TBRS models show increased V-NPC proliferation and mTOR signaling.

a, iPSCs carrying a p.R882H mutation in DNMT3A (882), generated from a TBRS patient and corrected to generate an isogenic control (C-WT). b, Control hESCs (WT), modified to mimic a TBRS patient with heterozygous p.P904Lmutation (904) via knock-in, which also created a model with small <10 bp mutations in both DNMT3Aalleles (KO). c, Two clonal iPSC lines were generated from a TBRS patient with heterozygous deletion of the DNMT3Agene (Del1/2) and compared with sex-matched unrelated control iPSCs (Con1/2). d, CRISPRi hPSC models were generated using two gRNAs targeting DNMT3A (G1/2) and compared with hPSCs without gRNA (KRAB). e-h, Quantification of DNMT3A protein expression in (e) 882, (f) 904, (f) KO, (g) Del1/2, and (h) G1/2 TBRS models versus matched controls. i-j, Schematic representation of (i) D-NPC specification and U) V-NPC specification. k-I, Representative images and quantification of relative neurosphere outgrowth (dotted line) from plated neurospheres (bold line) at D12 of (k) 822 and (1) 904 V-NPC specification versus matched controls. m-p, Representative images and quantification of Kl67 immunopositivity in (m) 822, (n) 904, (o) Del1 and (p) G1/2 V-NPCs versus matched controls. q-r, Quantification of relative (q) abundance and (r) phosphorylation of AKT, mTOR, and ribosomal protein S6 in TBRS V-NPCs (882, 904 and DeI1) versus matched controls (dotted line). s, Representative images and quantification of Kl67 positivity in 882 and C-WT V-NPCs after treatment with rapamycin (RAP) or vehicle control (DMSO). Data is represented as mean +/- SEM (e-h,m-s) or as distributions, with median (bold line) and upper and lower quartiles (dotted lines) indicated (k-I), and analyzed by Students I-test, versus matched controls (e-h,k-s). n=4 biological replicate experiments for all conditions, pValues: *p<0.05;**p<0.01;***p<0.001;****p<0.0001. Scale bars=200 µm (k-I) and 100 µm (m-p,s).

DNMT3A mediates epigenetic repression of pro-neuronal genes in V-NPCs

a, Principal component analysis of control (grey) and TBRS (882-blue and 904-orange) V-NPCs. b, Heatmap of shared DEGs across both 882 and 904 versus control (C-WT/WT) V-NPCs. c, Summary of GO enrichment analysis of DEGs upregulated in TBRS (882/904) versus control V-NPCs. d, Expression changes of genes associated with ‘cell proliferation’ and ‘synapse’ GO terms across TBRS (882, 904, G1/2) versus control V-NPCs. e, Changes in mCG/CG levels in C-WT versus 882 V-NPCs. f, Summary of GO enrichment analysis of 882 DEGs associated with an 882 hypo-DMR. g, Genomic browser views of mCG in C-WT and 882 V-NPCs, highlighting hypo-DMRs (red) associated with promoters of DEGs upregulated in 882 V-NPCs. h, mCG/CG differences at shared-DMRs (All), qualified by directionality of DMRs in both TBRS models (both-hyper or both-hypo) or specifically in the 882 model (882-hyper or 882-hypo). i, Summary of GO enrichment analysis of DEGs upregulated in TBRS (904 and 882) V-NPCs and associated with shared hypo-DMRs.Data is represented as mean +/-SEM (d,e,g,h) with individual biological replicates indicated (e,h). Data was analyzed by Student’s t-test versus isogenic controls (d,e) or by differential expression analysis (g). n=4 biological replicate experiments for all conditions, pValues: *p<0.05 (d) and *p<0.05; **p<0.01; ****p<0.0001(e,g).

TBRS V-ORGs show increased proliferation and neurogenesis.

a-b, Schematic representation of (a) D-ORG or (b) V-ORG differentiation. c, Representative images and quantification of Kl67+ cell fraction across TBRS and control V-ORGs. d-g, Representative images of (d) SOX2 and TUJ1 staining in TBRS and control V-ORGs, alongside quantification of (e) TUJ1+ area, (f) SOX2+ fraction (% of DAPI+ nuclei) and (g) TUJ1+ area normalized to SOX2+ cells. h-i, Representative images and quantification of the (h) ASCL1+ and (i) SST+ cell fraction across TBRS and control V-ORGs. All data is represented as distributions with median (bold line) and upper and lower quartiles (dotted lines) indicated and was analyzed by Mann-Whitney tests, comparing each TBRS model to its matched control. A minimum of 3 batches of organoids were prepared for each condition, with 3-9 organoids assessed per batch. pValues: *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001; n.s.-non-significant. Scale bars=100 µm (c,h-i) and 500 µm (d).

Increased H3K27me3 compensates for loss of DNA methylation specifically in R882H NPCs.

a-b, H3K27me3 peaks with significantly increased (hyper-OKS) or decreased (hypo-OKS) enrichment in 882 versus control (a) D-NPCs or (b) V-NPCs.c, Expression of genes associated with hypo-DMRs and hypo-DKSs in 882 NPCs (D and V) across neuronal differentiations. d, Summary of GO enrichment analysis of non-DEGs associated with hypo-DMRs and hypo-DKSs in 882 NPCs (D and V). e, Genomic browser views of mCG and H3K27me3 in 882 and control V-NPCs, highlighting (red) hypo-DMRs and hyper-DKSs associated with promoters of two genes. f, Quantification of EZH2 mRNA levels in C-WT D-NPCs after treatment with either scrambled siRNA (scram) or siRNAs targeting EZH2 (1-3), highlighting selection of siRNA 1 for further study (red). g-h, Quantification of expression changes in 10 genes associated with 882 D-NPC hypo-DMRs and hyper-DKSs, after treatment with (g) siRNAs or (h) an EZH2-specific inhibitor. i, Quantification of EZH2 mRNA levels in C-WT and 882 V-NPCs after lentiviral transduction with a construct overexpressing either EZH2 (OE) or GFP (GFP). j, Quantification of expression changes in 10 genes upregulated in 882 V-NPCs under OE or GFP conditions. Data represented as mean+/-SEM (f-j) and was analyzed by one-way ANOVA, with comparison of treatment versus (f-g) scram,(h) DMSO or (i-j) GFP control treated conditions. n=4 biological replicate experiments for all conditions, pValues: *p<0.05; **p<0.01; ****p<0.0001 (f,i) or *p<0.05 (g-h,j).

DNMT3A constrains neuronal maturation during interneuron differentiation.

a, Experimental paradigm for generating immature GABAergic neurons (V-INs) from ventrally patterned neuronal progenitors (V-NPCs). b, Principal component analysis of control (grey) and TBRS (882-blue and 904-orange) V-INs. c, Summary of GO enrichment analysis of DEGs upregulated in TBRS V-INs.d, Heatmap of genes associated with the ‘synapse’ GO term and upregulated in TBRS V-INs. e, Changes in mCG/CG at 882, 904, or shared (TBRS) V-IN hypo-DMRs and matched control V-INs. f, Changes in mCG/CG at TBRS-shared hypo-DMRs identified in V-NPCs or V-INs across both TBRS and control V-NPCs and V-INs. g, Representative example traces of control and TBRS D30 V-INs used to quantify neuronal morphology. h-i, Quantification of neuronal morphology in control and TBRS D30 V-INs, including (h) number of secondary neurites and (i) total neurite length. j-k, Representative images and quantification of the proportion of U) SST or (k) CALB1 positive neurons in day (D) 40 C-WT or 882 V-IN cultures. 1-p, Representative images and quantification of pre-synaptic marker (l,n) SYN1 and postsynaptic marker (1,o) PSD95 puncta density, and (m,p) VGAT, across TBRS and control D50 V-INs. Data was analyzed by Students I-test versus matched controls (e,f,h,i,n-p). n=4 biological replicate experiments for all conditions with a minimum of 20 neurons quantified for morphological measures per biological replicate, pValues: *p<0.05;**p<0.01;***p<0.001;****p<0.0001. Scale bars=100 µm (g), 50 µm U,k) and 10 µm (1,m).

TBRS-associated mutations in DNMT3A cause GABAergic neuron hyperactivity.

a, Representative example traces of action potentials in C-WT (grey) or 882 (blue) GABAergic neurons induced by 1 second of stimulation. b, Quantification of maximum spike number across C-WT and 882 GABAergic neurons during 1 second of stimulation. c, Resting membrane potential of C-WT and 882 GABAergic neurons. d-e, Quantification of action potential (AP) characteristics, including (d) AP halfwidth and (e) AP height of C-WT and 882 GABAergic neurons. f-g, Measurements of (f) sodium (Na) current density at 0 mV and (g) transient outward current after voltage step, both normalized by capacitance (pA/pf).h, Example traces of current changes in C-WT and 882 GABAergic neurons during voltage step from -80 to -10 mV. i, Representative example traces of action potentials in WT (grey) or 904 (orange) GABAergic neurons induced by 1 second of stimulation. j, Quantification of maximum spike number, measured in both glutamatergic and GABAergic neurons (combined data analysis), during 1 second of stimulation. Significance was calculated by Rank Sum Test and results are presented as mean+/-SEM with data points representing measurements taken from individual neurons, *pValue<0.05;**pValue<0.01;***pValue<0.001.

TBRS neuronal network dysfunction is driven by GABAergic neuron hyperactivity.

a, Representative raster plots of LD-MEA activity in C-WT or 882 GABAergic/glutamatergic neuron co-cultures from early (D55, left) to late (D80, right) time points, with synchronized activity indicated by red boxes. b-d, Characteristics of C-WT and 882 neuronal networks, highlighting increased (b) firing rate (F1,6=6.579, pValue<0.05, n=4) and (c) frequency of network bursting (F1,6=12.18, pValue<0.05,n=4) in 882 neuronal networks with no substantial changes in (d) network burst duration. e-g, Representative raster plots of HD-MEA activity (e) assessing the contribution of TBRS GABAergic neurons to neuronal network dysfunction after 50 days of differentiation, highlighting effects of TBRS GABAergic neurons on (f) firing rate and (g) appearance of neuronal bursting activity. h-o, HD-MEA activity assessing the contribution of TBRS GABAergic neurons to neuronal network dysfunction after 82 days of differentiation. h-i, Representative raster plots and quantification of changes in firing rate in cultures containing TBRS GABAergic neurons. j-k, Characterization of neuronal bursting parameters altered by TBRS GABAergic neurons, including U) the number of spikes per burst and the (k) inter-spike-interval (ISI) between spikes within neuronal bursts. 1-o, Characterization of neuronal network parameters altered in the presence of TBRS GABAergic neurons, including network burst (I) frequency and (m) duration alongside the (n) proportion of activity within network bursts and the (o) ISi of spikes within network bursts. Data was analyzed by: two-way ANOVA, with posthoc Tukey’s multiple comparison testing (b-d), Student’s t-test (f,g,i,j,k,l,n) or Mann-Whitney test (m,o), versus isogenic controls. Data is represented as mean+/-SEM (b-d,f,g,i,j,k,l,n) with data points representing individual biological replicates (f,g,i,j,k,l,n) or as box-whisker plots with means indicated with dotted lines (m,o). For all experiments n=3 biological replicates; *pValue<0.05; **pValue<0.01; ***pValue<0.001;****pValue<0.0001; n.s.=not significant.