Abstract Oxidatively damaged DNA caused by reactive oxygen species (ROS) in the fetal brain contributes to neurodevelopmental disorders (NDDs), but the mechanism is unclear. We investigated the impact of this DNA damage on the developing fetal brain using a DNA repair-deficient oxoguanine glycosylase 1 (Ogg1) knockout (KO) mouse model, exposed during pregnancy to physiological (saline)- and ethanol (EtOH)-enhanced ROS levels. Oxidatively damaged DNA in saline-exposed Ogg1 KO vs. wild-type (WT) fetal brains was increased, and further enhanced by EtOH exposure. These OGG1-and EtOH-dependent patterns of DNA damage were reflected in: (a) increased gene dysregulation in saline-exposed KO brains, and greatly exacerbated by EtOH, notably in the long-term synaptic potentiation pathway, crucial for learning and memory; (b) impaired mitochondrial metabolism in cultured primary Ogg1 KO neurons; and, (c) sex-dependent learning and memory disorders. These results suggest oxidatively damaged DNA in DNA repair-deficient fetal brains contributes to NDDs by altering gene expression and mitochondrial metabolism. Keywords: Alcohol (ethanol, EtOH); Neurodevelopmental disorders; DNA damage and repair; Fetal alcohol spectrum disorders (FASD); Gene dysregulation; Knockout (KO) mouse model; Oxoguanine glycosylase 1 (OGG1); Reactive oxygen species (ROS) Graphical abstract [35]Image 1 [36]Open in a new tab Highlights * • Ogg1 KO fetal brains had more DNA damage and differentially expressed genes (DEGs). * • One prenatal ethanol (EtOH) exposure increased the above OGG1-dependent effects. * • Altered mitochondrial bioenergetics in neurons were OGG1-and EtOH-dependent. * • OGG1-and EtOH-dependent DEGs reflected postnatal disorders in learning and memory. * • OGG1 effects imply epigenetic roles of DNA damage in neurodevelopmental disorders. 1. Introduction Oxidative DNA damage by reactive oxygen species (ROS) in the fetal brain has been linked to several neurodevelopmental disorders (NDDs), but the underlying mechanisms remain unclear [[37]1]. ROS are small, highly reactive molecules derived from molecular oxygen, ranging from hydrogen peroxide (least reactive) and superoxide radicals to hydroxyl radicals (most reactive) [[38]2]. Low levels of ROS serve as physiologically important signalling molecules, but high ROS levels can cause pathogenic oxidative damage to DNA [[39]2]. The brain, particularly the fetal brain, is highly vulnerable to the negative effects of ROS due in part to high levels of ROS-forming enzymes, including NADPH oxidases (NOXs) [[40]3,[41]4] and prostaglandin H synthases (PHSs) [[42]5,[43]6]. Additionally, the brain has the highest rate of oxygen metabolism among organs, accounting for 20 % of all oxygen consumption, causing increased ROS generation during oxidative phosphorylation in the mitochondria [[44]7]. Conversely, the protective antioxidant system is relatively limited in the fetal brain due to lower activities of some antioxidative enzymes like catalase [[45]8,[46]9] and superoxide dismutase [[47]10], and antioxidants like glutathione (GSH), compared to the liver or kidney [[48]11]. Enhanced levels of ROS are associated with neurodegenerative disorders later in life, such as Alzheimer's [[49]12] and Parkinson's disease [[50]13]. Neurodegeneration is similarly enhanced in aged mice with deficiencies in enzymes important for antioxidative protection, such as glucose-6-phosphate dehydrogenase (G6PD) [[51]14,[52]15]. However, the impact of ROS, and more specifically ROS-initiated DNA damage, on the development and function of the fetal brain have not been fully elucidated. Oxidative damage to DNA sustained during important developmental windows may disrupt or dysregulate important processes like neurogenesis, neuronal differentiation and synaptogenesis, causing altered functional brain development and NDDs [[53]16,[54]17], with worse outcomes when DNA repair is deficient [[55]18,[56]19]. In fetal alcohol spectrum disorders (FASD), elevated ROS levels from alcohol (EtOH) metabolism and increased activity of ROS-forming enzymes like NOXs are likely contributing mechanisms of cognitive dysfunction [[57]4,[58]20]. Other NDDs with potential in utero origins include autism spectrum disorders (ASD), where postmortem human samples from ASD subjects showed excessive ROS production, decreased antioxidant capacity, and mitochondrial dysfunction [[59]21,[60]22]. Similarly, in attention deficit-hyperactivity disorders (ADHD), markers for oxidative stress are elevated in serum samples of children and adults diagnosed with ADHD [[61][23], [62][24], [63][25]]. Thus, ROS and oxidative stress may be critical contributors to the initiation of NDDs, warranting further characterization of the molecular mechanisms. Oxidative damage to DNA, as distinct from damage to proteins and lipids, can initiate developmental disorders [[64][26], [65][27], [66][28], [67][29], [68][30], [69][31]], as shown by their increase in knockout (KO) mice with deficiencies in various DNA damage recognition and repair enzymes [[70]4,[71][32], [72][33], [73][34], [74][35], [75][36], [76][37]]. The most common molecular lesion resulting from ROS-induced DNA damage involves oxidation of the C-8 residue on guanine, forming 8-oxo-7,8-dihydroguanine (8-oxoGua), due to the low oxidation potential of guanine [[77]38]. Repair of 8-oxoGua is primarily catalyzed by the base excision repair enzyme oxoguanine glycosylase 1 (OGG1) [[78]39]. The mitochondrial electron transport chain is a substantial source of endogenous ROS, as 1–3 % of electrons transferred through the chain escape and produce superoxide radicals [[79]40]. Other ROS sources proximate to the nucleus contributing to oxidatively damaged DNA include PHS and NOX isozymes [[80][3], [81][4], [82][5], [83][6]]. In utero exposure to EtOH increases NOX activity, thereby increasing ROS levels [[84]3,[85]41,[86]42]. Accumulation of oxidatively damaged DNA also causes DNA strand breaks (DSBs) [[87]43] leading to replication stress [[88]44], cell cycle arrest [[89]45], induction of apoptosis [[90]46], and alterations in epigenetic markers like DNA methylation [[91]47] and histone methylation and acetylation [[92]47,[93]48], resulting in morphological birth defects [[94]49,[95]50] and NDDs [[96]19,[97]48,[98]50]. It is unclear how oxidatively damaged DNA causes NDDs, but the well-known mutational role of 8-oxoGua in carcinogenesis is an unlikely mechanism. Rather, emerging evidence has demonstrated that 8-oxoGua, and its recruitment of OGG1, play a role in gene expression by functioning as an epigenetic marker [[99]51,[100]52]. Altered gene expression patterns from the loss of Ogg1 in the brain have been reported [[101]48,[102]53], with similar results in Ogg1 KO lung airway epithelial cells exposed to TGFβ1-induced ROS [[103]54]. Several epigenetic roles of 8-oxoGua have been proposed [[104]51], but one well-characterized mechanism involves the recruitment of NF-kB to 8-oxoGua formed within transcription factor binding sites via an interaction with OGG1 protein, resulting in gene transcription [[105][55], [106][56], [107][57]]. The roles of ROS-initiated 8-oxoGua and DSBs in physiological and dysregulated gene expression involved in normal and abnormal brain development are largely unexplored. The goal of this study was to investigate the molecular mechanisms by which oxidatively damaged DNA contributes to NDDs. In the fetal brain, we modeled enhanced oxidatively damaged DNA by using a DNA repair-deficient Ogg1 KO mouse model, without or with a single prenatal exposure to saline for physiological ROS levels or EtOH for enhanced ROS levels. In one cohort of mice, we measured molecular changes in fetal brain following prenatal saline or EtOH exposure, and a separate cohort was evaluated for NDD-associated phenotypes using a battery of behavioural tests ([108]Fig. 1A) The remarkable spectrum of dysregulated genes and mitochondrial dysfunction observed in DNA repair-deficient Ogg1 KO progeny, prenatally exposed to only saline, provides new insights into the molecular mechanisms by which even physiological levels of fetal ROS formation can lead to oxidatively damaged DNA that contributes to NDDs ([109]Fig. 2A). Furthermore, the distinct OGG1-dependent mechanisms of NDDs initiated by prenatal EtOH exposure reveal novel ways by which enhanced oxidative stress can dysregulate neurodevelopmental outcomes. Fig. 1. [110]Fig. 1 [111]Open in a new tab Prenatal EtOH dose optimization and oxidatively damaged DNA in the fetal brain in Ogg1 knockout (KO) mice. (A) Overview of treatment paradigm, and molecular and behavioural tests. (B) Blood concentrations following administration of the reactive oxygen species (ROS) enhancer ethanol (EtOH) were determined by high-performance liquid chromatography with flame ionization detection. Male Ogg1 mice received an i.p dose of either 4 g/kg EtOH or 2 doses of 2.5 g/kg EtOH administered 4 h apart. Blood samples were collected from 4 to 5 mice at each time point. (C) Representative histological analysis of hippocampal sections by H&E staining in young adult Ogg1 KO vs. wild-type (WT) progeny exposed to a single prenatal dose of EtOH (4 g/kg) or saline on gestational day (GD) 17. No gross morphological changes in the adult hippocampi were observed following EtOH exposure (n = 3, sex balanced). (D) Formation of the ROS-initiated DNA lesion 8-oxo-7,8-dihydroguanine (8-oxoGua) in fetal brains 6 h after maternal administration of EtOH (4 g/kg) or saline on GD 17. 8-OxoGua lesions determined by ELISA in whole brain homogenates (n = 5–6, sex balanced) were increased in an Ogg1-and EtOH-dependent manner. (E) ROS-initiated DNA strand breaks (DSBs) in GD 17 fetal brains 6 h after exposure to saline or various single doses of EtOH. Saline-exposed Ogg1 KO vs. WT brains exhibited increased DSBs (p < 0.05), while a 4 g/kg EtOH dose was required to increase DSBs in the brains of Ogg1 WT (p = 0.051) and Ogg1 KO (p < 0.001) fetuses, compared to respective saline (Sal) controls. (F) Images of representative comets for each treatment in Ogg1 WT and KO fetal brains. Statistical significance was determined by 2-way ANOVA and Tukey's post-hoc tests (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001). Fig. 2. [112]Fig. 2 [113]Open in a new tab Ogg1 KO fetal brains exposed prenatally to saline showed changes in gene expression compared to Ogg1 WT brains. (A) Overview of the postulated molecular mechanism of neurodevelopmental disorders (NDDs) caused by oxidatively damaged DNA. In the developing fetal brain, ROS-initiated DNA damage increased by deficient DNA repair or drug-enhanced ROS formation causes epigenetic changes and gene dysregulation, and/or mitochondrial dysfunction, adversely altering the postnatal development of brain function. Outcomes following prenatal saline treatment reflect the pathogenic potential of physiological ROS levels in DNA repair-deficient progeny. The effect of deficient DNA repair on gene expression in whole fetal brains of Ogg1 KO vs. WT progeny was measured using microarray analysis. Differentially expressed genes (DEGs) were defined by a minimal 2-fold change in expression and p < 0.05. (B) Microarray results showing the total number of DEGs in Ogg1 KO vs. WT fetal brains 6 h following saline or EtOH exposure (n = 3). The numbers in parentheses indicate the number of dysregulated genes. (C) Volcano plot showing DEGs for saline-exposed Ogg1 KO vs. WT fetal brains. The blue and red dots indicate significant decreases and increases in gene expression, respectively. (D) Table showing key genes that were dysregulated in saline-exposed Ogg1 KO vs. WT brains with roles in brain function. (E) Gene ontology (GO) pathway analysis showing pathways that are significantly represented in the list of DEGs in Ogg1 KO brains. Dysregulated genes were enriched in pathways involving kidney development and small nucleolar RNAs (snoRNA) in rRNA processing. The numbers in parentheses indicate the number of dysregulated genes/total genes in a pathway. 2. Materials and Methods 2.1. Animals and diet Heterozygous (+/-) Ogg1 mice used to establish our breeding colony were originally generated by Klungland and coworkers [[114]39]. They were engineered and generously gifted by Dr. Tomas Lindahl (Imperial Cancer Research Fund, UK), and provided by Dr. Christi Walter at the University of Texas Health Science Center at San Antonio. The Ogg1 strain background is comprised of 58% C57BL/6J and 42% 129SV (Bhatia and Wells, analysis by The Jackson Laboratory, unpublished). Housing details are outlined in the online Supplementary Materials. Colony breeder pairs consisted of an Ogg1 +/- male breeder and an Ogg1 +/- virgin female. Experimental mice consisted of Ogg1 +/+ males mated with Ogg1 +/+ females, and Ogg1 -/- males mated with Ogg1 -/- females, to maximize the number of KO and WT genotypes within a single litter. All animal protocols were approved by the institutional animal care committee and in conformance with the guidelines established by the Canadian Council on Animal Care. 2.2. Ogg1 genotyping Ear notches and fetal tail samples were used to identify the Ogg1 genotype. DNA extraction and PCR details are indicated in the Supplementary Materials. 2.3. Treatment paradigm To obtain the prenatal saline- and EtOH-exposed fetal brain samples, Ogg1 male and female mice were housed together overnight. Each morning, the females were checked for a visible copulation vaginal plug, and inseminated dams were weighed and placed into separate cages. This day was designated as GD 1, and females were monitored for signs of pregnancy. The timepoint of GD 17 was optimized by our lab to reproducibly produce brain functional disorders in saline-exposed Ogg1 KO progeny, and in Ogg1 progeny exposed to EtOH [[115]48]. An EtOH dose of 4 g/kg or 2 doses of 2.5 g/kg (4 h apart) were administered intraperitoneally (i.p.) to pregnant dams on gestational day (GD) 17. Treatment via the i.p. route rather than oral administration reduces the variability in plasma drug concentrations and in molecular and behavioural outcomes, optimizing reproducibility and the sensitivity for detecting differences among treatment groups in mechanistic studies. Treatment on a single gestational day enables a focused investigation into mechanisms of FASD initiation rather than confounding downstream processes. For most biochemical studies, EtOH was administered as a single dose of 4 g/kg, which we have shown causes NDDs similar to human FASD in several mouse models without causing apparent maternal or fetal acute toxicity [[116]19,[117]26], and results in maximal oxidatively damaged DNA (8-oxoGua, DSBs) in fetal brains 6 h following maternal administration [[118]58]. A single or daily 4 g/kg dose of EtOH is commonly used in mouse models, and single or daily doses of 5–6 g/kg via gavage are often used to model human binge exposure, characterized as 4 or more alcoholic drinks per occasion [[119]59,[120]60]. A 2019 self-reporting study of 6814 pregnant women by the Centers for Disease Control and Prevention found that 11.5% of pregnant women had at least 1 alcoholic drink in the previous 30 days, and 3.9 % had at least one episode (average of 4.5 episodes) of binge drinking [[121]59]. Moreover, the incidence of binge drinking was 5.1% in some racial groups and 5.8 % in pregnant women ages 18–24, while the incidence in unmarried pregnant women (6.1 %) was almost triple that in married women (2.2 %). For the behavioural studies of NDDs, due to unexpected cannibalization of pups by some dams treated with the 4 g/kg dose, and based in part upon the results of blood EtOH concentrations shown in [122]Fig. 1B, EtOH was administered as a divided dose of 2 × 2.5 g/kg given 4 h apart on GD 17. While this dosing regimen produced lower peak blood EtOH concentrations than the single 4 g/kg dose ([123]Fig. 1B), it nevertheless caused drug- and OGG1-dependent NDDs. 2.4. Blood EtOH concentration measurement Ogg1 male mice were treated i.p. with EtOH, either as a single 4 g/kg dose, or with 2 × 2.5 g/kg doses administered 4 h apart. Mice were anesthetized using isoflurane for cardiac puncture blood sampling. Blood samples at each timepoint were collected from 4 to 5 mice in heparinized vacutainers (lithium heparin 68 USP units per tube; Becton–Dickinson, Oakville, ON, Canada) at 0, 0.5, 1, 3, 5 and 7 h. Blood EtOH concentrations were measured by the British Columbia Provincial Toxicology Centre using headspace gas chromatography with flame ionization detection. Ethanol was measured using an Agilent 7697A headspace autosampler with an Agilent 9000 Intuvo gas chromatography system equipped with dual flame ionization detectors. Separation was achieved on DB-BAC1 (30 m × 320 μm × 1.8 μm) and DB-BAC2 (30 m × 320 μm × 1.2 μm) columns with helium replacing nitrogen as the carrier gas. Tertiary butanol was used as the internal standard. The method ([124]https://www.agilent.com/cs/library/applications/5991-7217EN.pdf) was validated in accordance with the Academy Standards Board (ASB) of the American Academy of Forensic Sciences: ASB 036: Standard Practices for Method Validation in Forensic Toxicology and ISO/IEC 17025: General Requirements for the Competence of Testing and Calibration Laboratories. 2.5. Histology Histological analyses were carried out to investigate gross morphological changes from prenatal EtOH exposure in the adult hippocampus. Following the behavioural studies, Ogg1 progeny (approximately 4 months of age) were sacrificed, and brain tissues were collected (n = 3) and fixed in 10 % neutral buffered formalin (Sigma-Aldrich, Burlington, MA) for 48 h. Brain tissues were embedded in paraffin wax, sectioned, stained with hematoxylin and eosin, and imaged by the CILIA histology core at the Toronto Princess Margret Hospital. 2.6. Measurement of the ROS-initiated DNA lesion 8-oxoGua by ELISA The formation of 8-oxoGua in whole fetal brain DNA was quantified using ELISA. Fetal brains exposed in utero to saline or 4 g/kg EtOH were collected 6 h after maternal treatment. Nuclear and mitochondrial DNA were isolated using a protocol to reduce artifactual oxidation [[125]61,[126]62]. Fetal brain tissue was homogenized in 1 mL cold lysis buffer (320 mM sucrose, 5 mM MgCl2, 10 mM Tris, 0.1 mM deferoxamine, 1 % Triton X-100, pH 7.5) and centrifuged twice at 1000×g for 10 min at 4 °C to isolate the nuclear pellet containing DNA. The pellet was then incubated with enzyme reaction solution (1 % w/v SDS, 5 mM EDTA-Na2, 0.15 mM deferoxamine 10 mM Trish-HCl, pH 8.0) for 1 h at 50 °C with RNase A/T1 mix (624 and 312 U/mL final activity, respectively). Proteinase K (13.3 μL of 10 mg/ml solution) was added to the samples and samples were left overnight at 50 °C. The next morning, 300 μl of NaI solution (7.6 M NaI, 40 mM Tris, 20 mM EDTA-Na2, 0.3 mM deferoxamine, pH 8.0) was added to each sample followed by 500 μl of 100 % isopropanol to precipitate the DNA. Samples were then centrifuged at 10,000×g for 10 min to obtain the pellet. The DNA pellet was washed 5 times with 70 % EtOH. DNA was resuspended in 135 μL of TE buffer (1 mM EDTA in 10 mM Tris-HCL, pH 8.0). To aid resuspension, DNA was left at 4 °C overnight and heated at 60 °C for 1–2 h. DNA concentrations were measured using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Waltham, MA). 8-Oxo-dGua was detected using an 8-OH-dGua Check ELISA kit (JaICA, KOG-HS10E, Shizuoka, Japan) according to the manufacturer's protocol. A 50 μg aliquot of DNA from each sample was digested into single nucleotides with nuclease P1 (6 U/sample, 1 h at 37 °C), and 15 μL of 200 mM sodium acetate, pH 5.2, was added to each sample to lower the pH. The pH was increased with 15 μL of Tris-HCL buffer (1 M, pH 7.4), and samples were incubated with calf intestinal alkaline phosphatase (6 U/sample, 1 h at 37 °C) to remove the phosphate group. The reactions were filtered and concentrated using Amicon Ultra filter units (YM-10, 10,000 MW cutoff, Millipore, Billerica, MA, USA) for 30 min to 1 h. A total of 6 fetal brains, balanced for sex, were used for each genotype and treatment group, except for the EtOH-exposed Ogg1 WT group, which had 5 samples. Data were analyzed using 2-way ANOVA with multiple comparisons and Tukey's correction for multiple hypothesis testing. 2.7. Comet assay DNA strand breaks (single- and double-strand breaks) were detected using single-cell gel electrophoresis (comet assay) with neutral conditions, as previously described [[127]63,[128]64]. Ogg1 WT and KO dams were treated i.p. on GD 17 with a single dose of saline or EtOH (2, 3 or 4 g/kg) and whole fetal brains were collected 6 h later and snap frozen in liquid nitrogen. Within a month of collection, whole fetal brains were thawed on ice and homogenized in homogenization buffer (Hanks' Balanced Salt Solution, Ca++, Mg++ free, and phenol free, pH 7.5, 20 mM Na[2]EDTA, with 10 % DMSO added prior to use). The homogenized mixture was pushed through a 26 ½ gauge needle three times to obtain dissociated cells. A cell suspension containing single cells was mixed with 1 % low melting agarose (prepared in Dulbecco's phosphate-buffered saline, Ca++, Mg++ free, and phenol free, pH 7.4) and layered on a microscopy slide (CA48323-190L, VWR), precoated with 1.5 % normal melting agarose (dissolved in ddH2O). The cells were lysed overnight (∼22–24 h) in a lysis buffer (100 mM Na[2]EDTA, 2.5 M NaCl, 10 mM Tris-HCl, pH 10, with 1 % Triton X-100, and 10 % DMSO added prior to use) at 4 °C. Slides were then incubated in a neutral solution (100 mM Tris, 300 mM sodium acetate, pH 8.3) for 1 h at 4 °C. Electrophoresis was then conducted using neutral buffer at ∼3 mV/cm, 300 mAmp for 40 min at 4 °C to separate the damaged DNA fragments. Slides were dehydrated in 100 % ethanol for 5 min, and air dried for about 30 min. DNA was stained and incubated with SYBR Gold stain for ∼30 min. The comets were visualized using a fluorescence microscope (BioTek Cytation 5 Cell Imaging Multimode Reader) and analyzed for the percentage of DNA in the tail using the CometScore™ Freeware v2.0. For each sample, 50 comets from two technical replicate slides each were analyzed and averaged, n = 3, sex not determined, from at least 3 different litters to minimize litter effects. Data were analyzed using 2-way ANOVA with multiple comparisons and Tukey's correction for multiple hypothesis testing. 2.8. RNA extraction and microarray analysis On GD 17, dams were treated i.p. with 4 g/kg EtOH or saline, and whole fetal brains were collected 6 h later and snap frozen in liquid nitrogen. Three female biological replicates from each genotype and treatment from at least 3 different litters were analyzed. RNA isolation is outlined in the online Supplementary Materials. Microarray analysis (RNA labelling, hybridization, staining, scanning and background correction) was carried out by The Centre for Applied Genomics (Toronto, ON) using Affymetrix GeneChip Mouse 2.0 ST arrays. Details for analysis of the raw data are in the online Supplementary Materials. Raw data were analyzed using Transcriptome Analysis Console (Version 4.0.3.1.4, 2022, Thermo Fisher, Waltham, MA) with RMA normalization. Differential expression analysis was performed using Linear Models of Microarray Data (LIMMA) with the Benjamini-Hochberg procedure to control the false discovery rate (FDR) from multiple hypothesis testing. However, for the saline-exposed Ogg1 WT vs. KO gene set, a non-adjusted p-value was used to identify key dysregulated genes. Differentially expressed genes (DEGs) were defined by a ≥ 2-fold change. Only annotated genes were analyzed. Gene ontology (GO) pathway overrepresentation analysis and transcription factor binding analysis were carried out using g:Profiler following a protocol described elsewhere [[129][65], [130][66], [131][67]]. The following parameters were used for the g:Profiler analysis: organism “mus musculus (mouse)”, highlighted driver terms in GO, statistical domain scope “only annotated genes”, significance threshold “g:SCS threshold” at user threshold “0.05” and numeric IDs treated as “ENTREZGENE_ACC”. All data sources were used and included GO molecular function, GO cellular component, GO biological process, no electronic GO annotations, KEGG, reactome, wikipathwyas, TRANSFAC, miRTarBase, human protein atlas, CORUM and HP. 2.9. Quantitative PCR (qPCR) qPCR was performed using a TaqMan Gene Expression Assay (Thermo Fisher Scientific). Reactions were set up in 20 μL volumes containing 10 μL of TaqMan Universal PCR Master Mix (Thermo Fisher Scientific, 4444557), 1 μL of TaqMan Gene Expression Assay mix (containing the primer and probe specific for the target gene: Acot10 [4448892], Acot2 [4448892], Gabra2 [4448892] and Slc18 [4448892]), 2 μL of diluted cDNA, and 7 μL of nuclease-free water. Relative gene expression levels were calculated using the ΔΔCt method, where the threshold cycle (Ct) values of target genes were normalized to an endogenous control gene, GAPDH (Thermo Fisher Scientific, 4453320). 2.10. Primary cell culture Primary cortical and hippocampal neuronal cell cultures were prepared from GD 17 mouse fetuses. Tissue isolation and cell culture details are outlined in the online Supplementary Materials. 2.11. Mitochondrial stress test Primary cortical and hippocampal cells were collected from GD 17 fetal brains and cultured in a 96-well Seahorse plate (Agilent, 103725-100, Santa Clara, CA) coated with laminin/poly-l-lysine. The cells were cultured for 14 days to allow for differentiation into mature neurons. A mitochondrial stress test was carried out using a XF96 Seahorse Metabolic Analyzer (Agilent, Santa Clara, CA) [[132]68] and the oxygen consumption rate was recorded (n = 5 for each group). The ports and injection times were used per the manufacturer's protocol (Agilent, Santa Clara, CA). Various inhibitors were used to assess different components of mitochondrial metabolism. Carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) is an uncoupler of oxidative phosphorylation. It dissipates the proton gradient across the mitochondrial inner membrane by allowing protons to flow freely back into the mitochondrial matrix, preventing ATP synthesis. FCCP is used to measure the maximum respiratory capacity of the mitochondria. Oligomycin is an inhibitor of ATP synthase (complex V), and prevents protons from returning to the mitochondrial matrix. Oligomycin, an ATP synthase inhibitor, is used to determine the portion of mitochondrial respiration coupled to ATP production. Antimycin A and rotenone inhibit complexes III and I of the electron transport chain, respectively. Inhibiting the electron transport chain reduces oxygen consumption to determine the spare respiratory capacity and non-mitochondrial respiration. Each well was normalized to cell count, using a Hoechst stain (Thermo Scientific, 62249, Waltham, MA) and the Cytation 5 Cell Imaging Multimode Reader (BioTek, Winooski, VT). 2.12. Behavioural testing Pregnant Ogg1 dams were treated i.p. on GD 17 with 2 doses of saline or 2.5 g/kg EtOH, 4 h apart, beginning at about 10 a.m. In the behavioural studies, this EtOH dosing regimen was used because we found that a single 4 g/kg EtOH dose was unexpectedly associated with some cannibalization when pups were delivered on GD 21. Based in part on the results of blood EtOH concentrations ([133]Fig. 1B), we chose to administer 2 × 2.5 g/kg EtOH (4 h apart). While this dosing regimen produced lower peak blood EtOH concentrations than the single 4 g/kg dose ([134]Fig. 1B), it nevertheless caused drug- and OGG1-dependent NDDs. The pups were allowed to deliver spontaneously and were weaned on postnatal day 21. Three behavioural tests for learning and memory were performed at different postnatal weeks noted in their respective Methods subsection immediately below. At least 3 litters were used in each group to minimize litter effects. On the day of testing, the mice were habituated in the room for at least 1 h. The female estrus cycle has known effects on learning and memory [[135]69], so female mice were visually monitored for estrus [[136]70] and tested when the cycle was complete. All behavioural testing was performed during the daylight cycle between the hours of 10 a.m. and 4 p.m. The individual performing the tests was blinded to the genotype and treatment group. 2.13. Puzzle box test The puzzle box test was performed on 8-week-old mice, with at least 3 litters tested for each group. The puzzle box test protocol has been previously described and was used here with minor modifications [[137]71]. Testing details are outlined in the online Supplementary Materials. For each trial, the time to task completion was analyzed by 2-way ANOVA with multiple comparisons assessed by post-hoc Tukey's tests. Sex differences for each trial were evaluated by 1-way ANOVA with Fisher's LSD test. 2.14. Fear conditioning test Fear conditioning assessment of naive mice was determined at 10 weeks of age using progeny from at least 3 litters for each group. Mice were tested using a cued and contextual fear conditioning system (Ugo Basile, 46000), and video recordings were analyzed with ANYmaze software (V7.1, Stoelting Co.). The fear conditioning protocol is shown in [138]Fig. 7A, and conducted following published protocols [[139]72,[140]73]. Testing details are outlined in the online Supplementary Materials. In females and males, the freezing time was analyzed by 2-way ANOVA with multiple comparisons assessed by post-hoc Tukey's tests. Fig. 7. [141]Fig. 7 [142]Open in a new tab Female Ogg1 KO mice showed enhanced OGG1-and/or EtOH-dependent freezing behaviour and/or reduced passive avoidance performance. (A) Schematic of the fear-conditioning paradigm conducted over 2 days. Fear memory behaviour was tested using contextual and conditioned stimulus cues. Audio, visual and olfactory cues were changed based on testing the context vs. auditory conditioned cue fear memory retrieval. Further details are provided in the Methods. (B) The saline-exposed Ogg1 KO vs. WT female progeny displayed increased freezing time in the novel context control test (p < 0.01) and auditory cue retrieval test (p < 0.05), which was not seen in the respective males, indicating a sex effect. With prenatal EtOH exposure, female Ogg1 KO vs. WT female progeny displayed increased freezing time in the auditory cue test (p < 0.05), and similar but non-significant increases in the context retrieval and novel context control tests, which again were not seen in the respective males, indicating a sex effect. Progeny were selected from at least 3 different litters, and the number of mice tested in each group is indicated in parentheses. Statistical significance was determined by 2-way ANOVA and Tukey's post-hoc tests (∗p < 0.05, ∗∗p < 0.01). 2.15. Passive avoidance test The passive avoidance test was performed on 9-week-old mice from at least 3 different litters. The apparatus was a rectangular box with a light chamber and a red cellophane chamber that appears dark to the mouse. The chambers were separated by a manual guillotine door. The apparatus floor was comprised of metal rods to deliver a foot shock. Testing details are outlined in the online Supplementary Materials. Data for days 1 and 2 for females and males are shown in [143]Suppl. Fig. 3. The latency time to enter was analyzed using 2-way ANOVA with multiple comparisons and Tukey's correction for multiple hypotheses testing on each day in females and males. 2.16. Statistical analysis Statistical analysis was performed using GraphPad Prism 10 (Version 10.2.3, GraphPad Software Inc. San Diego, CA). The data are represented as the mean ± SD (standard deviation), as indicated in the figures. The sample numbers (n) are indicated in each figure and/or their legend. Data were analyzed using 2-way ANOVA with multiple comparisons assessed by post-hoc Tukey's tests, unless otherwise specified. Differences in means were considered statistically significant at p < 0.05. Significance levels are as follows: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. 3. Results 3.1. Overview of EtOH dosing regimens for biochemical and behavioural studies A 4 g/kg dose of EtOH was administered intraperitoneally (i.p.) to pregnant dams on gestational day (GD) 17 (see the Methods and Materials for details and rationale). For biochemical studies, EtOH was administered primarily as a single 4 g/kg dose, which reproduces NDDs similar to human FASD in several mouse models [[144]26], and maximal 8-oxoGua formation in fetal brains occurs 6 h following maternal administration [[145]58]. For the behavioural studies of NDDs, due to unexpected cannibalization of pups by some dams, and based in part upon the results of blood EtOH concentrations shown in [146]Fig. 1B, EtOH was administered as a divided dose of 2 × 2.5 g/kg given 4 h apart on GD 17. Although this dosing regimen produced lower peak blood EtOH concentrations than the single 4 g/kg dose ([147]Fig. 1B) used in the biochemical studies, it nevertheless caused drug- and OGG1-dependent NDDs without causing apparent acute maternal or fetal toxicity. 3.2. Study design and optimization of EtOH exposure in the Ogg1 fetal brain model We studied the impact of ROS during brain development by exposing DNA repair-deficient, Ogg1 homozygous (−/−) KO and congenic Ogg1 WT (+/+) fetuses with normal DNA repair to saline or the ROS-initiating drug EtOH ([148]Fig. 1A). In addition to serving as the vehicle control for EtOH, saline was used to assess the pathogenic potential of physiological levels of ROS formation in Ogg1 KO vs. WT progeny. To determine a maternal EtOH dose that caused sufficient oxidative DNA damage in fetal brains, without causing apparent maternal or fetal toxicity, we tested 2 dosing regimens based on the literature, including our own studies in Ogg1 KO mice. In the Ogg1 strain, a single i.p. dose of 4 g/kg produced a peak blood EtOH concentration of 504 ± 89 mg/dL ([149]Fig. 1B). Administration of 2 doses of 2.5 g/kg EtOH, 4 h apart, produced a peak of 265 ± 40 mg/dL. High prenatal exposures of EtOH are known to cause alterations in hippocampal gross morphology, which in turn can affect functions in memory [[150]74]. Gestational day (GD) 17 exposure of progeny to maternal EtOH, 2 × 2.5 g/kg, did not grossly alter adult hippocampal morphology (n = 3) ([151]Fig. 1C). 3.3. Increased oxidatively damaged DNA in Ogg1 KO fetal brains following prenatal saline exposure, and exacerbation by EtOH exposure 8-oxoGua is the most prevalent ROS-initiated lesion, and it is widely used as a biomarker for oxidatively damaged DNA [[152]75]. With an ELISA, we assessed 8-oxoGua levels in fetal brains 6 h after saline or a single EtOH (4 g/kg) exposure, which was previously determined to be the time of peak 8-oxoGua formation caused by a ROS-inducing drug ([153]Fig. 1D) [[154]58]. A 2-way ANOVA showed significant effects of both Ogg1 genotype (F(1,19), p < 0.0001) and EtOH (F(3,19), p < 0.0001) on 8-oxoGua formation. The interaction between the Ogg1 genotype and EtOH trended towards significance (p = 0.054). A Tukey's post hoc test determined saline-exposed Ogg1 KO fetal brains had 31 % more 8-oxoGua lesions than brains of saline-exposed Ogg1 WT progeny (p < 0.001). EtOH exposure increased 8-oxoGua lesions by 58 % in Ogg1 WT (p < 0.0001) and 24 % in KO (p < 0.0001) fetal brains compared to brains of the respective genotypes exposed to saline. The formation of 8-oxoGua lesions can lead to both single and double DNA strand breaks [[155]9,[156]76], which we characterized using single-cell gel electrophoresis, known as the “comet assay” [[157]63] ([158]Fig. 1E). Using neutral conditions, we detected DNA strand breaks (DSBs) in single cells by measuring the percentage of DNA in the comet tail in cells from fetal brains 6 h after saline or EtOH exposure. A 2-way ANOVA showed that both OGG1 deficiency in fetal brain (F(3,17), p < 0.001) and EtOH exposure (F(1,19), p < 0.001) affected DSBs, as did the interaction between Ogg1 genotype and EtOH (F(3,17), p < 0.001). There was a 61.5 % increase in DSBs in Ogg1 KO vs. WT brains exposed only to saline (p < 0.05), consistent with the 8-oxoGua ELISA results. On GD 17, a maternal dose of 4 g/kg EtOH was necessary to increase DBSs in Ogg1 WT fetal brains (trending to significance, p = 0.050) and Ogg1 KO brains (p < 0.001), compared to brains of the respective genotypes exposed to saline. These findings confirm that: (a) deficient DNA repair due to the loss of Ogg1 in fetal brain leads to increased oxidatively damaged DNA from physiological levels of ROS; and (b) the ROS-initiating drug EtOH increased oxidatively damaged DNA in fetal brain. Images of representative comets for each treatment in Ogg1 WT and KO fetal brains are shown in [159]Fig. 1F. 3.4. Gene dysregulation in the fetal brains of female Ogg1 KO progeny In addition to causing mutations, 8-oxoGua can function as an epigenetic mark in regulatory elements in promoters, thereby regulating gene expression [[160]51]. To determine if the OGG1-and EtOH-dependent increases in 8-oxoGua levels are associated with altered gene expression, we characterized the transcript profile in female fetal brains using a microarray approach ([161]Fig. 2A). Differentially expressed genes (DEGs) were selected based upon a minimal 2-fold change in expression and a false discovery rate (FDR)-adjusted p-value of <0.05. The total number of DEGs for each group is plotted in [162]Fig. 2B. We used qPCR to validate significant changes in gene expression observed in the microarray studies. The representative key genes Acot10, Acot2, Gabra2 and Slc18 were chosen for validation based on their degree of dysregulation and/or role in brain development ([163]Suppl. Fig. 1). 3.4.1. Physiological ROS levels Based upon the standard FDR correction, there were no statistically significant differences in gene expression in the saline-exposed Ogg1 KO (−/−) vs. WT (+/+) brains exposed to saline. However, based on: (a) similar patterns of OGG1-dependent gene expression below for EtOH; and, (b) preliminary results for saline from a similar study in DNA repair-deficient Brca1 KO mice [[164]77], the standard number of brains (n = 3) used in this study may have been insufficient for the variability in the data for saline-exposed brains. Accordingly, the small sample size may have limited the ability to detect subtle or less common patterns, which could be more evident with a larger cohort. We therefore also conducted a less stringent analysis using a non-FDR-adjusted p-value (p < 0.05) for saline-exposed brains to analyze genes exhibiting a 2-fold or greater dysregulation ([165]Suppl. Table 1), although this less stringent approach increases likelihood of false positives. In fetal brains of saline-exposed Ogg1 KO vs. WT fetal brains, 47 genes were upregulated, and 17 genes were downregulated ([166]Fig. 2C). Dysregulated genes with known roles in brain function are summarized in [167]Fig. 2D. A gene ontology (GO) pathway analysis of the dysregulated genes indicated enrichment for pathways involving kidney development and small nucleolar RNAs (SnoRNA) involved in rRNA processing ([168]Fig. 2E). A rigorous confirmation of these results will require further analysis using a greater number of brain samples, which may also reveal more genes dysregulated by the loss of OGG1. These results suggest that the loss of OGG1 and a resulting increase in oxidatively damaged DNA even at physiological ROS levels may adversely impact gene expression and cause NDDs. 3.4.2. EtOH-enhanced ROS levels Next, we characterized the impact of EtOH-enhanced levels of ROS formation and oxidatively damaged DNA on gene expression in the Ogg1 WT and KO fetal brains. Gene expression levels were normalized to saline exposure for each genotype, and significant changes were determined by the stringent FDR-adjusted p-value for genes exhibiting a 2-fold or greater dysregulation. In Ogg1 WT fetal brains, EtOH exposure caused the upregulation of 380 genes and the downregulation of 8 genes, compared to saline exposure ([169]Fig. 3A). Compared to WT brains, EtOH exposure in the Ogg1 KO brains resulted in 51 % more DEGs, 523 of which were upregulated and 63 were downregulated ([170]Fig. 3B). The top 100 genes that were dysregulated in Ogg1 KO fetal brains with EtOH exposure are listed in [171]Suppl. Table 2. Notably, an overall effect of gene upregulation was seen in both Ogg1 WT and KO fetal brains with EtOH exposure. Selected representative DEGs with known roles in brain development or oxidative stress responses are listed in [172]Fig. 3C. Given that non-coding microRNAs (miRNAs) are known to be affected by EtOH exposure in fetal brains in FASD studies [[173]78,[174]79], we classified the types of DEG using sequence ontology analysis. Among the DEGs, 74 % and 73 % were protein-coding transcripts in Ogg1 WT and KO fetal brains, respectively, while non-coding transcripts (including lncRNA, miRNA and rRNA) accounted for 16 % and 19 % of DEGs in Ogg1 WT and KO fetal brains, respectively ([175]Fig. 3D and 3E). Remarkably, miRNAs dysregulated in the Ogg1 fetal brains did not overlap with human maternal plasma miRNA used to predict FASD outcomes [[176][78], [177][79], [178][80]]. A GO pathway enrichment analysis revealed similarities in enriched pathways in WT and KO fetal brains, including genes/proteins with functions in the extracellular region and RNAi effector complex ([179]Fig. 3F). Interestingly, the long-term synaptic potentiation pathway, a process involved in learning and memory, was enriched only in the Ogg1 KO fetal brains with EtOH exposure. Taken together, these findings show that the loss of OGG1 can impact the expression of genes in response to EtOH exposure in the fetal brain. Fig. 3. [180]Fig. 3 [181]Open in a new tab Ogg1 KO fetal brains exposed prenatally to EtOH showed exacerbated changes in gene expression compared to EtOH-exposed Ogg1 WT brains. The effect of prenatal EtOH exposure on gene expression 6 h later on GD 17 in whole fetal brains was measured using microarray analysis (n = 3). DEGs were defined by a minimal 2-fold change in expression and a false discovery rate (FDR)-adjusted p-value < 0.05. Volcano plots show DEGs for EtOH vs. saline exposure in (A)Ogg1 WT brains and (B)Ogg1 KO brains. The blue and red dots respectively indicate significant decreases and increases in gene expression. (C) Table for EtOH-exposed Ogg1 KO vs. WT brains showing key dysregulated genes with roles in pathways relevant to oxidatively damaged DNA and neurodevelopment. (D) Analysis of DEGs by genome feature type using the Sequence Ontology database in Ogg1 WT brains, and (E)Ogg1 KO brains, each exposed to EtOH vs. saline. Ogg1 KO brains had 4 % more miRNA dysregulated than Ogg1 WT brains (non significant). (F) Gene Ontology (GO) pathway analysis shows pathways that are significantly represented in the list of DEGs in Ogg1 WT Ogg1 KO fetal brains exposed to EtOH vs. saline. Overlapping enriched pathways include genes with protein function in the extracellular region and RNAi effector complex. The numbers in parentheses indicate the number of dysregulated genes/total genes in a pathway. Notably, a pathway unique to Ogg1 KOs involves genes with roles in long-term synaptic potentiation, an important mechanism in learning and memory. 3.5. Long-term potentiation gene expression pathway is predominantly, if not exclusively, altered in Ogg1 KO fetal brains Our aim here was to distinguish OGG1-dependent mechanisms from EtOH-dependent mechanisms causing altered gene expression. We identified genes that were dysregulated in both Ogg1 KO and WT fetal brains exposed to EtOH, suggesting a common EtOH effect ([182]Fig. 4A). Furthermore, these results also identified uniquely dysregulated genes in Ogg1 KO vs. WT brains ([183]Fig. 4A). Genes that are shared or uniquely dysregulated in Ogg1 WT and KO fetal brains are listed in [184]Suppl. Table 3. A transcription factor (TF) binding analysis was used to detect transcription factor motifs overrepresented in each gene set. Remarkably, the loss of Ogg1 reduced the number of enriched transcription factor motifs in EtOH-exposed progeny from 22 in the Ogg1 WT fetal brains to 4 in the Ogg1 KO brains ([185]Fig. 4B); namely, Chicken Ovalbumin Upstream Promoter-Transcription Factor II (Arp-1), Glucocorticoid Receptor (Gr), Progesterone Receptor (Pr), and Pituitary-specific positive transcription factor 1 (Pit-1) ([186]Suppl. Table 3). This provides new potential targets for 8-oxoGua/OGG1/TF interactions leading to altered gene expression potentially relevant to NDDs. Considering the role of OGG1 in the epigenetic function of 8-oxoGua, we investigated if the Ogg1 KO brains responded differently than WT brains to EtOH exposure. A GO pathway analysis was performed for the following three gene sets: unique DEGs for Ogg1 WT fetal brains with EtOH ([187]Fig. 4C), unique DEGs for Ogg1 KO brains with EtOH ([188]Fig. 4D) and shared DEGs ([189]Fig. 4E). The shared DEGs had roles in cell signaling and in the extracellular region and RNAi effector complex. When comparing enriched pathways between Ogg1 WT vs. KO fetal brains, genes involved in long-term synaptic potentiation were uniquely dysregulated in Ogg1 KO brains. The specific genes that are dysregulated in each enriched pathway are listed in [190]Suppl. Tables 4–6. Fig. 4. [191]Fig. 4 [192]Open in a new tab Bioinformatic analysis of Ogg1-and EtOH-dependent effects on gene expression in fetal brains. (A) Venn diagram showing DEGs that are either overlapping or unique to Ogg1 WT or Ogg1 KO fetal brains 6 h after exposure to EtOH or saline on GD 17. (B) Transcription factor (TF) binding site enrichment analysis within the gene sets revealed that Ogg1 KO brains exhibit only 4 TF sites compared to 22 sites in Ogg1 WT brains. This provides new potential 8-oxoGua/OGG1/TF interactions leading to altered gene expression. (C) GO pathway enrichment analysis of DEGs unique to Ogg1 WT brains have roles in the immune response with antigen-binding activity, among other pathways. (D) GO pathway enrichment analysis of DEGs unique to Ogg1 KO brains have roles in long-term synaptic potentiation, among other pathways. (E) GO pathway enrichment analysis of shared EtOH-dependent DEGs have roles in the extracellular region and the RNAi effector complex. The numbers in parentheses indicate the number of dysregulated genes/total genes in a pathway. 3.6. Comparison of OGG1-dependent dysregulation of genes in fetal brains exposed to saline with brains exposed to EtOH Among the genes apparently dysregulated in saline-exposed Ogg1 KO vs. WT fetal brains (using the less stringent standard p-value), the majority were not dysregulated by EtOH exposure ([193]Suppl. Table 7). This suggests that physiological ROS levels in DNA repair-deficient fetal brains may uniquely influence specific gene responses. 3.7. Impaired mitochondrial metabolism in Ogg1 KO cells In addition to nuclear DNA, the 8-oxoGua lesion can accumulate in mitochondrial DNA (mtDNA), resulting in mitochondrial dysfunction [[194]81,[195]82]. We investigated the effect of physiological and EtOH-enhanced ROS levels on mitochondrial bioenergetics by conducting a mitochondrial stress test in cultured Ogg1 KO (−/−) vs. WT (+/+) primary neurons using a Seahorse XFe96 extracellular flux analyzer [[196]83,[197]84]. Primary cortical neurons were obtained from GD 17 Ogg1 WT and KO fetal brains, cultured in vitro for 14 days and imaged at 8 days in vitro (DIV) ([198]Fig. 5A). We measured oxidative phosphorylation capacity by evaluating the oxygen consumption rate (OCR) before and after the addition of a series of mitochondrial complex inhibitors (see Methods for details). The OCR trace ([199]Fig. 5B) was used to quantify key parameters of mitochondrial function. Fig. 5. [200]Fig. 5 [201]Open in a new tab Ogg1 KO primary neurons exhibit dysfunctional mitochondrial metabolism following saline and EtOH exposure. (A)Ogg1 KO primary cortical and hippocampal neurons were obtained from GD 17 fetal brains and grown for 14 days in vitro (DIV). Image shows the cells at 8 DIV. The scale bar indicates 50 μm. At 14 DIV, Ogg1 WT and KO primary neurons were treated with saline or EtOH and subjected to mitochondrial stress analysis using the Seahorse assay (n = 5). (B) Trace showing the oxygen consumption rate (OCR) over the 90-min study duration (n = 5 for each group). (C) Basal OCR in Ogg1 WT and KO primary neurons treated with saline or EtOH (n = 5). EtOH reduced the basal respiration in Ogg1 KO neurons, compared to saline-exposed Ogg1 KO neurons (p < 0.01). (D) OCR attributed to proton leakage (n = 5). Proton leakage was similarly increased in both the saline- and EtOH-exposed Ogg1 KO vs. WT neurons (p < 0.0001 and p < 0.0001). (E) Maximal OCR that was stimulated by the addition of the mitochondrial oxidative phosphorylation uncoupler, FCCP. The maximal respiration was reduced in both an OGG1-and EtOH-dependent manner, with the maximal reduction in EtOH-exposed Ogg1 KO neurons (p < 0.0001 for all comparisons). (F) OCR dedicated for mitochondrial ATP production (n = 5). ATP production was reduced in Ogg1 KO neurons exposed to EtOH vs. saline (p < 0.0001), and in EtOH-exposed neurons, ATP production was reduced in Ogg1 KO vs. WT neurons p < 0.0001). Statistical significance was determined by 2-way ANOVA and Tukey's post-hoc tests (∗∗p < 0.01, ∗∗∗∗p < 0.0001). Abbreviations: FCCP, carbonyl cyanide p-trifluoro methoxyphenylhydrazone. See Methods for the roles of oligomycin, antimycin and rotenone. 3.8. Mitochondrial dysfunction in saline-exposed Ogg1 KO cells Physiological levels of ROS altered mitochondrial metabolism parameters in Ogg1 KO vs. WT cells exposed only to saline. While the apparent increase in basal respiration in saline-exposed Ogg1 KO vs. WT cells was not statistically significant (p = 0.07) ([202]Fig. 5C), there was a 2.7-fold increase in oxygen consumed but lost to proton leakage in Ogg1 KO vs. WT cells (p < 0.0001) ([203]Fig. 5D), and conversely highly significant decreases in both the maximal respiration (p < 0.0001) ([204]Fig. 5E) and spare respiration capacity (p < 0.0001) ([205]Suppl. Fig. 2). The dysfunctional outcomes in Ogg1 KO cells reveal that oxidatively damaged DNA from even physiological levels of ROS can substantially impair mitochondrial function when DNA repair is deficient. 3.9. Mitochondrial dysfunction in Ogg1 KO cells was worsened by exposure to EtOH EtOH exposure had a more pronounced impact than saline on the effect of Ogg1 genotype for several mitochondrial metabolism parameters. With EtOH exposure, basal respiration was decreased 21 % (p = 0.059) in the Ogg1 KO vs. WT cells (opposite of the trend for saline, p = 0.07) ([206]Fig. 5C). EtOH also reduced the basal respiration in Ogg1 KO neurons compared to saline-exposed Ogg1 KO neurons (p < 0.01) ([207]Fig. 5C). Oxygen consumption for ATP production decreased 54 % in EtOH- vs. saline-exposed Ogg1 KO cells (EtOH effect) (p < 0.0001), and also deceased 67 % in EtOH-exposed Ogg1 KO vs. WT cells (p < 0.0001), which was not observed with saline exposure ([208]Fig. 5F). This decrease caused by EtOH in oxygen consumed for ATP production in Ogg1 KO vs. WT cells was inversely related to a 3.4fold increase by EtOH in oxygen consumed but lost to proton leakage (p < 0.0001), which was virtually identical to the Ogg1 KO vs. WT increase with saline exposure ([209]Fig. 5D). The maximal respiration capacity was significantly diminished with EtOH exposure in both an OGG1-and EtOH-dependent manner ([210]Fig. 5E). In Ogg1 WT cells, EtOH exposure reduced the maximal respiration by 30 % compared to saline exposure (p < 0.0001) ([211]Fig. 5E). Also, EtOH-exposed Ogg1 KO vs. WT cells sustained a further 36 % reduction in maximal respiration (p < 0.0001), showing a Ogg1-dependent effect ([212]Fig. 5E). Together these results indicate decreased energy production in the mitochondria due to proton leakage, a mechanism that uncouples oxidative phosphorylation, in the EtOH-exposed Ogg1 KO cells and to a lesser extent in the Ogg1 WT cells. Overall, mitochondrial dysfunction was exacerbated by EtOH vs. saline in an OGG1-dependent fashion for basal respiration ([213]Fig. 5C), maximal respiration ([214]Fig. 5E) and ATP production ([215]Fig. 5F), but not for proton leakage ([216]Fig. 5D). 3.10. Sex-dependent effects of OGG1 deficiency and prenatal EtOH exposure on brain function To investigate whether brain function was affected by oxidatively damaged DNA caused by a single prenatal exposure on GD 17 to physiological (saline treatment) or drug-enhanced (EtOH, 2 × 2.5 g/kg, 4 h apart) levels of ROS formation, we compared the performance of Ogg1 KO vs. WT progeny in 3 behavioural tests. The selection of these tests was based in part on previous behavioural studies of DNA repair-deficient mice in our lab [[217]4,[218]19,[219]62], along with our microarray results implicating gene dysregulation in genes involved in long-term potentiation, which is involved in learning and memory. 3.10.1. Puzzle box test The puzzle box test assesses general cognition and executive function by evaluating the capacity of a mouse to solve progressively more challenging tasks and recall each learned skill after 2 min (short-term) and 24 h (long-term) [[220]71]. For mice with intact high-order executive functions, such as working memory, planning and problem solving, the task completion time will decrease after each trial. The time to task completion reflects the time each mouse was near the underpass as opposed to the end of the apparatus. The results of the puzzle box test are shown in [221]Fig. 6, separated by problem solving, working memory and long-term memory trials for females (left panels) and males (right panels). A 2-way ANOVA was conducted to determine the effects of Ogg1 genotype and/or prenatal EtOH exposure on the task completion time for each trial ([222]Suppl. Table 8). Fig. 6. [223]Fig. 6 [224]Open in a new tab Sex-dependent differences in reduced long-term memory and problem solving following prenatal EtOH exposure in Ogg1 WT and KO progeny. This test is based upon mouse preference for the dark vs. light chamber in a two-chambered box, where the progeny perform 9 problem-solving trials with increasing difficulty (underpass < burrow task < plug task) to enter a dark chamber, with 3 trials/day performed over 3 days. A higher time to task completion indicates a disorder depending on the trial. Trial 1 assesses (A) problem solving ability, while trial 2, conducted 2 min after trial 1, evaluates (B) working memory. Trial 3, administered 24 h after trials 1 and 2, evaluates (C) long-term memory function. The male and female EtOH-exposed Ogg1 KO vs. WT progeny had deficits in long-term memory and took longer to complete the underpass task after 24 h (p < 0.05 and p < 0.05). The EtOH-exposed vs. saline-exposed Ogg1 WT female progeny also took significantly longer to complete the burrow task after 24 h (p < 0.05). Only the male EtOH-exposed Ogg1 KO progeny showed reduced problem-solving in the underpass task, compared to saline-exposed Ogg1 KO progeny and EtOH-exposed Ogg1 WT progeny (p < 0.05 and p < 0.05, respectively). The number of mice tested, selected from at least 3 different litters, is indicated in parentheses. Statistical significance was determined by a 2-way ANOVA and Tukey's post-hoc tests (∗p < 0.05). Problem solving (PS) – The results of a 2-way ANOVA showed the females did not exhibit any significant OGG1-or EtOH-dependent effects in the underpass, burrowing or plug PS trials. In contrast, the males had a significant effect of EtOH on completion time in the underpass and burrowing PS trials, but not the plug PS trial [Underpass: EtOH effect (F(1,34) = 4.86, p < 0.05) interaction (F(1,34) = 4.82, p < 0.05); Burrowing: EtOH effect (F(1,34) = 4.30, p < 0.05), [225]Suppl. Table 8]. A Tukey's post hoc test showed the completion time in the PS underpass task was higher for EtOH- vs. saline-exposed Ogg1 KO progeny (p < 0.05), and vs. EtOH-exposed Ogg1 WT progeny (p < 0.05), respectively, revealing both OGG1-and EtOH-dependent effects ([226]Fig. 6A). Working memory (2 min) – There were no significant effects of Ogg1 or prenatal EtOH exposure in females or males in the short-term working memory trials ([227]Fig. 6B). Long-term memory (LTM) (24 h) – Females demonstrated impairment in the LTM trials, indicated by statistically significant main effects of Ogg1 genotype and prenatal EtOH exposure ([228]Suppl. Table 8). In the LTM underpass task, there was a significant main effect of Ogg1 genotype (F(1,34) = 4.70, p < 0.05), a significant main effect of EtOH exposure (F(1,34) = 5.27, p < 0.05), and a trend toward significance for the interaction (F(1,34) = 6.52, p=0.052). A post hoc Tukey's test indicated that task completion time in the underpass test for Ogg1 KO female progeny was non-significantly higher for EtOH vs. saline exposure (p = 0.07) (drug effect), and in EtOH-exposed progeny was higher in Ogg1 KO vs. WT progeny (p < 0.05) (genotype effect) ([229]Fig. 6C). In the LTM burrowing task, the task completion time for Ogg1 WT female progeny was higher with EtOH vs. saline exposure (p < 0.05) (drug effect), and in EtOH-exposed progeny, there was a trend for a worse performance by Ogg1 KO vs. WT progeny (p = 0.06) (genotype effect) ([230]Fig. 6C). In male progeny, 2-way ANOVA indicated the males had a significant main effect of Ogg1 genotype and prenatal EtOH exposure in the LTM underpass task, a significant main effect of Ogg1 genotype (F(1, 34) = 4.70, p < 0.05); a significant main effect of EtOH exposure (F(1, 34) = 5.27, p < 0.05) and a significant interaction between Ogg1 genotype and EtOH exposure (F(1, 34) = 6.52, p < 0.05) ([231]Suppl. Table 8). A post hoc Tukey's test indicated the task completion time in the LTM underpass task for Ogg1 KO male progeny, similar to females, was higher for EtOH vs. saline exposure (p < 0.05) (drug effect), and in EtOH-exposed progeny was higher in Ogg1 KO vs. WT progeny (p < 0.05) (genotype effect) ([232]Fig. 6C). Unlike in females, there was no effect of EtOH or genotype in the burrowing test. 3.10.2. Fear conditioning We tested fear memory behaviour using contextual and conditioned stimulus cues, and measured the freezing time upon re-exposure, known as fear conditioning ([233]Fig. 7A). Fear conditioning is a commonly used test to measure the recall of fear memory following conditioning, which is a measure of learning and memory, but also involves an emotional component [[234]68]. When female or male progeny were exposed to the fear-conditioned context, there were no significant differences in freezing time among the groups for habituation or context retrieval ([235]Fig. 7B). Unexpectedly, when exposed to a novel context (visual environment, control), the saline-exposed Ogg1 KO vs. WT female progeny exhibited longer freezing behaviour (p < 0.01), and a 2-way ANOVA indicated a significant main effect of Ogg1 genotype (F(1,31) = 6.70, p < 0.05). This result was unexpected because the context (visual environment) in which the conditioning took place was changed, yet the females still showed a freezing response in a novel context, which was not seen with the WT mice. This altered freezing behaviour in females was not seen in the saline-exposed Ogg1 KO male littermates, indicating a significant effect of sex (p < 0.05). When exposed to the conditioned auditory cue, female progeny, but not their male littermates, exhibited a significant effect of Ogg1 genotype (F(1,31) = 9.97, p < 0.01). In females, a Tukey's post hoc test showed a significant increase in freezing time in the conditioned auditory cue test in both the saline- and EtOH-exposed Ogg1 KO progeny vs. WT progeny (p < 0.05). In males, there were no significant changes in freezing time among the groups. These results demonstrate an Ogg1 genotype effect that is independent of EtOH exposure. Increased freezing time for the female Ogg1 KO progeny was associated with increased fear memory to the auditory cue conditioned stimulus; however, the abnormal freezing response in the neutral novel context in females could alternatively mean that they may have failed to distinguish between the shock-paired context and a neutral context. 3.10.3. Passive avoidance The passive avoidance test is similar to the fear conditioning test, but instead measures an avoidance response after conditioning [[236]85]. The passive avoidance data for day 3 are shown for females and males ([237]Fig. 8), and the data for days 1 and 2 are plotted in [238]Suppl. Fig. 3. For females, on day 3, a 2-way ANOVA showed a significant effect of Ogg1 genotype (F(1,31) = 25.08, p < 0.0001) and prenatal EtOH exposure (F(1,31) = 7.67, p < 0.001), but the interaction term was not significant. An Ogg1 genotype effect was seen in females with both saline and EtOH exposure. Tukey's post hoc test showed saline-exposed Ogg1 KO vs. WT progeny had a reduced latency time (p < 0.05), as did EtOH-exposed Ogg1 KO vs. WT progeny (p < 0.05). In addition, EtOH significantly affected Ogg1 WT progeny, in which latency time was reduced by EtOH vs. saline exposure (p < 0.05). There were no statistically significant differences in latency time in males, although the pattern of Ogg1 genotype effects appeared similar to those in females. Fig. 8. [239]Fig. 8 [240]Open in a new tab Sex-, OGG1-and EtOH-dependent learning and memory disorders in the passive avoidance test. Saline-exposed female, but not male, Ogg1 KO progeny, performed worse than their female WT controls, revealing an OGG1-dependent effect at physiological levels of ROS formation (p < 0.05). Similarly, prenatally EtOH-exposed female, but not male, Ogg1 KO progeny, exhibited a worse performance than their EtOH-exposed female WT controls (p < 0.05). Female Ogg1 WT progeny prenatally exposed to EtOH also exhibited a worse performance compared to saline-exposed female WT progeny, revealing the effect of a ROS-enhancing drug (p < 0.05). Males exhibited a similar but lesser and non-significant pattern for an OGG1 effect, but not an EtOH effect. Progeny were selected from at least 3 different litters, and the number of mice tested in each group is indicated in parentheses. Statistical significance was determined by a 2-way ANOVA and Tukey's post-hoc tests (∗p < 0.05). 4. Discussion and conclusions 4.1. Overview Although the molecular etiology of NDDs is not well characterized, oxidatively damaged DNA is likely involved in some of these and other developmental disorders [[241]1]. Herein, using a DNA repair-deficient Ogg1 KO mouse model, we have shown that oxidatively damaged DNA in the form of 8-oxoGua lesions and DSBs was increased in the saline-exposed fetal brains of Ogg1 KO vs. WT progeny, and was further exacerbated by prenatal EtOH exposure. Increased levels of oxidatively damaged DNA were associated with altered gene expression in fetal brains, mitochondrial dysfunction and cognitive NDDs. Gene dysregulation in saline-exposed Ogg1 KO vs. WT fetal brains revealed the pathogenic potential of physiological ROS levels when DNA repair is deficient. Further dysregulation of these outcomes in EtOH-exposed brains showed a substantial impact of environmental enhancement of oxidatively damaged DNA. Ogg1 KO mice exposed to EtOH-enhanced levels of ROS exhibited more pronounced dysregulation than Ogg1 WT mice, notably displaying a unique alteration in genes involved in the long-term synaptic potentiation pathway, a critical mechanism in learning and memory. At the cellular level, both OGG1 deficiency and EtOH exposure altered multiple parameters of mitochondrial metabolism in primary Ogg1 neurons, which can impact axonal and dendritic development and synaptic function [[242]86]. Prenatal exposure to EtOH on GD 17 caused sex-dependent deficits in several behavioural tests of learning and memory that correlated with OGG1-and EtOH-dependent oxidatively damaged DNA, gene dysregulation and mitochondrial dysfunction, consistent with the involvement of these processes in the mechanism of ROS-mediated NDDs. 4.2. Epigenetic effects of oxidatively damaged DNA There are increasing reports documenting an epigenetic role of 8-oxoGua [[243]48,[244]51,[245]54,[246]57,[247]87], as distinct from its commonly known mutational role in carcinogenesis. In airway epithelium, 8-oxoGua/OGG1 complexes recruit transcription factors at sites of 8-oxoGua formation to facilitate gene expression [[248]54]. In the brain, adult Ogg1 KO hippocampi, with physiological ROS exposure, exhibited dysregulated gene expression, which were involved in pathways relating to learning and memory, and anxiety [[249]53]. Our study is the first to evaluate the impact of oxidatively damaged DNA on the developing fetal brain, providing support that 8-oxoGua lesions and DSBs can cause gene dysregulation. Among the dysregulated genes, we found miRNA-9 (miR-9) was uniquely dysregulated in EtOH-exposed Ogg1 KO fetal brains. As one of the most highly expressed miRNAs in the developing brain, miR-9 plays a key role in the proliferation, neurogenesis, migration, maturation and differentiation of neurons [[250]88]. miR-9 is EtOH-sensitive in fetal mouse cortical neurospheres, and may contribute to the teratogenic effects of EtOH in FASD [[251]89]. Our results with prenatal EtOH exposure suggest a role for additional ROS-initiated, DNA damage-mediated mechanisms of NDDs independent of OGG1, possibly involving DSBs, as gene dysregulation was more pronounced in the Ogg1 KO fetal brains ([252]Fig. 3B). For example, we have found that prenatal EtOH exposure can significantly decrease the fetal brain levels of BRCA1, another DNA repair protein [[253]4]. An EtOH-dependent role of oxidatively damaged DNA independent of OGG1 is further evidenced by the reduced number of genes with TF binding site enrichment in the dysregulated genes in the Ogg1 KO vs. WT fetal brains ([254]Fig. 4B). We identified 4 TFs that were uniquely upregulated in Ogg1 KO fetal brains Arp-1, Gr, Pr, and Pit-1 ([255]Suppl. Table 3), warranting further investigation. Several other mechanisms by which 8-oxoGua can alter gene expression have been demonstrated [[256]53]. 8-OxoGua mechanisms of gene dysregulation independent of OGG1 include 8-oxoGua physically blocking transcription factor binding within promoter regions [[257]90,[258]91]; disrupting the stability and function of G-quadruplex structures, potentially compromising their role in regulating gene expression [[259]51]; and, inhibiting the binding of MeCP2, a DNA methylation “reader”, which can recruit histone deacetylases to alter chromatin conformation [[260][92], [261][93], [262][94]]. 4.3. Cell fate At the global level, our gene ontology analysis did not reveal significant enrichment for pathways relating to cellular senescence, apoptosis and the determination of cellular fate. It is possible that such OGG1-dependent changes occur, but are restricted to specific brain regions and/or cell types, for which changes are obscured in the whole-brain analysis used in our study. In progeny exposed to EtOH, it also is possible that dysregulation of these pathways occurred outside of the time window surrounding the sampling time 6 h post-EtOH exposure. However, at the gene level, the transcription of apoptotic genes, including Trp53 (downregulated) and Mdm2 (downregulated), were dysregulated in EtOH-exposed Ogg1 KO fetal brains, as was the expression of the cellular senescence genes Comp (upregulated), Mirlet7f-1 (downregulated) and Trp53 (downregulated). Of potential relevance to EtOH-exposed Ogg1 KO brains was the report of elevated expression of inflammatory cytokines, which have been linked to cellular senescence, particularly the senescence-associated secretory phenotype [[263]95]. Apoptotic and senescence pathways also may be more prominently affected at a later stage of development, when oxidatively damaged DNA has had sufficient time to trigger downstream cellular responses. These changes in apoptotic and senescence genes suggest that pathway enrichment analysis of cells from specific brain regions and cell types at multiple time points would be similarly revealing. Our gene expression analysis suggests that apoptosis and senescence caused by transient prenatal oxidative damage to DNA might contribute to the mechanism of ROS-mediated NDDs, providing that protein and cellular outcomes are shown to reflect the changes in gene expression observed herein. However, some NDDs initiated in Ogg1 KO progeny by a single prenatal exposure to EtOH can be blocked by postnatal treatment with a histone deacetylase inhibitor [[264]96]. This protection by postnatal therapy would appear to rule out an obligatory role for transient prenatal apoptosis in at least some NDDs, although senescence remains a possibility, as those cells do not die, and might exert postnatal pathogenic effects. 4.4. Mitochondrial impact of oxidatively damaged DNA OGG1 protects the nuclear and mitochondrial DNA (mtDNA) from oxidative DNA damage [[265]97]. 8-oxoGua lesions accumulate 3 times more in the mtDNA than nuclear DNA, possibly due to proximity to the electron transport chain [[266]81]. In the developing fetal brain, efficient mitochondrial function is required for energy-demanding processes like neurite outgrowth [[267]98] and synaptogenesis [[268]86]. Moreover, mitochondrial dysfunction has been documented in individuals with FASD [[269]99] and ASD [[270]100]. We found that cultured primary Ogg1 KO neurons exposed only to saline, which approximates physiological levels of ROS production, had reductions in maximal respiration ([271]Fig. 5E) and spare respiration capacity ([272]Suppl. Fig. 2), which indicate inefficiencies in oxidative phosphorylation under cellular stress conditions [[273]101]. This dysfunction in saline-exposed Ogg1 KO vs. WT cells reveals the pathogenic mitochondrial impact of physiological levels of ROS production when DNA repair is deficient. The further reduction in maximal respiration in EtOH- vs. saline-exposed Ogg1 WT cells showed that environmentally enhanced ROS formation can exacerbate mitochondrial pathology, which in turn is further exacerbated by DNA repair deficiency, as shown in EtOH-exposed Ogg1 KO vs. WT cells ([274]Fig. 5E). A reduced maximal capacity may also be a result of reduced mitochondrial mass, which was not measured herein. However, there was a substantial OGG1-dependent increase in proton leakage, consistent with an uncoupling mechanism, which during oxidative stress may serve to reduce oxidant production in the mitochondria to decrease mtDNA damage [[275]102]. This suggests that the Ogg1 KO mitochondria may have high levels of oxidative stress causing them to dissipate the proton gradient, which uncouples proton transport from ATP synthesis. Others have found similar results in primary neural stem cells (NSCs) from the hippocampal dentate gyrus of Ogg1 KO vs. WT brains, where enhanced oxidatively damaged DNA reduced neurogenesis, interestingly causing NSCs to shift from neurogenesis to astrogliosis [[276]103]. In a double DNA repair Mth1/Ogg1 KO model, primary adult cortical neurons had a decreased mitochondrial membrane potential associated with mitochondrial dysfunction and cell death, and these cells required antioxidant supplementation for proper neuronal outgrowth and arborization [[277]82]. These findings suggest that oxidatively damaged DNA in neurons can significantly disrupt mitochondrial bioenergetics, thereby adversely affecting neuronal growth. 4.5. Impact on behavioural disorders Oxidatively damaged DNA in the developing brain, particularly when unrepaired, is a key driver of NDD-like phenotypes, as evidenced herein. We investigated the impact of oxidatively damaged DNA on learning and memory using 3 different behavioural tests of cognition. Our previous work showed that EtOH-initiated DNA oxidation in Ogg1 KO progeny led to disorders in the passive avoidance test, compared to WT progeny (sex not determined) [[278]18]. In the current study, we found similar disorders in the passive avoidance test in EtOH-exposed Ogg1 KO progeny. However, we also found significant sex-dependent differences affecting females more severely, determined outside of proestrus and estrus stages to avoid potential confounding estrus-mediated brain effects [[279]104]. The observation that female EtOH-exposed Ogg1 KO mice performed worse in cognitive tests than their male counterparts was also seen in a KO model for another DNA repair protein, BRCA1, in which the females exhibited an increased risk of NDDs, including disorders in learning and memory [[280]105]. Moreover, when ROS formation was blocked by pretreatment with phenylbutylnitrone, a NADPH oxidase inhibitor [[281]4] and free radical scavenging agent, some NDD phenotypes in the EtOH-exposed Ogg1 progeny were prevented, consistent with a ROS-dependent mechanism [[282]50]. Our current findings provide further evidence that some EtOH-initiated behavioural disorders are initiated, at least in part, by ROS-mediated oxidative DNA damage in the developing fetal brain. If not repaired by OGG1, and perhaps by other DNA repair proteins like BRCA1, this damage can result in abnormal postnatal brain function, implicating ROS formation and repair of oxidatively damaged DNA in determining the risk of NDDs. The initiating mechanisms for altered brain formation may occur during in utero brain development, affecting gene expression and mitochondrial energetics. Several limitations should be considered when interpreting our findings. First, the microarray analysis tested 3 biological replicates, which is common. However, this small sample size may have limited the generalizability of the results and reduced our statistical power, particularly in the saline exposure group. A larger sample size may reveal more nuanced and significant effects of oxidatively damaged DNA on gene expression, particularly at physiological ROS levels approximated by saline treatment. Furthermore, the transcriptome analysis was conducted using whole fetal brains, which may obscure critical regional differences in gene expression across distinct brain regions. Additionally, to focus on mechanisms of NDD initiation, we analyzed gene expression on GD 17 at one early timepoint, corresponding to the peak of oxidatively damaged DNA 6 h after EtOH exposure, and our results may not reflect dynamic changes at later times. This study focused on the impact of oxidatively damaged DNA on gene expression and mitochondrial dysfunction in the fetal brain, and it would be worth knowing in our model whether cellular outcomes such as apoptosis, autophagy and cellular senescence are involved in OGG1-and EtOH-dependent mechanisms in NDDs. 4.6. Molecular basis of sex differences in NDDs Behavioural testing showed that enhanced oxidatively damaged DNA caused more severely disordered brain function in females than males. The determinants of this sex difference appear to be downstream of oxidatively damaged DNA, as there was no sex difference in DNA damage, at least as measured at the global rather than gene-specific level. This also suggests that upstream pathways of ROS formation and detoxification, and DNA repair, which altogether regulate the levels of oxidatively damaged DNA, cannot explain sex differences observed in some NDDs. 5. Conclusions The molecular mechanisms of NDDs are not well understood, but may involve ROS-initiated DNA damage in the fetal brain. We have shown, at the molecular level, that increases in the oxidative DNA lesions 8-oxoGua and DSBs in the fetal brain due to OGG1 deficiency (prenatal saline exposure in Ogg1 KO progeny) or following prenatal exposure to the ROS enhancer EtOH, are all associated with altered transcription of developmentally important genes, and postnatal NDDs. Similarly, neuronal mitochondria from Ogg1 KO brains showed dysfunctional bioenergetics, even at physiological levels of oxidatively damaged DNA, some of which were exacerbated by exposure to EtOH. Notably, female Ogg1 KO mice exposed prenatally to either saline or EtOH had a higher risk than males for learning and memory disorders, apparently due to factors downstream of oxidatively damaged DNA. The association of NDDs with enhanced oxidatively damaged DNA, altered gene expression and mitochondrial dysfunction, due to both OGG1 deficiency and prenatal EtOH exposure, suggest that these molecular alterations constitute important mechanisms of ROS-mediated NDDs. These new insights may prove useful in developing diagnostic biomarkers and novel therapeutic approaches for some NDDs. CRediT authorship contribution statement Ashley P. Cheng: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. Piriththiv Dhavarasa: Data curation, Formal analysis, Investigation, Methodology. Jana van Heeswyk: Data curation, Formal analysis, Investigation. Sophia M. Richards: Data curation, Formal analysis, Investigation, Methodology, Writing – review & editing. Xuan Li: Data curation, Formal analysis, Investigation. Aaron M. Shapiro: Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Supervision. Peter G. Wells: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – review & editing. Funding This work was supported by grants from the Canadian Institutes of Health Research (PJT-156023) and a New Initiatives & Innovations Award from the University of Toronto Faculty of Pharmacy. APC received a scholarship from the University of Toronto Centre for Pharmaceutical Oncology. Declaration of competing interest None. Acknowledgements