Article14 December 2015Open Access Cell cycle networks link gene expression dysregulation, mutation, and brain maldevelopment in autistic toddlers Tiziano Pramparo Tiziano Pramparo Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Michael V Lombardo Michael V Lombardo Department of Psychology, University of Cyprus, Nicosia, Cyprus Center for Applied Neuroscience, University of Cyprus, Nicosia, Cyprus Autism Research Centre, Department of Psychiatry, University of Cambridge, Cambridge, UK Search for more papers by this author Kathleen Campbell Kathleen Campbell Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Cynthia Carter Barnes Cynthia Carter Barnes Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Steven Marinero Steven Marinero Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Stephanie Solso Stephanie Solso Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Julia Young Julia Young Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Maisi Mayo Maisi Mayo Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Anders Dale Anders Dale Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Clelia Ahrens-Barbeau Clelia Ahrens-Barbeau Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Sarah S Murray Sarah S Murray Scripps Genomic Medicine & The Scripps Translational Sciences Institute (STSI), La Jolla, CA, USA Department of Pathology, University of California San Diego, La Jolla, CA, USA Search for more papers by this author Linda Lopez Linda Lopez Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Nathan Lewis Nathan Lewis Novo Nordisk Foundation Center for Biosustainability at the UCSD School of Medicine, and Department of Pediatrics, University of California San Diego, La Jolla, CA, USA Search for more papers by this author Karen Pierce Karen Pierce Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Eric Courchesne Corresponding Author Eric Courchesne Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Tiziano Pramparo Tiziano Pramparo Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Michael V Lombardo Michael V Lombardo Department of Psychology, University of Cyprus, Nicosia, Cyprus Center for Applied Neuroscience, University of Cyprus, Nicosia, Cyprus Autism Research Centre, Department of Psychiatry, University of Cambridge, Cambridge, UK Search for more papers by this author Kathleen Campbell Kathleen Campbell Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Cynthia Carter Barnes Cynthia Carter Barnes Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Steven Marinero Steven Marinero Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Stephanie Solso Stephanie Solso Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Julia Young Julia Young Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Maisi Mayo Maisi Mayo Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Anders Dale Anders Dale Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Clelia Ahrens-Barbeau Clelia Ahrens-Barbeau Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Sarah S Murray Sarah S Murray Scripps Genomic Medicine & The Scripps Translational Sciences Institute (STSI), La Jolla, CA, USA Department of Pathology, University of California San Diego, La Jolla, CA, USA Search for more papers by this author Linda Lopez Linda Lopez Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Nathan Lewis Nathan Lewis Novo Nordisk Foundation Center for Biosustainability at the UCSD School of Medicine, and Department of Pediatrics, University of California San Diego, La Jolla, CA, USA Search for more papers by this author Karen Pierce Karen Pierce Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Eric Courchesne Corresponding Author Eric Courchesne Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA Search for more papers by this author Author Information Tiziano Pramparo1, Michael V Lombardo2,3,4, Kathleen Campbell1, Cynthia Carter Barnes1, Steven Marinero1, Stephanie Solso1, Julia Young1, Maisi Mayo1, Anders Dale1, Clelia Ahrens-Barbeau1, Sarah S Murray5,6, Linda Lopez1, Nathan Lewis7, Karen Pierce1 and Eric Courchesne 1 1Department of Neurosciences, UC San Diego Autism Center, School of Medicine University of California San Diego, La Jolla, CA, USA 2Department of Psychology, University of Cyprus, Nicosia, Cyprus 3Center for Applied Neuroscience, University of Cyprus, Nicosia, Cyprus 4Autism Research Centre, Department of Psychiatry, University of Cambridge, Cambridge, UK 5Scripps Genomic Medicine & The Scripps Translational Sciences Institute (STSI), La Jolla, CA, USA 6Department of Pathology, University of California San Diego, La Jolla, CA, USA 7Novo Nordisk Foundation Center for Biosustainability at the UCSD School of Medicine, and Department of Pediatrics, University of California San Diego, La Jolla, CA, USA *Corresponding author. Tel: +1 858 534 6914; E-mail: [email protected] Molecular Systems Biology (2015)11:841https://doi.org/10.15252/msb.20156108 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Abstract Genetic mechanisms underlying abnormal early neural development in toddlers with Autism Spectrum Disorder (ASD) remain uncertain due to the impossibility of direct brain gene expression measurement during critical periods of early development. Recent findings from a multi-tissue study demonstrated high expression of many of the same gene networks between blood and brain tissues, in particular with cell cycle functions. We explored relationships between blood gene expression and total brain volume (TBV) in 142 ASD and control male toddlers. In control toddlers, TBV variation significantly correlated with cell cycle and protein folding gene networks, potentially impacting neuron number and synapse development. In ASD toddlers, their correlations with brain size were lost as a result of considerable changes in network organization, while cell adhesion gene networks significantly correlated with TBV variation. Cell cycle networks detected in blood are highly preserved in the human brain and are upregulated during prenatal states of development. Overall, alterations were more pronounced in bigger brains. We identified 23 candidate genes for brain maldevelopment linked to 32 genes frequently mutated in ASD. The integrated network includes genes that are dysregulated in leukocyte and/or postmortem brain tissue of ASD subjects and belong to signaling pathways regulating cell cycle G1/S and G2/M phase transition. Finally, analyses of the CHD8 subnetwork and altered transcript levels from an independent study of CHD8 suppression further confirmed the central role of genes regulating neurogenesis and cell adhesion processes in ASD brain maldevelopment. Synopsis Analyses of the relationship between blood gene expression and brain size in 142 Autism Spectrum Disorder (ASD) and control male toddlers reveal peripheral blood signatures of ASD and genetic mechanisms underlying abnormal early neural development. In ASD, the correlation of brain size measures with cell cycle and protein folding gene networks is lost, while cell adhesion networks significantly correlate with brain size. Cell cycle networks detected in blood are highly preserved in the human brain and are upregulated during prenatal states of development. In ASD, cell cycle networks display changes in topological organization and these alterations are more pronounced in bigger brains. A predicted high-confidence network indicates dysregulation of neurogenesis and cell adhesion processes in ASD brain development. Introduction Autism Spectrum Disorder (ASD) is a heritable disorder involving early brain maldevelopment (Courchesne et al, 2011a). The brain at young ages is abnormal in a myriad of ways including brain overgrowth with an anterior/frontal to posterior cortical gradient in the majority, but undergrowth in a minority, during the first years of life (Courchesne et al, 2007); this shift upward in brain size distribution is quantitative and not categorical. Brain weight at autopsy is also shifted upward with heavier than the normal mean for an estimated 80% of 2–16 year olds, but lighter for a minority (Redcay & Courchesne, 2005; Courchesne et al, 2011b). A small sample of young ASD boys with heavy brain weight exhibited an excess of 67% neurons in the prefrontal cortex, which mediates social, communication and cognitive development (Courchesne et al, 2011b). The excess of neurons in enlarged brains points to potential dysregulation of mechanisms that govern cerebral cortical neuron number during second trimester development. Indeed, gene expression studies of prefrontal cortex in young ASD postmortem cases report dysregulation of gene expression associated with cell production, DNA-damage response, and apoptosis (Chow et al, 2012). Recently, alterations were detected in cell cycle timing and excess cell proliferation in neuroprogenitor cells derived from fibroblasts of living ASD patients who displayed early brain overgrowth (Marchetto et al, unpublished data). Disrupting mechanisms regulating cell number in the second trimester has long been theorized to play a role in brain maldevelopment in ASD (Courchesne et al, 2001) because mutant mouse model studies show that cell cycle molecular machinery governs the overall size of the brain (Nakayama et al, 1996; Ferguson et al, 2002). In fact, several pathological changes characteristic of ASD were recently modeled in mouse (WDFY3 loss of function) and displayed abnormal decreases in cell cycle timing, excess radial glia cell proliferation, prenatal brain overgrowth, and an abnormal anterolateral to posteromedial gradient of cortical overgrowth; interestingly, it also displays focal laminar dysplasia associated with mis-migrated cells (Orosco et al, 2014). This latter pathology may also parallel the report of focal prefrontal and temporal cortical laminar defects in 91% of young ASD males and females (Stoner et al, 2014). However, additional genes are known to be associated with either abnormal brain enlargement or reduction in animal models (Ellegood et al, 2014) and/or rare individual ASD cases (O'Roak et al, 2012), suggesting additional mechanisms underlying ASD. Recent genomic analyses of high-confidence genes in ASD (Parikshak et al, 2013; Willsey et al, 2013; De Rubeis et al, 2014) also point to dysregulation of cortical neuron number, laminar development, and cell cycle in the prefrontal cortex during second and third trimesters. While each of these genes occurs only in rare individual ASD cases, cycle cell dysregulation functions may commonly disrupt development in the second trimester in ASD. While many high-confidence ASD genes regulate downstream transcriptional programs including cell cycle functions and proliferation, such as CHD8, in general they are not cell cycle genes per se. This suggests that effects of high-confidence ASD genes on cell proliferation and brain size may be quantitative and continuous and not categorical. Many genetic and non-genetic defects could disrupt cell cycle, with changes in signaling and transcriptional activity, which could lead to variations in cell number and brain size. Unfortunately, it remains infeasible to directly test the impact of cell cycle changes on cell number and brain size in ASD in vivo or with postmortem approaches. This is because cell cycle activity changes with development, and assays that test cell cycle activity in older postmortem tissue provide only indirect information about its function during fetal development. Moreover, the scarcity of postmortem ASD cases further limit getting even indirect evidence of cell cycle dysfunction on brain size from this avenue. These barriers hinder the study of relationships between cell cycle disorganization and brain size variance in ASD during early development. However, we note that genetic disruption of cell cycle network organization could be detectable in multiple tissues at different ages. While the physiological response of a genetic perturbation often varies with tissue type and age, the presence of disruption may nonetheless be detectable and quantifiable across types and ages. That is, detection in one tissue type at one age, such as leukocytes in infants, may help the search for the presence of disruption in other inaccessible cell types and ages, such as fetal neuroprogenitor cells. Of note, the GTEx Consortium reported in Science that cell cycle gene expression networks are present in all tissues, including brain and blood (GTEx Consortium, 2015). Therefore, we took a systems biology approach to analyzing gene co-expression patterns in blood leukocyte samples of ASD and control infants and toddlers in order to examine how variation in co-expression modules are associated with variation in brain size at very young ages in ASD. Here, we show that gene expression profiles from leukocytes at very young ages may be a biomarker of early brain growth deviance in ASD. Furthermore, we use cell cycle networks as an entry point to elucidate perturbation of transcriptional networks associated with smaller and bigger brains. Our findings of network dysfunction are integrated with recent genomic studies describing genes frequently mutated in ASD, thus providing compelling evidence that cell cycle networks may indeed be a point of convergence for gene expression dysregulation, mutation, and early brain maldevelopment in ASD. Results We tested the hypothesis that blood-based gene expression profiling may reveal biological signatures relevant to neurodevelopment and that such signatures may differ between ASD and control toddlers. Leukocyte RNA levels were analyzed in relationship to total brain volume (TBV) using an established approach based on gene co-expression (WGCNA; Fig 1A and B; Langfelder & Horvath, 2008). This method elucidates patterns of altered gene expression, organized as networks of co-expressed genes, and provides insights into relationships of genes with disease-related endophenotypes or traits. Furthermore, it provides metrics to understand the details of network perturbation (Langfelder & Horvath, 2007, 2008; Fig 1C). We leveraged network metrics to understand whether network perturbation differentially affected smaller and bigger brains in ASD toddlers as compared to controls (Fig 1D). Lastly, we used a reverse genetic approach to frame our findings with recent evidence from genomic studies reporting high-confidence genes of ASD (De Rubeis et al, 2014). Figure 1. Schematic of the approach used in the study Blood gene expression was analyzed in relationship with neuroanatomic measures using a co-expression network-based approach (WGCNA). The distribution of neuroanatomic measure was normal and not significantly different between ASD and control toddlers. The analysis of co-expression was combined with all available samples. Data from the combined network-based analysis was further investigated in each ASD and control group separately using a linear model. Network features, calculated from the WGCNA co-expression analysis in relationship to brain size, were used to dissect alterations of network patterns in ASD brains. Network features were also used to characterize smaller and bigger brains in each study group. Download figure Download PowerPoint Different gene networks associate with brain size in ASD and control toddlers Analyses were run using processed gene expression data (Pramparo et al, 2015) that included 12,208 unique gene-probes from 87 ASD and 55 control male subjects ages 1–4 years. The majority of subjects were of Caucasian origin and Pearson's chi-squared test showed no significant difference in race characteristics between ASD and control (X2[5] = 7.98, P = 0.1569). Multivariate regression analysis showed no variance explained by differences in race and ethnicity characteristics between ASD and control subjects and age variance was accounted in downstream analyses. Unsupervised WGCNA resulted in 22 co-expression modules (Appendix Fig S1). Each module was given an arbitrary color name and was summarized by a metric known as the module eigengene (ME), which is the first principal component of the module (i.e., axis capturing the majority of variation in expression in the module). After degree-preserving random shuffling, it was determined that all 22 modules were significantly detected above chance levels (see Materials and Methods and Appendix Fig S2). Module preservation analysis between the two separate datasets (ASD/control) displayed high-preservation scores, suggesting that the combined analysis was not confounded by differences in networks structure of the two datasets (Appendix Fig S3 and Appendix Table S1). Module eigengene values from each of the 65 ASD and 38 control subjects, who also had MRI scans, were used in linear correlation tests in which we related RNA levels to brain size. TBV measures were age-corrected (see Materials and Methods) and showed normal distributions with no statistically significant differences in mean and variance between the two groups (Fig 2A). Seven modules were significantly correlated with TBV measures across all subjects (FDR < 0.05) with the greenyellow and grey60 gene modules displaying the strongest correlations (Fig 2B). Permutation analysis with randomly generated MEs (see Materials and Methods) demonstrated that these associations were significant against chance for all but the yellow module (Appendix Fig S4). Figure 2. Analysis of gene networks associated with variation in brain size in ASD and control toddlers A. Distributions of brain size as indexed by total brain volume (TBV) in ASD and control toddlers used in the co-expression analysis (WGCNA). T, value from t-test; F, value from Levene's test. B. Module eigengenes (MEs) from the combined WGCNA are linearly correlated with TBV measures in all brains, ASD and control groups. P-value is in parenthesis and adjusted P-value (q-value) is < 0.05 for all seven modules. Significant associations after 10,000 permutation tests are provided in Appendix Figs S4 and S5. C. Metacore enrichment scores of the seven (7) modules initially related to brain size variation across all subjects. Each module is called by its assigned color and represents the top process network obtained by the enrichment analysis in Metacore GeneGO (see also Dataset EV1). D–F. (i) Linear modeling of module eigengenes (MEs) by TBV measures in control (blue) and ASD (red) toddlers. See also Fig 2B for cor and P-values. (ii) Linear modeling of GS by GC to display changes in network organization relevant to brain size. (iii) The top 30 genes with highest values for GS and GC were compared between ASD and control. Purple indicates the number of genes that moved away from the top 30 rank position between the two groups (Different genes), and grey indicates the number of genes that did not (Common genes). Significance codes: ***P-value < 0.001; **P-value < 0.01; cor, correlation coefficient; ns, not significant. Download figure Download PowerPoint To identify gene networks that correlated with brain size within each diagnostic group, we computed Pearson's r correlation statistics between each of the seven MEs and TBV measures in ASD and control toddlers, separately. In control toddlers, the greenyellow and grey60 MEs were significantly correlated with age-corrected TBV (Fig 2B), while brain size in ASD toddlers displayed significant linear correlations with the salmon, turquoise, and cyan MEs (Fig 2B; see Appendix Table S2 for bootstrapped 95% confidence intervals). We restricted further analyses to only modules showing the strongest effects on brain size (i.e. r > 0.3, P < 0.05, FDR < 0.05), which included greenyellow, grey60, and salmon modules. These effects were found to be independent of age (Appendix Table S3) and confirmed to be significant against chance after permutation analysis (Appendix Fig S5 and see Materials and Methods). WGCNA on the separate ASD and control datasets also confirmed that these three gene networks were the strongest signal associated with TBV variation in each group (Appendix Figs S6 and S7 and Dataset EV1). We next used permutation tests to examine whether the strength of these correlations (MEs-TBV) significantly differed between the two groups. However, correlations were not significantly different between groups for greenyellow P = 0.33, but were at or beyond trend level significance for grey60 (P = 0.06) and salmon (P = 0.01; see Fig 2Di–Fi and Appendix Table S4). To investigate the biological functions of these modules, we ran pathway enrichment analysis in Metacore GeneGO using a threshold of FDR q < 0.05. The greenyellow and grey60 modules were enriched in genes with cell cycle and protein folding functions, respectively, while genes in the salmon module were enriched in cell adhesion functions (Fig 2C and Dataset EV1). This enrichment remained significant after filtering for expression in fetal and adult brain tissue using the Metacore GeneGo database (Dataset EV1). The other modules with modest correlations displayed enrichment in translation, inflammation, and cytoskeleton rearrangement functions (Fig 2C and Dataset EV1). Network perturbation in ASD affects gene connectivity and relevance for brain size In addition to quantifying gene module summary measures like the module eigengene and its relationship to brain size, we also used two gene-level metrics (gene significance and gene connectivity) to assess associations with brain size. Gene significance (GS) is defined as the correlation between gene expression and a trait (i.e., TBV), thus providing a measure of “significance or relevance” of a particular gene to variation in a trait such as TBV. Gene connectivity (GC) is a connectivity measure indicating how strongly connected (i.e., correlation strength) is a particular gene with all other genes within the module. Higher GC values are indicative of central or ‘hub’ genes, whereas genes with lower GC are oriented around the periphery of the co-expression module. Examining the correlation between GC with GS values for each gene allows for insight into understanding how metrics of a gene's organization within a network (i.e. gene connectivity) may be associated to its relevance with brain size (i.e., gene significance). GS-GC correlations were stronger in control compared to ASD (i.e. more positive) in both the cell cycle (control: r = 0.64; ASD: r = 0.42; z = 3.47) and protein folding modules, (control: r = 0.64; ASD r = 0.40; z = 3.19) (see Fig 2Dii–Fii). Thus, as a gene becomes more highly connected with other genes within the cell cycle and protein folding modules, it also becomes more relevant to (or has stronger impact on) TBV, and this relationship is stronger in control than ASD. For the cell adhesion module, the ASD group showed a stronger correlation between GS and GC than the control group (ASD: r = 0.22; control: r = 0.002; z = 2.48; see Fig 2Dii–Fii). Thus, as a gene becomes more highly connected within the cell adhesion module, it becomes more relevant to brain size in ASD than in the control group. Along with the evidence showing generalized atypicality in GS in ASD (i.e., reductions in GS in cell cycle and protein folding modules, but increase in cell adhesion; Appendix Fig S8), this evidence supports the idea that GS is accompanied by a modest alteration in GC between groups, indicating that a gene's relevance to brain size covaries with changes in network organization in ASD (Fig 2Dii–Fii). With regard to the cell cycle network in particular, this network re-organization in ASD can be described as many high GC genes (i.e., hub-genes located more centrally within the network) with a reduced GS, but also many low GC genes (i.e. low-connectivity genes located around the periphery of the network) which displayed some of the highest GS levels (Fig 2Dii–Fii). Of the three modules, the cell cycle module displayed the most severe network re-organization. This can be shown through further analyses of the top 30 genes on each metric (GS and GC; see Materials and Methods). First, we ran Venn analyses to determine the gene overlap between ASD and control toddlers and found that the majority of the genes with highest GS were unique to each group, especially for the cell cycle gene network with 29 out of 30 genes being different between groups (Fig 2Diii and Dataset EV1). Then we investigated whether the top 30 GC genes (i.e., hub-genes) in the co-expression network were also the top GS genes in each group. Within the cell cycle module, 16 of the 30 hub-genes also possessed the top GS scores for controls, while in ASD only 5 of the 30 hub-genes were top GS genes (OR = 5.71, P = 0.004 CI = 1.72–18.94). The remaining 25 top GS genes in ASD had lower GC scores, and thus were considered ‘peripheral’ in the cell cycle co-expression network. While in controls these 25 peripheral genes displayed a strong positive association between GS and GC (r = 0.74, P = 2.7e-5), in ASD the directionality of the association flipped (r = −0.34, P = 0.098; Appendix Fig S9), resulting in a substantial group difference in correlation strength (z = 4.33, P = 1.51e-5). For the protein folding module, there were similar proportions of hub-genes displaying top GS scores in both groups (OR = 1.96, P = 0.198, CI = 0.70–5.48). However, among the peripheral genes (i.e. top GS genes with low GC scores), again there was a flip in directionality of GS-GC correlation (controls r = 0.89, P = 1.8-e-6; ASD r = −0.19, P = 0.45; z = 0.35, P = 8.60e-8; Appendix Fig S9). Likewise, the cell adhesion module also displayed similar proportions of hub-genes with top GS scores (OR = 1, P = 1, CI = 0.30–3.30). Peripheral genes (i.e., top GS genes with low GC scores) display similar GS-GC correlations across controls and ASD (controls r = 0.04, P = 0.85; ASD r = −0.35, P = 0.094; z = 1.34, P = 0.17 Appendix Fig S9). This evidence reinforced the findings of network re-organization and revealed a trend in gene expression relevance for brain size that shifts from central genes (hub-genes) in control toddlers to peripheral genes in ASD toddlers particularly within the cell cycle network. The overall network perturbation may underlie potential downstream consequences in overall transcriptional regulation. Cell cycle module is preserved and highly expressed during early stages of normal fetal brain development We next reasoned that if correct neurodevelopment relies on the tight modulation of gene networks driving brain size, changes in expression levels of these networks would likely be most damaging at early developmental stages. It is also expected that biological processes involved in cell proliferation (e.g., during the neuronal progenitor pool expansion) would be expressed at high levels at earliest ages and lowest during postnatal life when the brain structures have already been formed. Based on these hypotheses, we
