Multiscale modeling uncovers 7q11.23 copy number variation-dependent changes in ribosomal biogenesis and neuronal maturation and excitability

Abstract

Copy number variation (CNV) at 7q11.23 causes Williams-Beuren syndrome (WBS) and 7q microduplication syndrome (7Dup), neurodevelopmental disorders (NDDs) featuring intellectual disability accompanied by symmetrically opposite neurocognitive features. Although significant progress has been made in understanding the molecular mechanisms underlying 7q11.23-related pathophysiology, the propagation of CNV dosage across gene expression layers and their interplay remains elusive. Here we uncovered 7q11.23 dosage-dependent symmetrically opposite dynamics in neuronal differentiation and intrinsic excitability. By integrating transcriptomics, translatomics, and proteomics of patient-derived and isogenic induced neurons, we found that genes related to neuronal transmission follow 7q11.23 dosage and are transcriptionally controlled, while translational factors and ribosomal genes are posttranscriptionally buffered. Consistently, we found phosphorylated RPS6 (p-RPS6) downregulated in WBS and upregulated in 7Dup. Surprisingly, p-4EBP was changed in the opposite direction, reflecting dosage-specific changes in total 4EBP levels. This highlights different dosage-sensitive dyregulations of the mTOR pathway as well as distinct roles of p-RPS6 and p-4EBP during neurogenesis. Our work demonstrates the importance of multiscale disease modeling across molecular and functional layers, uncovers the pathophysiological relevance of ribosomal biogenesis in a paradigmatic pair of NDDs, and uncouples the roles of p-RPS6 and p-4EBP as mechanistically actionable relays in NDDs.

Keywords: Neurodevelopment; Neuroscience; Psychiatric diseases; Stem cells; Translation.

Figure 1. 7q11.23 isogenic iNeurons preserve dosage.

(A) Scheme of experimental design and generation of 7q11.23 isogenic lines. (B) Two-color FISH analyses using 7 alpha satellite probes (see Supplemental Methods) as a control for the chromosomal number (yellow) and ELN, a WBSCR gene (red). ELN showed 3 signals in 7Dup, 2 in isoCTL, and 1 in isoWBS, corresponding to the 7q11.23 copy number in respective clones. (C–E) WBSCR genes maintain the dosage at the RNA and protein levels. RNA-seq data for WBSCR genes are shown for both patient-derived and isogenic neurons for all 3 genotypes. Although GTF2I transcripts were not downregulated in isoWBS in the RNA-seq analysis, both the translatome and proteome data showed 7q11.23 dosage–dependent expression (Supplemental Figure 1, F and G) that was also confirmed by Western blot (Figure 1D and Supplemental Figure 1H), suggesting that the upregulation observed at the mRNA level is probably an artifact of sequencing of repetitive regions. Western blot results from the same neuronal preparation, run on 2 gels, are shown in D. GAPDH was used as a loading control. Quantification of Western blots (shown in D and Supplemental Figure 1H) is shown as relative expression in E. Non-normalized data are shown as mean ± SEM (n = 4). The statistical comparisons were done with 1-way ANOVA followed by Tukey’s multiple-comparison test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Figure 2. 7q11.23 hemideletion delays, whereas hemiduplication accelerates, neuronal differentiation.

(A) Diagram showing the timing of neuronal differentiation in 3 different neuronal models: STEMdiff-driven (dual-Smad-based; see Supplemental Methods) and Ngn2-driven iNeurons, and cortical brain organoids. The expected change in profiled markers in each model is schematized. Red, iPSCs; green, NPCs; blue, neurons. (B) Expression of stem markers (OCT4, SOX2, and NANOG) and NPC marker PAX6 in early NPCs (n = 3) measured by qPCR. (C) Representative immunofluorescence images of 30-day-old Ngn2-iNeurons from isoWBS, isoCTL, and 7Dup stained for the mature neuronal marker MAP2B (red) and with DAPI (blue). Scale bars: 100 μm. (D) Quantification of MAP2B fluorescence intensity versus the cell number in Ngn2-iNeurons assessed at 10, 20, and 30 days of 2 independent differentiations (14–18 fields of view). IsoWBS and 7Dup were normalized to controls. (E) Immunofluorescence in cryosections of cortical organoids from isogenic lines on days 18 and 50 for proliferative marker Ki67, neural progenitor marker PAX6, and neuronal postmitotic marker CTIP2. Scale bars: 50 μm. First row, quantification of Ki67: isoWBS n = 3 organoids, isoCTL n = 4, 7Dup n = 3; second row, quantification of PAX6: isoWBS n = 5 organoids, isoCTL n = 4, 7Dup n = 3; third row, quantification of CTIP2: isoWBS n = 5 organoids, isoCTL n = 4, 7Dup n = 3. Data points are organoids’ sections from 2 independent experiments. All data are shown as mean ± SEM. The statistical comparisons were done with 1-way ANOVA followed by Tukey’s multiple-comparison test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Figure 3. Robust transcriptional changes in translation- and neural transmission–related genes.

(A) Fold changes of DEGs in the merged analysis of isogenic and patient-derived lines (in either WBS vs. CTL, 7Dup vs. CTL, or regression on 7q11.23 copy number), showing robust transcriptional signatures that are largely symmetrically opposite between genotypes. SRP, signal recognition particle. (B) Enriched GO terms in the regression on 7q11.23 copy number. Similar terms are clustered (denoted by colors) and only the top term per cluster is shown. (C–E) Top DEGs associated with translation (C), ion channels and their regulation (D), or ASD (E). (F) Comparison of fold changes at the RNA and protein levels (for the union of genes found significant at either level), in each of the 3 comparisons performed. (G) Expression of the transcriptionally dysregulated translation genes that could also be measured at the proteome level, highlighting a buffering in 7Dup.

Figure 4. Posttranscriptional regulation.

(A) Cross-layer clustering of DEGs reveals distinct patterns of transcriptional and translational regulation, emphasizing condition-specific buffering and genes forwarded to the proteome. Buffering coefficients (capturing the reduction or amplification of the fold change at the protein level) for each condition are shown in the center. (B and C) Enriched GO terms in the forwarded (B) and buffered (C) clusters. (D) Translational buffering opposes the transcriptional dysregulation of genes encoding TOP mRNAs in 7Dup. Cumulative distribution plots comparing the fold changes in both conditions and across gene expression layers of genes encoding 5′ TOP mRNAs to that of other genes are shown. RPF, ribosome-protected fragments; TE, translation efficiency.

Figure 5. Genotype-specific dysregulation of p-RPS6, but not p-4EBP, in 30-day-old iNeurons.

(A) Simplified scheme of mTOR signaling. (B–G) Quantification of Western blot analyses for total RPS6 (B), p-RPS6 S240/S244, and p-RPS6 S235/S236 normalized to RPS6 levels (C and D respectively), 4EBP (E), p-4EBP Thr37/Thr46 (F), and p-4EBP Thr37/Thr46 normalized to 4EBP levels (G). The experiment was done on 6 different iNeuron preparations, differentiated in 2 different rounds of differentiation. (H) Representative Western blot quantification for B–G. Other quantified blots are shown in Supplemental Figure 5. (I and J) Quantification of puromycin incorporation assay (I; n = 5) with representative Western blot (J). Other quantified blots are shown in Supplemental Figure 4G. All data are shown as mean ± SEM. The statistical comparisons were done with 2-way ANOVA followed by Tukey’s multiple-comparison test (B–G) or 1-way ANOVA followed by Tukey’s multiple-comparison test (I). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Black asterisks indicate significance between treatments, whereas the red asterisks indicate significance between genotypes.

Figure 6. 7q11.23 CNV causes symmetrically opposite neuronal excitability dynamics.

(A) Bright-field images of CTL, WBS, and 7Dup patient-derived iNeurons. Scale bars: 200 μm. (B) Representative AP trains in response to steps of 5-pA depolarizing current lasting 500 ms from –60 mV in iNeuron recordings. (C) Quantitative analysis depicting the number of elicited APs in the current-clamp configuration in the 3 genotypes (WBS: 4 lines, n = 29 neurons; CTL: 3 lines, n = 40; 7Dup: 4 lines, n = 16) in response to increasing current steps (CTL vs. WBS: current step 35*, 45–60*, 75–85*; WBS vs. 7Dup: 35–40*, 45–50**, 55–80***, 85–95*). (D) Bar graph depicting the amplitude of elicited APs. (E–G) Membrane resistance was calculated in the current-clamp mode without current injection. Input resistance was calculated in voltage-clamp mode using a pulse test of 10 mV. Rheobase was calculated as the minimum current required to elicit 1 AP. (H) Bright-field images of isogenic iNeurons. Scale bars: 200 μm. (I) Representative AP trains in response to steps of 5-pA depolarizing current lasting 500 ms from –60 mV in isogenic iNeurons. (J) Quantitative analysis depicting the number of elicited APs in the current-clamp configuration in the isogenic iNeurons (isoWBS, n = 23 neurons; isoCTL, n = 22 neurons; 7Dup, n = 25 neurons) (isoCTL vs. isoWBS: 15–95****; isoCTL vs. 7Dup: 25–75****, 80–95**; isoWBS vs. 7Dup: 15–95****). (K) Bar graph depicting the amplitude-elicited APs. (L–N) Passive properties and rheobase of the iNeurons recordings from isogenic lines, calculated as above. Data are shown as mean ± SEM and are the average of 3 independent experiments. For comparing AP frequency, we used 2-way ANOVA followed by Tukey’s multiple-comparison test, while for comparing passive properties we used 1-way ANOVA followed by Tukey’s multiple-comparison test. *P < 0.05; ***P < 0.001.

Figure 7. REST mediates WBS pathophysiological phenotypes.

(A) Master regulator analysis of the 7q11.23 dosage–dependent genes based on transcription factor–curated targets. The x and y axes respectively indicate the magnitude and significance of the inferred changes in the activity, while the color and size respectively indicate the magnitude and significance of the change in expression of the factor at the RNA level. Factors in a box are consistent and statistically significant in both activity and expression. nES, normalized enrichment score. (B) The genes altered in WBS versus control and rescued by REST inhibition in isoWBS iNeurons are especially associated with potassium ion transmembrane transport and extracellular structural organization. (C) Heatmap showing potassium transport and translation-related genes that were consistently differentially expressed in WBS iNeurons and were rescued by REST inhibitor (RESTi) in isogenic lines. Note that for ease of comparison between the inhibition experiment and 7q11.23 CNV, the fold changes are shown relative to WBS.