Altered hippocampal neurogenesis in a mouse model of autism revealed by genetic polymorphisms and by atypical development of newborn neurons

Individuals with autism spectrum disorder (ASD) often exhibit atypical hippocampal anatomy and connectivity throughout their lifespan, potentially linked to alterations in the neurogenic process within the hippocampus. In this study, we performed an in-silico analysis to identify single-nucleotide polymorphisms (SNPs) in genes relevant to adult neurogenesis in the C58/J model of idiopathic autism. We found coding non-synonymous (Cn) SNPs in 33 genes involved in the adult neurogenic process, as well as in 142 genes associated with the signature genetic profile of neural stem cells (NSC) and neural progenitors. Based on the potential alterations in adult neurogenesis predicted by the in-silico analysis, we evaluated the number and distribution of newborn neurons in the dentate gyrus (DG) of young adult C58/J mice. We found a reduced number of newborn cells in the whole DG, a higher proportion of early neuroblasts in the subgranular layer (SGZ), and a lower proportion of neuroblasts with morphological maturation signs in the granule cell layer (GCL) of the DG compared to C57BL/6J mice. The observed changes may be associated with a delay in the maturation trajectory of newborn neurons in the C58/J strain, linked to the Cn SNPs in genes involved in adult hippocampal neurogenesis.


Altered hippocampal neurogenesis in a mouse model of autism revealed by genetic polymorphisms and by atypical development of newborn neurons
Isabel Barón-Mendoza1,2 , Montserrat Mejía-Hernández 1,2 , Karina Hernández-Mercado 1 , Jessica Guzmán-Condado 1 , Angélica Zepeda 1* & Aliesha González-Arenas 1* Individuals with autism spectrum disorder (ASD) often exhibit atypical hippocampal anatomy and connectivity throughout their lifespan, potentially linked to alterations in the neurogenic process within the hippocampus.In this study, we performed an in-silico analysis to identify single-nucleotide polymorphisms (SNPs) in genes relevant to adult neurogenesis in the C58/J model of idiopathic autism.We found coding non-synonymous (Cn) SNPs in 33 genes involved in the adult neurogenic process, as well as in 142 genes associated with the signature genetic profile of neural stem cells (NSC) and neural progenitors.Based on the potential alterations in adult neurogenesis predicted by the in-silico analysis, we evaluated the number and distribution of newborn neurons in the dentate gyrus (DG) of young adult C58/J mice.We found a reduced number of newborn cells in the whole DG, a higher proportion of early neuroblasts in the subgranular layer (SGZ), and a lower proportion of neuroblasts with morphological maturation signs in the granule cell layer (GCL) of the DG compared to C57BL/6J mice.The observed changes may be associated with a delay in the maturation trajectory of newborn neurons in the C58/J strain, linked to the Cn SNPs in genes involved in adult hippocampal neurogenesis.
Autism spectrum disorder (ASD) comprises a group of neurodevelopmental disorders characterized by persistent deficits in social interaction/communication and the presence of repetitive and stereotyped behaviors 1 .It is estimated that 1 in 100 children worldwide has ASD 2 .The etiology of autism is polygenic and, in most cases, it is idiopathic 3 .
Interestingly, dysregulation in the neurogenic process has been proposed as one of the underlying mechanisms associated with changes in brain structure in ASD 4 .There is evidence of aberrant cortical lamination in children with ASD, which may arise from disturbances in cell proliferation, migration, and differentiation during prenatal development 5 .Moreover, toddlers with autism exhibit brain overgrowth during the first years of life, which may result from altered cell proliferation 4 .In line with this, in vitro experiments using induced pluripotent stem cells (iPSCs)-derived neural progenitors from individuals with autism with increased brain volume have demonstrated accelerated proliferation 6 , decreased proliferation when derived from individuals who do not have macrocephaly 7 , and impairments during the determination of neuronal subtype 8 .There is also strong evidence that genes with autism-associated variants are involved in the neurogenic regulation 3,4,9 .
Adult neurogenesis in humans is still a topic of debate; however, in mice, the subgranular zone (SGZ) of the hippocampal dentate gyrus (DG) is one of the well-recognized adult neurogenic niches 10,11 .Adult neurogenesis in rodents plays a crucial role in hippocampal functionality and plasticity 12 , while individuals with autism often display atypical hippocampal anatomy and connectivity throughout their lifespan 13 , as well as changes in the volume and cytoarchitecture of the DG, particularly in the granule cell layer (GCL) 14,15 .Hence, the evaluation of the neurogenic process in the hippocampus of autism models is a relevant topic for understanding the potential contribution of this process in ASD.
Adult neurogenesis in the mouse DG involves a series of tightly regulated cellular events that allow progenitor cells to differentiate and develop into fully functional mature granule cells (GC) [16][17][18] .Type 1 neural stem cells (NSC), also known as radial glial-like progenitors, express stem markers nestin and GFAP (glial fibrillary acidic protein).Upon specific signaling, NSC residing in the SGZ of the DG exit their quiescent state and start dividing, giving rise to intermediate progenitor cells (Type 2a cells) which exhibit high proliferative activity and a nonradial glial-like morphology, while still expressing glial markers.Type 2a cells generate intermediate progenitor cells (Type 2b cells) that express early neuronal specification markers such as DCX (Doublecortin) and PSA-NCAM (Polysialylated-neural cell adhesion molecule) 16,19 .Type 2 cells that survive the critical period divide and differentiate into neuroblasts (Type 3 cells) 20 , which express DCX and PSA-NCAM and exhibit varying morphologies depending on their maturation stage.In early stages, they may display short, plump dendritic processes or none.In the final stage of morphological maturation, fully differentiated young neurons develop a long apical dendrite with branches reaching the molecular layer, parallel to the expression of calretinin 21 .Only a small subset of the newborn cells matures into calbindin-expressing granule cells, which are then available to integrate into pre-established DG circuits 16,19 .
The adult neurogenic process is also controlled by coordinated expression patterns of several genes 17,22,23 .Interestingly, common genetic variants in inbred mice are associated with strain differences in adult hippocampal neurogenesis, suggesting a strong contribution of polymorphisms in the regulation of the neurogenic process 23 .
Our group has previously identified single-nucleotide polymorphisms (SNPs) in genes essential for brain development and plasticity in the C58/J murine model of idiopathic ASD 24 .The C58/J strain exhibits alterations in hippocampal cytoarchitecture and hippocampus-dependent learning deficits [24][25][26][27] , which may be linked, among other factors, to impairments in hippocampal neurogenesis.
Given the impact that specific genetic alterations may have on the neurogenic process and the potential dysregulation of neurogenesis in ASD after birth, we conducted an in-silico study of single-nucleotide polymorphisms (SNPs) in genes relevant to adult neurogenesis in the C58/J model and analyzed whether the hippocampal neurogenic process in this strain could potentially exhibit alterations based on the predicted changes.

The C58/J strain displays non-synonymous SNPs in genes associated with the regulation of adult neurogenesis
According to in vivo and in vitro experimental data curated and reported by the MANGO database, a total of 397 genes are expressed during the stages of proliferation, differentiation, survival, migration, dendritogenesis, and maturation in adult neurogenesis 17 .These genes are also expressed in specific cell stages defined by MANGO as precursor cells, stem cells (Type 1), undetermined progenitors (Type 2a), determined progenitors (Type 2b), neuroblast-like cells (Type 3), new neurons, immature and mature neurons, and doublecortin (DCX)-positive cells 17 (See supplementary Table S1 for MANGO criteria and definitions).
Thus, to predict potential changes in adult neurogenesis in the C58/J model of autism, we conducted an in-silico analysis of the 397 genes reported by MANGO to identify single-nucleotide polymorphisms (SNPs) in the C58/J strain, compared to the control C57BL/6 J strain (WT).We used the Sanger4 Dataset available in the Mouse Phenome Database (MPD) [28][29][30] .Our analysis revealed that among the 397 genes, 33 of them carried at least one non-synonymous SNP in coding regions (Cn SNPs), leading to changes in the coding amino acid sequence, in the C58/J strain.Notably, the Disc1 gene showed the highest enrichment of coding non-synonymous (Cn) SNPs (Fig. 1A,C, Supplementary Table S2).Additionally, within the Cn SNPs of the Disc1 gene, three (rs31943453, rs31944226, rs215748054) were identified as highly prone to damaging the protein structure based on impact predictions conducted using the PolyPhen-2 platform 31 (Supplementary Table S3).
Based on the gene expression profile of the mouse DG available in the Hipposeq platform 32 , 26 of the 33 genes with Cn SNPs were also found to be differentially expressed in mature granule cells (GC) (Supplementary Fig. S1C).
In a recent study, Artegiani et al. (2017) reported a gene expression profile that characterizes neural stem cells (NSC) (591 genes) and neural progenitors (1065 genes) in the mouse DG 33 .Based on that report, we investigated In silico analysis of coding nonsynonymous (Cn) single-nucleotide polymorphisms (SNPs) in genes relevant to adult neurogenesis in the mouse DG. (A) Among the 397 genes associated with processes (left) and cell stages (right) involved in adult neurogenesis, according to the MANGO database (data inside the large circle), 33 genes carried at least one Cn SNP in C58/J mice (data inside the small circle).(B) Among the genes characterizing the genetic profile of neural stem cells (NSC) (591 genes) and neural progenitors (1065 genes) in the mouse DG, as reported by Artegiani (2017) (data inside the large circle), 142 genes carried at least one Cn SNP in C58/J mice, with 64 genes corresponding to NSC and 78 genes to the neural progenitors (data inside the small circle).(C) The 33 genes reported by the MANGO database display from 1 to 16 Cn SNPs in C58/J mice in comparison with the control C57BL/6 J strain.www.nature.com/scientificreports/whether C58/J mice exhibited any Cn SNP in the signature genes of NSC and neural progenitors.We identified a total of 142 genes with at least one Cn SNP in C58/J mice, with 64 genes corresponding to the NSC genetic profile and 78 genes corresponding to the neural progenitors' genetic profile (Fig. 1B, Supplementary Table S2).
To obtain a more comprehensive understanding of the potential mechanisms affected by the genes with Cn SNPs identified in the C58/J strain, we performed a Gene Ontology (GO) enrichment analysis using the g:Profiler platform 34 .
Among the top 20 enriched terms in the biological process category resulting from the GO enrichment analysis of the 33 genes with Cn SNPs in C58/J mice reported by MANGO, terms associated with the regulation of cell proliferation, development, nervous system development, neurogenesis, cell differentiation, cell signaling, and cell death were identified (Supplementary Fig. S1A, Supplementary Table S4).Other GO terms in the biological process category involved in the regulation of neurogenesis were enriched, including terms associated with neuron differentiation, neuron death, migration, neuron projection development, regulation of the ERBB, MAPK, and WNT signaling pathways, learning, memory, and behavior (Fig. 2A).Additionally, the enriched GO terms in the cellular component category were related to neuronal structures (Fig. 2B).
The GO enrichment analysis of the genes with Cn SNPs in C58/J mice corresponding to NSC and neural progenitors' datasets showed enriched terms associated with organic acid metabolism and biosynthesis in NSC, and cell signaling and heterocycle compound biosynthesis in neural progenitors in the biological process category (Fig. 2C).GO terms associated with cell compartments and structural support, such the membrane, nucleoplasm, and cytoskeleton, were also enriched (Supplementary Fig. S1B, Supplementary Table S4).

The spatial distribution and morphology of newborn cells are atypical features in the DG of the C58/J strain
We investigated whether the C58/J model of autism exhibited alterations in the neurogenic process in the hippocampus based on the predicted changes in adult neurogenesis revealed by the in-silico analysis.
We evaluated the number of newborn cells in the dorsal dentate gyrus (DG) of young adult C58/J and WT (C57BL/6J) mice.Using confocal microscopy, we detected newly generated neuronal-lineage cells (BrdU + / DCX +) that had undergone proliferation, survival, and differentiation within a 2-weeks period between the BrdU administration protocol and the time of sacrifice (Fig. 3).To obtain a representative sample of the number of double-positive cells in the dorsal DG of both groups, we imaged and analyzed one region of interest (ROI) from the crest, two from the suprapyramidal blade, and one from the infrapyramidal blade (Figs. 3 and 4A) (See methods section for detailed analysis procedures).
First, we estimated the total number of BrdU + /DCX + cells in the DG of WT and C58/J mice.We observed a lower number of double-positive cells in the whole DG of C58/J mice compared to the WT group (Fig. 4B, DG: *p = 0.0374, WT: 51,657.4± 12,723.6 vs. C58/J: 40,259.3± 4987.6 cells per mm 3 ).However, when we analyzed DG regions separately, we found no significant differences in the number of double-positive cells in the crest, and the suprapyramidal and infrapyramidal blades comparing both strains (Fig. 4B, crest, supra, infra).
The spatial positioning of the young neurons within the layers of the DG has been associated with their migration ability and maturation stage.Hence, we examined the distribution of BrdU + /DCX + cells within the subgranular zone (SGZ), granular cell layer (GCL), molecular layer (ML), and hilus of the DG (Supplementary Fig. S2A-D).Since only one double-positive cell was found in the ML of C58/J mice, we excluded this layer from the analysis.
There was a diminished number of BrdU + /DCX + cells within the GCL of the whole DG in the C58/J strain compared to WT mice (Supplementary Fig. S2A, GCL: *p = 0.0003, WT: 26,665.2± 7801.9 vs. C58/J: 14,116.9± 2741.8 cells per mm 3 ).We did not find any differences in the number of double-positive cells within the SGZ and the hilus between both groups (Supplementary Fig. S2A, SGZ, hilus).When we analyzed the DG by region, we observed that the number of double-positive cells was lower within the GCL of the suprapyramidal blade in the C58/J mice (Supplementary Fig. S2C, GCL: *p = 0.0339, WT: 12,966.6 ± 4,073.5 vs. C58/J: 7424.4 ± 2,435.6 cells per mm 3 ).We did not observe any changes in the number of double-positive cells within the SGZ or the GCL from the crest or from the infrapyramidal blade (Supplementary Fig. S2B, D).
To assess whether changes in the distribution of BrdU + /DCX + cells within the SGZ and GCL could not solely be attributed to the lower number of double-positive cells in the DG of C58/J mice, we calculated the proportion of BrdU + /DCX + cells present in each layer in relation to the total number of double-positive cells in the whole DG or in each specific region (Fig. 4C).
Changes in the distribution of BrdU + /DCX + neurons within the SGZ and GCL of C58/J mice may be linked to the maturation stage of the cells.Therefore, we categorized BrdU + /DCX + cells based on their morphology, as cell shape modifications are closely associated with the maturity stage of new neuroblasts transitioning into Figure 2. Gene Ontology (GO) enrichment analysis for neurogenesis-associated genes with coding nonsynonymous SNPs in the C58/J strain.Enriched GO terms resulting from the enrichment analysis of the 33 genes with Cn SNPs associated with adult neurogenesis in C58/J mice, according to the MANGO database, in the biological process (A) and cellular component (B) categories.(C) Enriched GO terms resulting from the analysis of the 64 and 78 genes with Cn SNPs associated with NSC and neural progenitors (PROG) in C58/J mice, according to the Artegiani database, in the biological process category.(D) The genes with Cn SNPs in C58/J mice from both databases (MANGO and Artegiani) were enriched in DNA-binding motifs for specific transcription factors (TFs).Adjusted p-values were obtained through g:SCS multiple testing correction method on g:Profiler.The number of genes found in each GO term is indicated in parentheses.Gene ratio: the number of requested genes found in the functional category divided by the number of genes from the background genome.Rich factor: the number of requested genes found in the functional category divided by the number of total genes comprised in the specific functional category.www.nature.com/scientificreports/granule cells 21 .We classified cells as early neuroblasts (ABC-type) if they exhibited short, plump dendritic processes or none.Cells in an intermediate maturation state corresponding to the D-type were those with a medium-length vertical dendrite extending towards the molecular layer (ML).Lastly, cells with the most mature morphology, characterized by a long apical dendrite or/and at least one branched dendrite oriented toward the ML, were classified as EF-type (Fig. 4D).We estimated the total number of ABC, D, and EF-type BrdU + /DCX + cells within each layer (SGZ, GCL) and region (crest, suprapyramidal and infrapyramidal blades) of the DG (Supplementary Fig. S2E-L).In the whole DG of C58/J mice, there was a lower number of ABC and EF-type BrdU + /DCX + cells within the GCL (Supplementary Fig. S2F, GCL-ABC: *p = 0.0106, WT: 14,953.4± 5,589.8 vs. C58/J: 9,515.8 ± 437.4; GCL-EF: *p = 0.0193, WT: 6,901.6 ± 2,598.5 vs. C58/J: 1,882.2 ± 1,506.3 cells per mm 3 ), but no differences within the SGZ compared to WT mice (Supplementary Fig. S2E).When we analyzed the DG by region, we observed a lower number of EF-type cells within the GCL of the suprapyramidal blade (Supplementary Fig. S2J, GCL-EF: *p = 0.0327, WT: 3,450.8 ± 1,718.2 vs. C58/J: 836.6 ± 792.9 cells per mm 3 ), and a reduced number of ABC-type cells within the GCL of the infrapyramidal layer of C58/J mice (Supplementary Fig. S2L, GCL-ABC: *p = 0.0386, WT: 3,450.8 ± 1,977.2 vs. C58/J: 1,673.1 ± 935.3 cells per mm 3 ).We did not observe changes in the number of any cell type within the SGZ and GCL of the crest, and the SGZ of the suprapyramidal and infrapyramidal blades comparing both strains (Supplementary Fig. S2G-I, K).
Then, to evaluate whether differences in the number of BrdU + /DCX + cells with a specific cell type could not be solely attributed to the lower number of double-positive cells in the DG of C58/J mice, we determined the proportion of ABC, D, and EF-type cells in relation to the total number of double-positive cells found within the SGZ and the GCL in the whole DG and in each region (Fig. 4E-H).

Genes associated with neurogenesis carrying non-synonymous SNPs in the C58/J strain are orthologous to human genes implicated in ASD
Finally, to gain insights into the relevance of the changes in adult DG neurogenesis found in the C58/J strain, we focused on evaluating the implication of all the identified genes with Cn SNPs (Supplementary Table S2) in the etiology of ASD.Thus, the genes with Cn SNPs in the C58/J mice were searched for human orthologous genes in the SFARI GENE database 37 , which provides current information on genes associated with ASD.
Among the genes with Cn SNPs in C58/J mice reported by MANGO, we identified 8 genes with reported human orthologs in the SFARI database (Supplementary Table S5).NF1 and CACNA1C exhibited the highest association scores (score 1) for ASD (Table 1).Moreover, we identified 15 genes with Cn SNPs from the NSC and progenitors' dataset with human orthologs in SFARI (Supplementary Table S5).Among these, ANK2, RFX3, CHD3, MYT1L, PHIP, and PHF21A display the highest association scores (score 1) for ASD (Table 1).MYT1L and RFX3 were also ranked as strong candidates specifically for ASD, rather than other neurodevelopmental Table 1.Genes with Cn SNPs involved in adult neurogenesis in C58/J mice are orthologous to human genes associated with autism spectrum disorder (high score according to SFARI)..NF1 and ANK2 were scored by EAGLE as moderate candidates (Table 1).Recently, Kim and collaborators reported a group of genes that were differentially expressed in radial glial cells (Pax6 + or Vimentin +) from the prefrontal cortex of individuals with ASD 39 .Based on this information, we identified 31 genes with Cn SNPs in C58/J mice associated with adult neurogenesis that were orthologous to the genes dysregulated in the radial glial cells of individuals with autism (Supplementary Table S2-2.4).

Discussion
Dysregulation in the neurogenic process is one of the underlying mechanisms associated with the neurobiological changes found in ASD 4,39 .Individuals with autism often exhibit atypical hippocampal anatomy and connectivity throughout their lifespan, which may be associated with alterations in the hippocampal neurogenic process 40 .
The adult neurogenic process in the mouse hippocampus is controlled by gene expression patterns and is strongly influenced by polymorphisms 17,22,23 .Therefore, we used an in-silico approach to identify singlenucleotide polymorphisms (SNPs) in genes relevant to adult hippocampal neurogenesis in the C58/J model of idiopathic autism.
Our results show that C58/J mice display non-synonymous SNPs (Cn SNPs) in 33 genes involved in the adult neurogenic process, according to the MANGO database.The majority of these genes are implicated in the regulation of proliferation, differentiation, and survival of neural precursors.This suggests that the C58/J strain may exhibit impairments in the early events of the hippocampal neurogenic process, as previously proposed for cortical neurogenesis in ASD 8,9,41,42 .
In line with this, C58/J mice also present Cn SNPs in 64 and 78 genes, corresponding to the signature genetic profile of NSC and neural progenitors, respectively, as reported by Artegiani et al. 33 .These Cn SNPs may interfere with the genetic regulation of NSCs during their transition from a dormant to a proliferative stage and subsequent differentiation into neural progenitors.Evidence of this is observed in Shank3b knockout mice, a model of genetic autism, where altered gene expression in NSCs resulted in increased quiescence and reduced neuroblast differentiation frequency 39 .
Our GO enrichment analysis of genes with Cn SNPs revealed enriched terms related to cell proliferation and the regulation of MAPK and WNT signaling pathways.Interestingly, experiments with iPSCs-derived neural progenitor cells from individuals with ASD have shown that dysregulation of cell proliferation is associated with alterations in the RAS/ERK and WNT signaling pathways 4,6 .
Other enriched GO terms were associated with metabolic processes and neuron death, critical mechanisms for neurogenesis progression and homeostasis 16,33,43 .We also identified genes with Cn SNPs that are important for later events in neurogenesis, such as migration, dendritogenesis, and maturation, according to the MANGO database.Additionally, we found enriched GO terms potentially involved in neuronal integration, including synaptic transmission and plasticity, as well as terms associated with neural processes that can be modulated by hippocampal adult neurogenesis, such as learning, memory, and behavior 12,44,45 .The genes with Cn SNPs were also found to be expressed in mature granule cells (GC) according to Hipposeq platform, suggesting their potential relevance for GC functionality.
Hence, these in-silico predictions based on the genetic variants of C58/J mice may imply that adult neurogenesis in the hippocampus is dysregulated at different stages of the neurogenic process: during the proliferation of NSC and neural progenitors, neuronal differentiation, neuronal death, migration, and morphological and functional maturation of neuroblasts into GC.
The GO enrichment analysis based on the TRANSFACT database revealed that the genes with Cn SNPs in C58/J mice possess binding motifs for specific transcription factors (TFs).This suggests that many of these genes may share regulatory pathways to control their expression during the adult neurogenic process.However, there is limited evidence regarding the specific participation of these TFs, such as E2F1 and Sp1, in the regulation of gene expression during adult neurogenesis 46,47 .
Interestingly, one of the genes with a high number of Cn SNPs in C58/J mice, Disc1 (Disrupted-inschizophrenia 1), has a strong participation in proliferation, morphogenesis, differentiation, and migration during hippocampal adult neurogenesis 48 .Another gene, Fmn2 (Formin-2), also showed enrichment with Cn SNPs and is essential for promoting neural progenitor proliferation 49 .Moreover, given that the Disc1 and Fmn2 genes carry Cn SNPs predicted to significantly damage the protein structure, it is possible that the function of both proteins is compromised in the C58/J strain.Therefore, future studies should include functional analyses to assess the potential protein alterations.
We further tested whether the in-silico predicted changes in adult neurogenesis of the C58/J strain could be related to in vivo hippocampal neurogenesis by analyzing the number and distribution of BrdU + /DCX + cells in young adult (6-weeks-old) C58/J mice compared to the WT strain C57BL/6J.
In agreement with the Cn SNPs we identified in genes involved in cell proliferation, differentiation, and survival, we observed an overall lower number of BrdU + /DCX + cells in the DG of C58/J mice.This suggests a reduced rate of neuronal-lineage cells (Type 2b and Type 3 cells) that underwent proliferation and differentiation, or that had a lower survival rate within the 2-weeks evaluation period.However, since DCX only allows us to identify Type 2b progenitors and Type 3 neuroblasts, we cannot rule out the possibility of impaired proliferation in NSC (Type 1 cells) and neural intermediate progenitors (Type 2a cells).Additionally, cell death also plays a crucial role in regulating the number of neuroblasts that survive and mature.Therefore, further experiments in autism models are required to evaluate the populations of NSC and Type 2a cells, as well as the death rate of hippocampal newborn neurons.
Maturation trajectories of newborn neurons are associated with their migration from the SGZ to the GCL, accompanied by morphological modifications 16,21  www.nature.com/scientificreports/mice indicated a high prevalence of immature ABC-type cells within the SGZ, which may correspond to Type 2b progenitors and early neuroblasts that remained close to their site of origin.Consistent with this, the C58/J strain exhibited a reduced proportion of EF-type cells in the GCL, which may represent late neuroblasts with morphological maturation signs, such as the outgrowth of apical branched dendrites and migration to the GCL.These observed changes may be associated with a delay in the maturation trajectory of newborn neurons in the DG of C58/J mice, potentially linked to the predicted alterations in the expression of relevant genes involved in hippocampal neurogenesis.Nevertheless, it is important to point out that DCX-expressing newborn neurons with complex morphological features do not represent cells in a fully mature state.Therefore, in future experiments, we will evaluate further maturation steps by sacrificing the animals at later time points after BrdU administration and by analyzing the expression of proteins that appear after DCX such as calretinin and calbindin.Additionally, we will further characterize the electrophysiological properties of these new neurons to analyze their potential incorporation into the established circuits of mature granule cells in the DG.
The C58/J strain may not exhibit changes in the migration trajectory of newborn neurons, as we did not observe differences in the number of BrdU + /DCX + cells within the hilus, a common characteristic of ectopic migration 50 .However, we cannot rule out the possibility of an altered migration ability of newborn cells in C58/J mice, considering the presence of Cn SNPs in the Disc1 gene.Proper migration of newly born neurons in the mouse DG relies on Disc1 expression during both hippocampal development and adult neurogenesis [51][52][53] .
Furthermore, we observed differential changes in the distribution and morphological maturation of neuroblasts in C58/J mice depending on the region of the DG (the crest, the suprapyramidal and infrapyramidal blades).It has been proposed that topographic variations in neurogenesis along the transverse axis of the DG (suprapyramidal vs. infrapyramidal blades) are necessary for the formation of a gradient of immature neurons/ mature granule cells, which is involved in the specific organization of hippocampal neural circuits [54][55][56] .
Alterations in hippocampal neurogenesis have been found in other murine models of autism 40,57 .For instance, the BTBR strain exhibits a reduced number of hippocampal newborn neurons 58 , and the valproic acid (VPA)induced mouse model shows a decreased ratio of neuronal differentiation in the hippocampus 59 .Interestingly, in the Auts2 knockout mice, the induction of the expansion of hippocampal newborn neurons rescued impaired social novelty recognition 60 .
Finally, we investigated whether changes in hippocampal adult neurogenesis in the C58/J model could be linked to ASD etiology by searching for orthologous human genes previously associated with autism.According to the SFARI GENE database 37 , we identified 22 ASD-related genes that are orthologous to the genes with variants in C58/J mice involved in adult neurogenesis.Among them, 8 genes showed a high score for ASD association.Additionally, we identified 31 genes with Cn SNPs in C58/J mice that were orthologous to genes dysregulated in the radial glial cells of individuals with autism, as reported by Kim et al. 39 .
In conclusion, these results suggest that the C58/J model of idiopathic autism exhibits a set of genes with nonsynonymous SNPs associated with the regulation of adult neurogenesis, some of which have human orthologous genes previously implicated in ASD.These genetic variants could contribute to the observed alterations in adult neurogenesis in the DG of C58/J mice.Thus, the dysregulation of the neurogenic process in the juvenile hippocampus should be further evaluated as a contributing mechanism involved in the etiology of ASD.

Animals
Two female and three male mice (n = 5 animals per group), 4 weeks old, from C57 BL/6J and C58/J strains, were used for the in vivo experiments.The animals were purchased from The Jackson Laboratory (BHB, ME, USA).The litters, grouped by sex, were housed in individual cages and kept on a commercial pelletized diet (T.G. rodent diet T2018S.15,Envigo), ad libitum.They were maintained under a reversed 12:12 h light/dark cycle, with lights on from 19:00 to 7:00.www.nature.com/scientificreports/ the various events and cell stages that comprise the adult neurogenic process 17 (See Supplementary Table S1 for MANGO criteria and definitions).
In addition, we consulted the gene dataset previously reported by Artegiani et al. 33 .This dataset provides a genetic characterization of neural stem cells (NSC) and neural progenitors from the neurogenic niche in the DG of 6-10-weeks-old female and male mice, obtained through single-cell RNA sequencing (data of the genes differentially expressed in NSC and neural progenitors is available in the supplementary material from the work of Artegiani) 33 .
To investigate whether genes associated with adult neurogenesis carry single-nucleotide polymorphisms (SNPs) in the genome of the C58/J model of autism, the MANGO, NSC, and progenitor's gene datasets were submitted to the SNP data retrieval utility tool of the Mouse Phenome Database (MPD) 28,66 .We used the Sanger4 Dataset (Sanger SNP and indel data, 89 + million locations, 37 inbred strains of mice, 2017) 29 to detect coding non-synonymous (Cn) SNPs, leading to changes in the coding amino acid sequence, in both adult females and males of the C58/J strain in comparison with the control C57 BL/6J strain.
Subsequently, we performed a Gene Ontology (GO) enrichment analysis using the g:Profiler platform (version e108_eg55_p17_9f356ae) 34 to identify additional biological functions associated with the neurogenesis-related genes with Cn SNPs in C58/J mice.The GO enrichment analysis covered the Biological process and Cellular component categories.We also used the TRANSFACT database to detect DNA-binding motifs for transcription factors [34][35][36] .
The potential impact on the protein structure of Cn SNPs was predicted using the PolyPhen-2 platform 31 .PolyPhen-2, trained by the Humdiv dataset model, extracts various sequence and structure-based features of an amino acid substitution to perform a probabilistic classification of the functional significance of an allele replacement.This classification is based on pairs of false positive rate (FPR) thresholds: benign (indicating a low probability of protein damage), possibly damaging (indicating a less confident prediction of protein damage), or probably damaging (a more confident prediction of protein damage) 31 .
To evaluate the expression of genes with Cn SNPs in the granule cells of the mouse DG, we used the Hipposeq platform containing the gene expression characterization of excitatory cell classes in the hippocampal formation of mice from both sexes 32 .An analysis for multiple genes expression based on the Dorsal-ventral survey of hippocampal principal cells dataset 32 was requested.The selected cell population was Dorsal DG granule cells.Data was retrieved as FPKM (fragments per kilobase of transcript per million reads mapped).
To identify whether genes with Cn SNPs display human orthologous genes implicated in the etiology of ASD, we searched for genes associated with autism spectrum disorder in the SFARI GENE database 37 containing relevant up-to-date information.In addition, we searched for these genes in the repository reported by Kim and collaborators, which contains a group of genes that were found to be differentially expressed in radial glial cells (Pax6 + or Vimentin +) from the prefrontal cortex of individuals with ASD in comparison to neurotypical subjects 39 .

Statistical analysis
The data obtained from the quantification of BrdU + /DCX + cells from both strains passed the Shapiro-Wilk test for normal distribution.
To compare the estimated number and percentage of BrdU + /DCX + cells based on their distribution within DG regions (crest, suprapyramidal and infrapyramidal blades), layers (SGZ, GCL, and hilus), and their morphological classification (ABC, D, and EF types) between control and C58/J mice, a two-way Analysis of Variance (ANOVA) was performed, followed by Sidak's correction for multiple comparisons.A p-value below 0.05 was considered statistically significant.The results were plotted and statistically analyzed using RStudio 4.2.3 and GraphPad Prism 8.0.2 software.
Data of the quantification of BrdU + /DCX + cells were also analyzed by comparing both sexes within each strain, but we did not observe statistically significant changes depending on the sex effect (Two-way ANOVA followed by Sidak's correction).
For the GO enrichment analysis, g:Profiler (version e108_eg55_p17_9f356ae) considers a result statistically significant if it corresponds to an experiment-wide threshold of a = 0.05 after applying the g:SCS multiple testing correction method, as reported by the platform 34 .The enriched GO terms were plotted using the log10 of p-adjusted values.The Gene Ratio was calculated by dividing the number of requested genes found in the functional category by the number of genes in the background genome (experimental gene set).The Rich Factor was calculated by dividing the number of requested genes found in the functional category by the total number of genes within that specific functional category.

Figure 1 .
Figure 1.Neurogenesis-associated genes with coding non-synonymous SNPs in the C58/J strain.In silico analysis of coding nonsynonymous (Cn) single-nucleotide polymorphisms (SNPs) in genes relevant to adult neurogenesis in the mouse DG. (A) Among the 397 genes associated with processes (left) and cell stages (right) involved in adult neurogenesis, according to the MANGO database (data inside the large circle), 33 genes carried at least one Cn SNP in C58/J mice (data inside the small circle).(B) Among the genes characterizing the genetic profile of neural stem cells (NSC) (591 genes) and neural progenitors (1065 genes) in the mouse DG, as reported by Artegiani (2017) (data inside the large circle), 142 genes carried at least one Cn SNP in C58/J mice, with 64 genes corresponding to NSC and 78 genes to the neural progenitors (data inside the small circle).(C) The 33 genes reported by the MANGO database display from 1 to 16 Cn SNPs in C58/J mice in comparison with the control C57BL/6 J strain.(D) The genes with Cn SNPs in C58/J mice are expressed during specific neurogenic processes (upper panel: proliferation, differentiation, survival, dendritogenesis, migration, maturation) and throughout different cell stages (inferior panel: stem cells [Type 1], undetermined progenitors [Type 2a], determined progenitors [Type 2b], neuroblast-like cells [Type 3], immature neurons, mature neurons, DCX-positive cells), according to the MANGO database.(E) 64 and 78 genes corresponding to the NSC (upper panel) and neural progenitors' (inferior panel) datasets, respectively, display from 1 to 23 Cn SNPs in C58/J mice.
Figure 1.Neurogenesis-associated genes with coding non-synonymous SNPs in the C58/J strain.In silico analysis of coding nonsynonymous (Cn) single-nucleotide polymorphisms (SNPs) in genes relevant to adult neurogenesis in the mouse DG. (A) Among the 397 genes associated with processes (left) and cell stages (right) involved in adult neurogenesis, according to the MANGO database (data inside the large circle), 33 genes carried at least one Cn SNP in C58/J mice (data inside the small circle).(B) Among the genes characterizing the genetic profile of neural stem cells (NSC) (591 genes) and neural progenitors (1065 genes) in the mouse DG, as reported by Artegiani (2017) (data inside the large circle), 142 genes carried at least one Cn SNP in C58/J mice, with 64 genes corresponding to NSC and 78 genes to the neural progenitors (data inside the small circle).(C) The 33 genes reported by the MANGO database display from 1 to 16 Cn SNPs in C58/J mice in comparison with the control C57BL/6 J strain.(D) The genes with Cn SNPs in C58/J mice are expressed during specific neurogenic processes (upper panel: proliferation, differentiation, survival, dendritogenesis, migration, maturation) and throughout different cell stages (inferior panel: stem cells [Type 1], undetermined progenitors [Type 2a], determined progenitors [Type 2b], neuroblast-like cells [Type 3], immature neurons, mature neurons, DCX-positive cells), according to the MANGO database.(E) 64 and 78 genes corresponding to the NSC (upper panel) and neural progenitors' (inferior panel) datasets, respectively, display from 1 to 23 Cn SNPs in C58/J mice.

Figure 3 .
Figure 3. Evaluation of BrdU + /DCX + newborn neurons in the DG of the C58/J and WT strains.The representative confocal images show immunofluorescence for BrdU (magenta) and DCX (green) in cells of the DG from WT mice (left panel) and C58/J mice (right panel).The panels correspond to the crest (A), suprapyramidal (B), and infrapyramidal (C) blades in both strains.Each panel shows I) DCX channel, II) BrdU channel, III) merged channels, and IV) orthogonal xz and yz views of the BrdU + /DCX + cells indicated by red arrows.The DG layers (SGZ, GCL, ML, and hilus) are indicated in blue legends.The images are maximal intensity projections from 60x magnification z-stacks.The orthogonal views are shown as enlarged insets from the 60x magnification.Scale bars: 50 µm.