprotection against sevoﬂurane-induced neurotoxicity extends to several brain regions in neonatal rats. Br. J. Anaesth. 119 (3), 506–516.

Poulter, M.O., Barker, J.L., O’Carroll, A.M., Lolait, S.J., Mahan, L.C., 1992. Diﬀerential and transient expression of GABAA receptor alpha-subunit mRNAs in the developing rat CNS. J. Neurosci. 12 (8), 2888–2900.

Ramachandra, R., Subramanian, T., 2011. Atlas of the Neonatal Rat Brain. CRC Press. Rice, D., Barone .Jr., S., 2000. Critical periods of vulnerability for the developing nervous system: evidence from humans and animal models. Environ. Health Perspect. 108 (Suppl. 3), 511–533.

Rizzi, S., Carter, L.B., Ori, C., Jevtovic-Todorovic, V., 2008. Clinical anesthesia causes permanent damage to the fetal guinea pig brain. Brain Pathol. 18 (2), 198–210. Sakata, K., Kito, K., Fukuoka, N., Nagase, K., Tanabe, K., Iida, H., 2019. Cerebrovascular reactivity to hypercapnia during sevoﬂurane or desﬂurane anesthesia in rats. Korean J Anesthesiol 72 (3), 260–264.

Scallet, A.C., Schmued, L.C., Slikker Jr., W., Grunberg, N., Faustino, P.J., Davis, H., Lester, D., Pine, P.S., Sistare, F., Hanig, J.P., 2004. Developmental neurotoxicity of ketamine: morphometric conﬁrmation, exposure parameters, and multiple ﬂuorescent labeling of apoptotic neurons. Toxicol. Sci. 81 (2), 364–370.

Schmued, L.C., Stowers, C.C., Scallet, A.C., Xu, L., 2005. Fluoro-Jade C results in ultra

high resolution and contrast labeling of degenerating neurons. Brain Res. 1035 (1), 24–31.

Shipley, M.T., Adamek, G.D., 1984. The connections of the mouse olfactory bulb: a study

using orthograde and retrograde transport of wheat germ agglutinin conjugated to horseradish peroxidase. Brain Res. Bull. 12 (6), 669–688.

Slikker Jr., W., Zou, X., Hotchkiss, C.E., Divine, R.L., Sadovova, N., Twaddle, N.C.,

Doerge, D.R., Scallet, A.C., Patterson, T.A., Hanig, J.P., Paule, M.G., Wang, C., 2007. Ketamine-induced neuronal cell death in the perinatal rhesus monkey. Toxicol. Sci. 98 (1), 145–158.

Sprung, J., Flick, R.P., Katusic, S.K., Colligan, R.C., Barbaresi, W.J., Bojanic, K., Welch, T.L., Olson, M.D., Hanson, A.C., Schroeder, D.R., Wilder, R.T., Warner, D.O., 2012. Attention-deﬁcit/hyperactivity disorder after early exposure to procedures requiring general anesthesia. Mayo Clin. Proc. 87 (2), 120–129.

Stratmann, G., Sall, J.W., May, L.D., Loepke, A.W., Lee, M.T., 2010. Beyond anesthetic properties: the eﬀects of isoﬂurane on brain cell death, neurogenesis, and long-term neurocognitive function. Anesth. Analg. 110 (2), 431–437.

Talpos, J.C., Chelonis, J.J., Li, M., Hanig, J.P., Paule, M.G., 2019. Early life exposure to extended general anesthesia with isoﬂurane and nitrous oxide reduces responsivity on a cognitive test battery in the nonhuman primate. Neurotoxicology 70, 80–90.

Todd, T.P., Fournier, D.I., Bucci, D.J., 2019. Retrosplenial cortex and its role in cue-

speciﬁc learning and memory. Neurosci. Biobehav. Rev. 107, 713–728.

S.M. Burks, et al.

von Bohlen und Halbach, O., 2011. Immunohistological markers for proliferative events, gliogenesis, and neurogenesis within the adult hippocampus. Cell Tissue Res. 345 (1), 1–19.

Vutskits, L., Culley, D.J., 2019. GAS, PANDA, and MASK: no evidence of clinical anes-

thetic neurotoxicity!. Anesthesiology 131 (4), 762–764.

Walters, J.L., Paule, M.G., 2017. Review of preclinical studies on pediatric general an- esthesia-induced developmental neurotoxicity. Neurotoxicol. Teratol. 60, 2–23. Walters, J.L., Chelonis, J.J., Fogle, C.M., Ferguson, S.A., Sarkar, S., Paule, M.G., Talpos, J.C., 2020. Acetyl-L-carnitine does not prevent neurodegeneration in a rodent model of prolonged neonatal anesthesia. Neurotoxicol. Teratol. https://www.ncbi.nlm.nih. gov/pubmed/32376384.

Wilder, R.T., Flick, R.P., Sprung, J., Katusic, S.K., Barbaresi, W.J., Mickelson, C., Gleich, S.J., Schroeder, D.R., Weaver, A.L., Warner, D.O., 2009. Early exposure to anesthesia and learning disabilities in a population-based birth cohort. Anesthesiology 110 (4), 796–804.

Workman, A.D., Charvet, C.J., Clancy, B., Darlington, R.B., Finlay, B.L., 2013. Modeling transformations of neurodevelopmental sequences across mammalian species. J. Neurosci. 33 (17), 7368–7383.

Wyss, J.M., Sripanidkulchai, B., 1985. The development of Ammon’s horn and the fascia dentata in the cat: a [3H]thymidine analysis. Brain Res. 350 (1–2), 185–198.

14

Neurotoxicology and Teratology 80 (2020) 106890

Zhang, X., Shen, F., Xu, D., Zhao, X., 2016. A lasting eﬀect of postnatal sevoﬂurane an- esthesia on the composition of NMDA receptor subunits in rat prefrontal cortex. Int. J. Dev. Neurosci. 54, 62–69.

Zheng, S.Q., An, L.X., Cheng, X., Wang, Y.J., 2013a. Sevoﬂurane causes neuronal apop- tosis and adaptability changes of neonatal rats. Acta Anaesthesiol. Scand. 57 (9), 1167–1174.

Zheng, H., Dong, Y., Xu, Z., Crosby, G., Culley, D.J., Zhang, Y., Xie, Z., 2013b. Sevoﬂurane anesthesia in pregnant mice induces neurotoxicity in fetal and oﬀspring mice. Anesthesiology 118 (3), 516–526.

Zhou, X., Li, W., Chen, X., Yang, X., Zhou, Z., Lu, D., Feng, X., 2016. Dose-dependent eﬀects of sevoﬂurane exposure during early lifetime on apoptosis in hippocampus and neurocognitive outcomes in Sprague-Dawley rats. Int J Physiol Pathophysiol Pharmacol 8 (3), 111–119.

Zou, X., Patterson, T.A., Divine, R.L., Sadovova, N., Zhang, X., Hanig, J.P., Paule, M.G.,

Slikker Jr., W., Wang, C., 2009. Prolonged exposure to ketamine increases neuro- degeneration in the developing monkey brain. Int. J. Dev. Neurosci. 27 (7), 727–731. Zurek, A.A., Yu, J., Wang, D.S., Haﬀey, S.C., Bridgwater, E.M., Penna, A., Lecker, I., Lei, G., Chang, T., Salter, E.W., Orser, B.A., 2014. Sustained increase in alpha5GABAA receptor function impairs memory after anesthesia. J. Clin. Invest. 124 (12), 5437–5441.