Overarching control of autophagy and DNA damage response by CHD6 revealed by modeling a rare human pathology

Abstract

Members of the chromodomain-helicase-DNA binding (CHD) protein family are chromatin remodelers implicated in human pathologies, with CHD6 being one of its least studied members. We discovered a de novo CHD6 missense mutation in a patient clinically presenting the rare Hallermann-Streiff syndrome (HSS). We used genome editing to generate isogenic iPSC lines and model HSS in relevant cell types. By combining genomics with functional in vivo and in vitro assays, we show that CHD6 binds a cohort of autophagy and stress response genes across cell types. The HSS mutation affects CHD6 protein folding and impairs its ability to recruit co-remodelers in response to DNA damage or autophagy stimulation. This leads to accumulation of DNA damage burden and senescence-like phenotypes. We therefore uncovered a molecular mechanism explaining HSS onset via chromatin control of autophagic flux and genotoxic stress surveillance.

Fig. 1. Generation of isogenic iPSCs and developmental impact of the I1600M CHD6 mutation.

a Schematic representation of the CHD6 protein highlighting key functional protein domains (top) and the HSS-specific I1600M missense mutation in the second SANT/SLIDE domain (bottom). b Reprogramming skin fibroblasts from an HSS patient carrying the I1600M CHD6 mutation into iPSCs exemplified by immunofluorescence (left) and pluripotency marker expression levels (mean log₂ fold-change; right) in two independent clones. Bar, 5 μM. c Generation of iPSCs monoallelically (left) or heterozygously expressing I1600M CHD6 (right) using two CRISPR/Cas9 editing strategies. d Derivation of wild-type, monoallelic mutant or patient-derived cardiomyocytes (CMs; top) and neural crest cells (NCCs; bottom) from iPSCs and detection of lineage-specific marker genes by immunofluorescence (right). Bar, 5 μM. e RT-qPCR mean mRNA level changes of CM markers (n = 2 independent replicates, except het) from all genotypes in (d). f Top KEGG pathways misregulated in both patient and mutant NCCs RNA-seq data. g As in (f), but using RNA-seq data from patient and mutant CMs. h Representative immunofluorescence images of CHD6 distribution in wild-type, patient-derived, and monoallelic mutant CMs. Nuclei were counterstained by DAPI. Bar, 5 μM. i Overexpression of wild-type (top row) and mutant CHD6 (bottom row) in chicken embryos and examination of lateral and ventral views at stage HH20 to assess proper development of branchial arches (blue), eyes (red), and jaw (white). The number of embryos analyzed (n) is also given. j As in (i), but for ventral views of the heart (left) and corresponding H-E tissue stainings (right). C conus, RV right ventricle, LV left ventricle, A atrium. Bar, 500 μM.

Fig. 2. CHD6 binds autophagy-related genes across different cell types.

a Genome browser views of CHD6 ChIP-seq data from wild-type (wt), monoallelic mutant (mut), heterozygous mutant (het) or patient-derived (pat) NCCs (light blue), CMs (magenta), iPSCs (dark blue) around the GAA promoter. hESC ENCODE ChIP-seq data for CHD1/7 and histone marks are aligned below. b Line plot (top) and heatmaps (bottom) showing ChIP-seq signal distribution in the 6 kbp around CHD6-bound sites from wild-type (gray), patient-derived (dark blue), or monoallelic mutant NCCs (light blue). Input data provide a baseline (dashed). c Venn diagrams showing overlap of CHD6 ChIP-seq peaks from the three indicated NCC genotypes (top) or across cell types (bottom). *: significantly more than expected by chance; P < 10⁻⁴, two-sided hypergeometric test. d Bar plot showing the percent of mutCHD6 ChIP-seq peaks located at increasing distances upstream or downstream of gene TSSs from all three cell types. e Significantly enriched GO terms associated with mutCHD6-bound genes in NCCs. f Heatmaps showing TF motif enrichment (over background; blue shades) and associated P-values (red shades) within accessible DNase I footprints overlapping NCC mutCHD6 ChIP-seq peaks. g Line plot showing TFEB and mutCHD6 ChIP-seq signal overlap at iPSC peaks. h Significantly enriched GO terms associated with the genes at TFEB/CHD6-shared peaks from (g). i Heatmap showing changes in mRNA levels (log₂) of mutCHD6-bound and differentially-regulated genes upon 2-h starvation of iPSCs. Those up- or downregulated (magenta/blue rectangles) in both monoallelic- (mut) and heterozygous-mutant iPSCs (het) are highlighted. j Significantly enriched GO terms associated with the genes highlighted in (i).

Fig. 3. CHD6 affects the DNA damage response through modulation of autophagy flux.

a Western blots (top) showing changes in ATG3 and LC3 levels in wild-type (wt) and monoallelic mutant iPSCs (mut) that were serum-starved (starv) and/or treated with chloroquine (chlo) or rapamycin (rapa); β-tubulin levels provide a loading control. Mean normalized band intensities from two experiments (±SD) were quantified and plotted relative to wt levels (bottom). b FACS profiles (left) of wild-type (wt) or monoallelic mutant iPSCs (mut) transfected with the GFP-LC3-RFP reporter. Plots (right) quantify the percent of RFP-only iPSCs over all RFP/GFP-double positive plus RFP-only cells in control and 2-h starvation conditions (mean ±SD). *: mean significantly different from control; P < 0.01, unpaired two-tailed Welsch t-test. c Representative images of wild-type (wt) or monoallelic mutant iPSCs (mut) immunostained for LC3 (green) and p62 (magenta) in the presence or absence of rapamycin from one experiment are shown. Box plots (mean with whiskers indicating 95th percentile intervals) quantify the number of puncta, mean signal and extent of LC3/p62 colocalization. *: mean significantly different from wt; P < 0.01, unpaired two-tailed Welsch t-test. Bar, 5 μM. d As in (c), but for γH2A.X levels (thick black lines indicate IQR, thin lines indicate 95% confidence intervals) after 30-min etoposide (ETO) treatment of wild-type (wt), heterozygous (het), or monoallelic mutant (mut) CMs. *: significantly different from wt; P < 0.01, Wilcoxon-Mann-Whitney test. Bar, 10 μM. e As in (d), but for γH2A.X (green) and CHD6 (magenta) after etoposide treatment of iPSCs. The line scan and colocalization index (ci) of the two signal profiles (bottom; γH2A.X foci indicated by arrows) exemplify the lack of signal overlap. Bar, 5 μM. f Comet assays and quantification of tails in wild-type (wt), heterozygous mutant (het), patient-derived (pat) or monoallelic mutant (mut) iPSCs treated with etoposide for 30 min and allowed 24 h to recover; numbers of cells analyzed (n) are indicated. *: significantly different from wt; P < 0.05, Wilcoxon-Mann-Whitney test. g As in (f), but for wild-type iPSCs treated with etoposide and allowed to 24 h recover in the presence or absence of autophagy inhibitors. *: significantly different from wt; P < 0.05, Wilcoxon-Mann-Whitney test. h Bar plots showing percentage (mean ±SD, n = 3 independent experiments) of wild-type (wt), heterozygous (het) or monoallelic mutant (mut) iPSCs that survived apoptosis, also in the presence or absence of autophagy inhibitors. *: mean significantly different from wt; P < 0.05, two-tailed unpaired Student’s t-test. i Bar plots showing the percentage (±SEM, n = 3 independent experiments) of wild-type (wt), heterozygous (het) or monoallelic mutant iPSCs (mut) in the G1, S, or G2 cell cycle phase upon 0, 10 or 30 min of etoposide treatment followed by 24 h recovery. *: significantly different from wt; P < 0.05, two-tailed unpaired Student’s t-test. j Significantly-enriched GO terms associated with the genes highlighted in (k). k Heatmap showing changes in mRNA levels (log₂) of genes differentially regulated in wild-type iPSCs upon 1 h of etoposide treatment. Those convergently up-/downregulated in monoallelic (mut) and heterozygous mutant (het) cells are highlighted (magenta/blue rectangles).

Fig. 4. The I1600M CHD6 mutation hinders co-factor recruitment at autophagy gene promoters.

a In silico rendering of the second putative SANT-SLIDE domain structure for wild-type and mutant CHD6 (based on the published Chd1 SANT-SLIDE structure) and zoomed views around the I1600M mutation (bottom). b NanoDSF melting profiles (first derivative of the 350/330 nm ratio; n = 3 independent runs) of FLAG-tag purified wild-type (dark blue) or I1600M (light blue) full-length CHD6 along a temperature gradient. *: significantly different mean; P < 0.01, two-tailed unpaired Student’s t-test. Scattering profiles along the gradient reveal no difference in aggregation (inset). c Overexpression of wild-type or mutant CHD6 cloned into a piggybac vector (left) in iPSCs exemplified by Venus-GFP levels or anti-HA western blotting of CHD6 (right); β-tubulin provides a control. Bar, 25 μM. d Network representation of GO terms associated with proteins co-purifying with wild-type CHD6 in steady-state iPSCs. e Heatmaps showing enrichment or depletion of GO terms/pathways linked to CHD6-interacting proteins upon starvation, etoposide treatment or in steady-state wild-type and mutant iPSCs. f Western blot of SMARCC1 co-immunoprecipitating with overexpressed wild-type (wt) or mutant CHD6 (mut) in iPSCs; anti-HA and anti-β-tubulin blots provide a control. g Representative genome browser views for SMARCA4/-B1/-C1 ChIP-seq aligned to CHD6 ChIP-seq from wild-type (wt) and monoallelic mutant iPSCs (mut) around the DNAAF1 and LAMP1 promoters. h Line plots showing average SMARCA4/-B1/-C1 ChIP-seq signal profiles in the 4 kbp around CHD6 peaks; signal from mutCHD6 ChIP-seq provides a reference. i Venn diagram (left) showing overlap of genes bound by mutCHD6 and SMARCA4/-B1/-C1. Bar plots (right) showing significantly-enriched GO terms associated with the 243 shared genes. *: significantly more than expected by chance; P < 10⁻³, hypergeometric test. j Bar plots showing changes in SMARCC1 or ACTL6A ChIP-qPCR signal (fold enrichment over wt ±SD; n = 2 independent experiments) from heterozygous (white) or monoallelic mutant iPSCs (blue). k Heatmaps (left) showing changes in mRNA levels (log₂) in response to iPSC starvation at genes co-bound by CHD6 and SMARCA4/-B1/-C1. Pearson’s correlation coefficients for each pair are shown. Bar plots (right) showing significantly-enriched GO terms associated with these genes.

Fig. 5. A model for HSS-relevant effects of the CHD6 mutation.

In individuals with wild-type CHD6 (top), co-recruitment of CHD6 and BAF/PBAF complexes to the promoters of autophagy genes ensure proper gene regulation in response to pro-autophagy and stress cues. In those carrying the I1600M mutation, the recruitment of BAF/PBAF co-factors is hindered and autophagy control is deregulated; as a result, the DNA damage response is compromised, senescent cell features emerge, and cell differentiation and specification processes can be impacted.