Epigenetic dysregulation from chromosomal transit in micronuclei

Abstract

Chromosomal instability (CIN) and epigenetic alterations are characteristics of advanced and metastatic cancers^1-4, but whether they are mechanistically linked is unknown. Here we show that missegregation of mitotic chromosomes, their sequestration in micronuclei^5,6 and subsequent rupture of the micronuclear envelope⁷ profoundly disrupt normal histone post-translational modifications (PTMs), a phenomenon conserved across humans and mice, as well as in cancer and non-transformed cells. Some of the changes in histone PTMs occur because of the rupture of the micronuclear envelope, whereas others are inherited from mitotic abnormalities before the micronucleus is formed. Using orthogonal approaches, we demonstrate that micronuclei exhibit extensive differences in chromatin accessibility, with a strong positional bias between promoters and distal or intergenic regions, in line with observed redistributions of histone PTMs. Inducing CIN causes widespread epigenetic dysregulation, and chromosomes that transit in micronuclei experience heritable abnormalities in their accessibility long after they have been reincorporated into the primary nucleus. Thus, as well as altering genomic copy number, CIN promotes epigenetic reprogramming and heterogeneity in cancer.

Fig. 1. Distinct histone PTMs in micronuclei.

a, Representative immunofluorescence images from three biological replicates of micronucleated MDA-MB-231 cells stained for DNA (blue) and histone post-translational modifications (PTMs; red). White outlined boxes show a magnified view of micronuclei. b, Percentage of primary nuclei (blue) and micronuclei with particular histone PTMs (red) from an immunofluorescence experiment using human MDA-MB-231 (top) and mouse 4T1 (bottom) cells. **P < 0.01, ***P < 0.001, ****P < 0.0001, two-sided t-test, n = 3 biological replicates; data represent mean ± s.d.; n.s., not significant. c, Percentage of micronuclei with particular histone PTMs from an immunofluorescence experiment in MDA-MB-231 cells treated with DMSO vehicle control or vorinostat (HDAC inhibitor, HDACi). Statistical details are the same as in b. d, Heat map of z-scores of the relative abundance of histone PTMs from histone mass spectrometry in isolated micronuclei and primary nuclei of mouse 4T1 cells; n = 3. e, Percentage of micronuclei with H3K27me3 and H3K27ac from immunofluorescence staining in human high-grade serous ovarian cancer (HGSOC) tumour samples; n = 16, statistical significance tested using two-sided paired t-test; p < 0.0001. f, Representative immunofluorescence images of 16 human HGSOC tumour samples stained for DNA (blue) and either H3K27me3 or H3K27ac (red). In a and f, arrows, micronuclei; scale bars, 2 µm. Source data

Fig. 2. Micronuclear rupture and chromosome missegregation alter histone post-translational modifications in micronuclei.

a, Representative images of MDA-MB-231 cells stained for DNA (blue), histone PTMs (red) and cGAS (green). White outlined boxes show a magnified view of micronuclei, either intact or ruptured (Rupt.). b, Percentage of intact (green bars) and ruptured (yellow) micronuclei with histone PTMs in MDA-MB-231 cells; *P < 0.05, **P < 0.01, ***P < 0.001, two-sided t-test, n = 3; bars show mean ± s.d. c, Percentage of intact and ruptured micronuclei with H3K27me3 (left, P = 0.0008) and H3K27ac (right, P < 0.0001) in human high-grade serous ovarian cancer (HGSOC) samples; n = 16, two-sided paired t-test. d, Representative images from 16 human HGSOC samples stained for DNA (blue), cGAS (green), H3K27me3 (red) and H3K27ac (cyan). e, Percentage of intact and ruptured micronuclei with histone PTMs in DMSO-treated (red bars) or vorinostat-treated (grey) MDA-MB-231 cells; **P < 0.01, ***P < 0.001, two-sided t-test, n = 3; data represent mean ± s.d. f, Percentage of micronuclei with histone PTMs in control (red bars) and lamin-B2-expressing (LMNB2; grey) MDA-MB-231 cells; *P < 0.05, **P < 0.01, ****P < 0.0001, two-sided t-test, n = 3; data represent mean ± s.d. g, Representative images of MDA-MB-231 cells with micronuclei stained for DNA (blue), H3K27ac (green) and nascent RNA (red). Violin plots show the z-score of ethynyluridine (EU) intensity; ****P < 0.0001, two-sided Mann–Whitney U-test, n = 3. Solid and dotted bars represent the median and quartiles, respectively. h, Representative fluorescence in situ hybridization (FISH) images of CEN-SELECT DLD-1 cells treated with DMSO (control, left) or Dox/IAA (right) stained for Y chromosome (red), DNA (blue) and H3K27ac (green). Violin plots show the H3K27ac immunofluorescence intensity of DLD-1 cells treated with either DMSO (control) or Dox/IAA; ****P < 0.0001, two-sided Kolmogorov-Smirnov test, n = 3. Scale bars, 2 µm (d), 5 µm (a and h) and 10 µm (g). Source data

Fig. 3. Chromatin accessibility is altered in micronuclei.

a, Representative fluorescence-lifetime imaging microscopy images from 3 biological replicates of cGAS–GFP-expressing 4T1 cells that express cGAS and green fluorescent protein stained with Hoechst (blue). Images show intact (left) and ruptured (right) cells. Fluorescence lifetime is shown in pseudocolour; scale bar, 2 µm. Transm., transmission light microscopy. b, Violin plots denote fluorescence lifetime in euchromatic (Euch.) and heterochromatic (Hetero.) regions of primary nuclei (blue), and intact and ruptured micronuclei of cGAS–GFP-expressing Trex1-knockout 4T1 cells.**P < 0.01, ****P < 0.0001, n = 3; two-sided Mann–Whitney U-test; solid and dotted lines represent the median and quartiles, respectively. c, Representative ATAC-see (red) fluorescence images from 3 biological replicates of an MDA-MB-231 cell with a micronucleus stained for DNA (blue); scale bar, 2 µm. d, Violin plots representing ATAC-see signal intensity quantification of primary nuclei (PN) and micronuclei (MN) of MDA-MB-231 cells treated with either DMSO, an EZH2 inhibitor (EZH2i) or HDACi; ***P < 0.001, ****P < 0.0001, two-sided Mann–Whitney U-test, n = 5; solid and dotted bars represent the median and quartiles, respectively. e, Schematic of the isolation of micronuclei showing intact and ruptured micronuclei and primary nuclei. f, Heat maps showing differentially accessible genomic peaks from ATAC-seq (left) and WGS (right) from primary nuclei, ruptured micronuclei and intact micronuclei isolated from 4T1 cells; n = 2. g, Pie charts representing positional information of differentially accessible peaks in micronuclei compared with primary nuclei from 4T1 cells; n = 2. h, Density plot comparing change in H3K4me3 CUT&RUN read counts and ATAC-seq read counts in the same genomic region in micronuclei and primary nuclei from 4T1 cells; comparison performed with two-sided Spearman’s rank correlation statistics; n = 2, r = 0.36, P = 2.2 × 10⁻¹⁶. i, Heat map (middle) of H3K4me3 CUT&RUN peaks performed on isolated intact and ruptured micronuclei and primary nuclei of 4T1 cells (n = 2) revealing 4 clusters. Pie charts denote the differential accessibility of the reads from individual clusters on the heat map in micronuclei and primary nuclei (left), as well as the positional composition of the H3K4me3 of these reads (right). j, Enrichment plots of genes with promoters that are more accessible in intact micronuclei (top) or ruptured micronuclei (bottom) compared with primary nuclei in 4T1 cells in comparison to human breast tumours belonging to the top (FGA^high) or bottom (FGA^low) quartile of fraction genome altered, according to the TCGA. NES, normalized enrichment score; FDRq, false discovery rate q-value. Source data

Fig. 4. CIN drives long-term changes in chromatin accessibility.

a, Heat map representing chromatin accessibility from ATAC-seq in control (C) or lamin-B2-overexpressing (L) TP53-knockout RPE-1 cells treated with reversine or DMSO; n = 3. b, Tornado plots from CUT&RUN analysis representing H3K27ac- and H3K4me3-bound chromatin in regions corresponding to the accessibility heat map shown in a; n = 3.Scales on the right side signify peak intensity. c, Log₂-transformed fold change in the number of H3K4me3 (top), H3K27me3 (middle) and H3K27ac (bottom) CUT&RUN peaks on long-term reversine-treated (Rev) versus DMSO-treated TP53-knockout RPE-1 cells and those expressing lamin B2; graphs show pairwise comparisons. d, Experimental schematic representation depicting CEN-SELECT DLD-1 cell system used to generated Y-chromosome missegregation. Chromosome (Chr.) 6 is an example of a control autosome. e, Violin plots representing intraclonal variance from ATAC-seq, H3K4me3 CUT&RUN, H3K27me3 CUT&RUN and H3K27ac CUT&RUN across 10-kb segments in each of 14 DLD-1 clones. Aut., autosomes; Y, Y chromosome, Chr. 1 and Chr. 16. *****P < 0.00001, two-sided Mann–Whitney U-test, n = 14 clones of CEN-SELECT DLD-1 cells. f, Genome viewer plot showing copy-number-normalized ATAC-seq counts in a 2-kb region of chromosome 6 (control autosome) and the Y chromosome of DLD-1 CEN-SELECT parental control cells and the individual clones. Top row represents parental cells that did not undergo missegregation; the remaining rows represent single cell clones that underwent missegregation and transient micronucleation of the Y chromosome. Red arrows denote differentially accessible peaks. g, Density plot showing the comparison of the log₂-transformed fold change in H3K4me3 CUT&RUN reads (top; r = 0.17, P = 2.6 × 10⁻¹¹) or H3K27ac CUT&RUN read counts (bottom; r = 0.35, P= 2.2 × 10⁻¹⁶) against ATAC-seq read counts in a given region between DLD-1 CEN-SELECT clones or parental cells; two-sided Spearman’s rank correlation statistics. Source data

Fig. 5. Chromosomal transit in micronuclei promotes heritable epigenetic abnormalities.

a, Experimental schematic depicting missegregation of Chr. 4 in RPE-1 TP53-knockout cells induced using a chromosome 4 telomeric CRISPR guide; sgRNA, single guide RNA. b, Representative chromosome paint images from 2 biological replicates in RPE-1 TP53-knockout single-cell clones from a chromosome-4 missegregation system showing chromosome 4 (green) and DNA (blue). Scale bar, 5 µm. c, Percentage of micronuclei in single-cell-derived clones derived after transfection of TP53-knockout RPE-1 cells with either chromosome-4-targeting sgRNA or non-targeting control sgRNA; *P < 0.05, **P < 0.01, two-sided Mann–Whitney U-test; bars represent medians; n = 5 fields of view under 63× magnification. d, Percentage of micronuclei that contain chromosome 4 in single-cell-derived clones derived after transfection of TP53-knockout RPE-1 cells with chromosome 4 targeting sgRNA or non-targeting control or the mixed cell population shortly after transfection with the chromosome 4 targeting sgRNA; **P < 0.01, two-sided Mann–Whitney U-test; bars represent median; n = 10 fields of view under 63× magnification. e, Box plot (10-Megabase bin) showing copy-number normalized change of ATAC-seq counts on chromosome 4 between clone 1 and the control clone treated with the non-targeting guide sgRNA. The line represents the median and error bars represent minimum and maximum values; n = 3; solid red line represents the median; the bounds of the box are the interquartile range (Q1 to Q3); error bars are defined by 1.5*interquartile range beyond Q1 and Q3. f, Genome viewer plot showing copy-number-normalized ATAC-seq counts in the region of chromosome 4 that is copy-number altered in clone 1 (left) or clone 2 (right). Red arrows denote differentially accessible peaks compared with the control clone. g, Model showing how continuous chromosomal missegregation followed by micronucleation can introduce heritable epigenetic dysregulation. Because chromosomally unstable cells tend to continuously undergo chromosomal missegregation, this could in turn propagate epigenetic abnormalities through the same mechanism, leading to epigenetic heterogeneity. Source data

Extended Data Fig. 1. Micronuclei have distinct histone PTM profiles compared with primary nuclei.

a, Representative immunofluorescence images from 3 biological replicates of micronucleated MDA-MB-231 cells stained for DNA (blue) and histone posttranslational modifications (PTMs) (red), arrows point to micronuclei, scale bars 2 µm. Insets show magnified view of micronuclei (white outlined box). b, Percentage of primary nuclei and micronuclei with given histone PTMs from immunofluorescence experiment in OVCAR3, MCF10A, and RPE-1 cells, ** p < 0.01, *** p < 0.001, **** p < 0.0001, two-sided t-test, n = 3 biological replicates, bars represent mean ± SD. c, Violin plots showing the distribution of micronuclei percentage to primary nuclei in various cell lines used in the immunofluorescence quantification experiments. Values were obtained from 3 biological replicates. In each biological replicate, 10 fields of view under 63x magnification were observed. Solid and dashed lines in the plot represent the median and quartiles, respectively. d, Violin plots showing the normalized immunofluorescence intensity distribution of histone PTMs in: 55 micronuclei (MN) and 104 accompanying primary nuclei (PN) for H3K9Me3, 52 MN and 213 accompanying PN for H3K27Me3, 56 MN and 162 accompanying PN for H3K36Me3, 52 MN and 213 accompanying PN for H3K4Me3, 13MN and accompanying PN for H3K36Me2. Experiments were done on MDA-MB-231 cells; ns not significant, *** p < 0.001, **** p < 0.0001, two-sided Mann-Whitney test. Solid and dashed lines in the plot represent the median and quartiles, respectively. Source data

Extended Data Fig. 2. Epigenetic modifying drugs and epigenetic enzymes alter histone PTMs in micronuclei.

a, Representative immunofluorescence images from 3 biological replicates of micronucleated MDA-MB-231 cells stained for DNA (blue) and histone PTMs (red) treated with vehicle control (DMSO) or GSK126 EZH2 inhibitor (EZH2i), arrows point to micronuclei, scale bars 2 µm. Insets show magnified view of micronuclei (white outlined box). b, Violin plot showing normalized immunofluorescence intensity distribution of H3K27Me3 from immunofluorescence experiment in MDA-MB-231 cells. Cells were either treated with vehicle control (DMSO) or GSK126 EZH2 inhibitor (EZH2i); n = 52 micronuclei (MN) and 222 accompanying primary nuclei (PN) for DMSO-treated cells; n = 58 MN and n = 196 accompanying PN for EZH2i-treated cells; ** p < 0.01, **** p < 0.0001, two-sided Mann-Whitney test. Solid and dashed lines in the plot represent the median and quartiles, respectively. c–e, Representative immunofluorescence images from 3 biological replicates of micronucleated MDA-MB-231 cells stained for DNA (blue) and histone PTMs (red) treated with vehicle control (DMSO) or vorinostat HDAC Inhibitor (HDACi). Arrows point to micronuclei, scale bars 2 µm. Insets show magnified view of micronuclei (white outlined box). f–h, Representative immunofluorescence image from 3 biological replicates of micronucleated MDA-MB-231 cells stained for DNA (blue) and various epigenetic enzymes/complex members; f: KDM1 (green), g: PHC2 (green), h: RNF40 (red). Arrows point to micronuclei. Scale bars 2 µm. i, Representative immunofluorescence image from 3 biological replicates of micronucleated MDA-MB-231 cells stained for DNA in blue and phospho-RPB1 carboxy-terminal domain (serine 5), a functionally active RNA polymerase II (RNAPII) subunit, in green. Arrows point to micronuclei. Scale bars 2 µm. j, Violin plots showing normalized immunofluorescence intensity distribution of phospho-RPB1 carboxy-terminal domain (serine 5) from immunofluorescence experiment in MDA-MB-231, n = 25 MN and PN, **** p < 0.0001, two-sided Mann-Whitney test. Solid and dashed lines in the plot represent the median and quartiles, respectively. Source data

Extended Data Fig. 3. Mass spectrometry analysis of histone PTMs and epigenetic enzymes profiling in primary nuclei and micronuclei.

a, Micronuclei purification experiment schematic using cell lysis followed by sucrose gradient ultracentrifugation and FACS (figure created on Biorender.com). b, Abundance of various histone H3 PTMs relative to total histone H3 levels obtained from bottom-up histone mass spectrometry performed on isolated micronuclei (MN) and primary nuclei (PN) of 4T1 cells, n = 3 biological replicates, statistical significance tested using two-sided paired t-test, no adjustments made for multiple comparisons. c, (Left) Dot blots of H2AK119Ub and H2BK120Ub as well as total histones H2A and H2B as control from isolated micronuclei (MN) and primary nuclei (PN) of MDA-MB-231 cells. (Right) Relative quantification of H2AK119Ub and H2BK120Ub level normalized to total histones H2A and H2B, respectively, * p < 0.05, two-sided t-test, n = 4 biological replicates (H2AK119Ub), n = 3 biological replicates (H2BK120Ub), bars represent mean ± SD. d, Representative immunoblotting result from 3 biological replicates of lamin A in MDA-MB-231 cells treated with scrambled siRNA control (−) and siRNA for lamin A (+) for 24, 48, and 72 h with β-actin as loading control. e, Representative image from 3 biological replicates of MDA-MB-231 cells treated with shRNA for lamin A, stained for DNA (blue), H3K27ac (red), and cGAS (green). Arrows point to cGAS-positive area which indicates lamin rupture. Scale bar 5 µm. f, Percentage of intact (green) and ruptured (rupt.; gold) primary nuclei (PN) or micronuclei (MN) that are positive for H3K27ac immunofluorescence staining from MDA-MB-231 cells, ** p < 0.01, two-sided t-test, n = 3 biological replicates, bars represent mean ± SD. Source data

Extended Data Fig. 4. Impact of micronuclear rupture on histone PTMs.

a, Sample-level correlation between percentages of micronuclei with H3K27ac and cGAS staining in human high grade serous ovarian cancer (HGSOC) tumor samples (n = 16), 100 micronuclei counted/sample. statistical significance tested using two-sided t-test, no adjustments made for multiple comparisons. b, Violin plot showing normalized immunofluorescence intensity of histone PTMs in MDA-MB-231 cells’ PN, intact MN or ruptured (rupt.) MN; * p < 0.05, two-sided Mann-Whitney test). For H3K9Me3: n = 209 PN, 45 intact MN, and 51 ruptured MN; for H3K27Me3: n = 388 PN, 36 intact MN, and 53 ruptured MN; for H3K36Me3: n = 213 PN, 54 intact MN, and 39 ruptured MN. Solid and dashed lines in the plot represent the median and quartiles, respectively. c, Representative immunofluorescence images from 3 biological replicates of micronucleated MDA-MB-231 cells stained for DNA (blue), histone PTMs (red), and cGAS (green), scale bars 10 µm. Insets show magnified view of micronuclei (white outlined box). d, Representative immunoblotting results from 3 biological replicates of lamin B2 in control and mCherry-lamin B2 overexpressing MDA-MB-231 cells with β-actin as loading control. e, Representative immunofluorescence image of control and mCherry-lamin B2 overexpressing MDA-MB-231 cells stained with DNA (blue), scale bar 10 µm. f, Percentage of micronuclei with cGAS positive immunofluorescence staining in control and mCherry-lamin B2 overexpressing MDA-MB-231 cells (LMNB2); *** p < 0.001, two-sided t-test, n = 3 biological replicates, bars represent mean ± SD. g, Violin plot showing normalized immunofluorescence intensity of histone PTMs in primary nuclei (PN) and micronuclei (MN) of control and mCherry-lamin B2 (lamin B2) overexpressing MDA-MB-231 cells. For H3K9Me3: n = 99 PN (WT), 53 MN (WT), 73 PN (lamin B2), and 55 MN (lamin B2); for H3K27Me3: n = 73 PN (WT), 50 MN (WT), 82 PN (lamin B2), and 56 MN (amin B2); for H3K36Me3: n = 50 PN (WT), 50 MN (WT), 68 PN (lamin B2), and 50 MN (lamin B2); **** p < 0.0001, two-sided Mann-Whitney test, solid and dashed lines in the plot represent the median and quartiles, respectively. Source data

Extended Data Fig. 5. Mitotic errors are associated with abnormalities in histone PTMs.

a, Representative immunofluorescence image of MDA-MB-231 cells undergoing anaphase with lagging chromosome (pointed by arrows) stained for DNA (blue) and histone PTMs (red). Scale bars 2 µm. b, Violin plots showing the normalized immunofluorescence intensity distribution of histone PTMs on normal chromosomes (NC) and lagging chromosomes (LC) of MDA-MB-231 cells during mitosis. For H3K9Me3: n = 40 NCs and LCs, for H3K27Me3: n = 42 NCs and 28 LCs, for H3K36Me3: n = 40 NCs and 31 LCs; ** p < 0.01, *** p < 0.001, **** p < 0.0001, two-sided Mann-Whitney test, solid and dashed bars in the plot represent the median and quartiles, respectively. c, Representative immunofluorescence images of MDA-MB-231 cells at different stages of the cell cycle, stained for DNA (blue) and H2BK120Ub (red), arrows point towards micronuclei, scale bars 2 µm. d, Percentage of H2BK120Ub-positive micronuclei immunofluorescence staining on vehicle (DMSO)-treated or CENP-E inhibitor (CENP-Ei, GSK923295)-treated MDA-MB-231 cells; ** p < 0.01, two-sided t-test, n = 3 biological triplicates, bars represent mean ± SD. e, Violin plot showing the fluorescence intensity quantification from EU staining experiment in MN and PN of MDA-MB-231 cells; n = 38 MN and PN, **** p < 0.0001, two-sided Mann-Whitney test. Solid and dashed bars in the plot represent the median and quartiles, respectively. Source data

Extended Data Fig. 6. Micronuclei have distinct accessibility as compared with primary nuclei.

a, Violin plots representing the normalized distribution of fluorescence lifetime in euchromatic (Euch.) and heterochromatic (Hetero.) regions of primary nuclei (PN), ruptured and intact micronuclei (MN) from 4T1 cells, *** p < 0.001, **** p < 0.0001, n = 73 cells (31 intact and 42 ruptured micronuclei) from five biological replicates, two-sided Mann-Whitney test, solid and dashed bars in the plot represent the median and quartiles, respectively. b, Violin plots representing the normalized distribution of fluorescence lifetime in euchromatic (Euch.) and heterochromatic (Hetero.) regions of primary nuclei (PN), ruptured and intact micronuclei (MN) from MDA-MB-231 cells, *** p < 0.001, **** p < 0.0001, n = 3 biological replicates, two-sided Mann-Whitney test, solid and dashed bars in the plot represent the median and quartiles, respectively. c, Representative fluorescence images from 3 biological replicates showing ATAC-see signal in MDA-MB-231 cells treated with (top) or without (bottom) fluorophore-tagged adaptors-loaded Tn5 transposase, scale bar 5 μm. d, Experimental schematic for the isolation of primary nuclei, intact micronuclei, and ruptured micronuclei that will be separated by fluorescence activated cell sorting represented in e. e, Density area plots of a representative micronuclei isolation flow cytometry experiment from 4T1 cells expressing H2B-mCherry and GFP-cGAS, right panel is the subset of the gate applied on the left panel (dashed box). f, Representative immunofluorescence image from 3 independent experiments showing isolates of primary nuclei and micronuclei from 4T1 cells after sucrose gradient ultracentrifugation stained for DNA (blue) and cGAS (green), scale bar 20 µm g, Relative copy number calling from MoCaSeq pipeline ordered by chromosome (x-axis) from isolated primary nuclei, intact micronuclei, and ruptured micronuclei of 4T1 cells. h, Principal component analysis (PCA) plot of intact micronuclei (MN), ruptured (rupt.) MN, and primary nuclei (PN) of 4T1 cells based on peak reads from ATAC-seq, n = 2 biological replicates. i, Scatter plot showing the comparison of the log 2 fold change values of genomic copy number vs. log2 fold change of ATAC-seq counts in a given region between intact MN (top) or ruptured MN (bottom) vs. PN of 4T1 cells. Two-sided Pearson’s rank correlation statistics are shown on the panel. j, Box plot showing genomic copy number distributions obtained from WGS of intact MN, ruptured MN, and PN of 4T1 cells categorized by differentially accessible regions from ATAC-seq peaks obtained from previous ATAC-seq experiment on MN and PN of 4T1 cells (blue = more accessible in PN, red = more accessible in MN). Line represents median value, n = 2 biological replicates for ATAC-Seq, single replicate for WGS, black line represents median value, the bounds of the box are the interquartile range (Q1 to Q3), the whiskers are defined by 1.5*interquartile range beyond Q1 and Q3, the minimum was 0.1 and the maximum was 3.011. Source data

Extended Data Fig. 7. Changes in chromatin accessibility in micronuclei exhibit a positional bias.

a, Enriched pathways from regions that are more accessible in micronuclei (obtained from ATAC-seq experiment) relative to primary nuclei, sorted by gene ratio. b, Enriched pathways from regions that are less accessible in micronuclei (obtained from ATAC-seq experiment) compared with primary nuclei, sorted by gene ratio. For a and b, pathway enrichment was determined using an over-representation statistical test (a one-sided version of Fisher’s exact test) and p-values were adjusted for multiple hypothesis testing via the Benjamini-Hochberg correction. c, Heat map representing chromatin accessibility from ATAC-seq peaks at transcription start sites (TSS peaks) and all other sites (non-TSS peaks) in primary nuclei (PN), intact (Intact MN) and ruptured micronuclei (Rupt. MN) from 4T1 cells, n = 2 biological replicates. d, Pie charts showing annotations of genomic regions from ATAC-seq experiment that are more accessible in MN (left charts) and less accessible in MN (right charts), stratified by the baseline expression of the genes within that region obtained from RNA-sequencing experiment of 4T1 dnMCAK CIN^high cells (top = high expression, middle = medium expression, bottom = low expression). e-f, Sequencing tracks in ZDHHC21 (d), AKTS1, and TBC1D17 (e) promoter loci in intact MN, ruptured MN, and PN of 4T1 cells, showing ATAC-seq peaks (top panel) and H3K4me3 CUT&RUN peaks (bottom panel). g, Experimental strategy schematic depicting the generation of 4T1 MCAK (CIN^low) and dnMCAK (CIN^high) cells, which are then subjected to ATAC-seq to compare their chromatin accessibility to the chromatin accessibility results obtained from ATAC-seq experiments on isolated micronuclei (MN) and primary nuclei (PN) of 4T1 cells. h, CIN profiling on 4T1 MCAK (CIN^low) and dnMCAK (CIN^high) cells. Left panel shows chromosomal missegregation percentage out of 50 cells per biological replicate in anaphase, **** p < 0.0001, two-sided t-test, bars represent mean ± SD, n = 3 biological replicates, 150 total number of anaphase cells observed. Right panel shows micronuclei percentage over primary nuclei in a field of view in 63x magnification, **** p < 0.0001, two-sided Mann-Whitney test, bars represent median, n = 3 biological replicates, 30 fields of view. i, Violin plots representing log2 fold change of ATAC-seq peaks from 3 biological replicates comparing 4T1 CIN^high to 4T1 CIN^low cells. The peaks are categorized by their relative accessibility in micronuclei vs. primary nuclei purified from 4T1 cells. The categories are: 1) peaks that are more accessible in MN vs. PN (MN>PN), 2) peaks that are more accessible in PN vs. MN (MN<PN), 3) and peaks that are equally accessible in MN vs. PN (MN = PN), n = 3 biological replicates, **** p < 0.0001, two-sided Mann-Whitney test, solid and dashed bars in the plot represent the median and quartiles, respectively. Source data

Extended Data Fig. 8. The generation of isogenic hTERT RPE-1 cell lines with different CIN rates.

a, Experimental schematic depicting the generation of isogenic hTERT p53 KO RPE-1 cells with different CIN rates. b, IC50 curves of reversine-treated control and Lamin B2 overexpressing RPE-1 cells, points represent mean ± SD, n = 3 biological replicates. c, Percentage of cells undergoing anaphase with chromosome mis-segregation in control or lamin B2 overexpressing RPE-1 cells long-term-treated vehicle (DMSO) or reversine (Rev.); n = 300 cells per condition, 6 biological replicates, *** p < 0.001, two-sided t-test, bars represent mean ± SD. d, Percentage of micronuclei relative to primary nuclei in control or lamin B2 overexpressing RPE-1 cells long-term-treated with vehicle (DMSO) or reversine (Rev.); n = 30 fields of view in 63x magnification, 3 biological replicates, **** p < 0.0001, two-sided t-test, bars represent median. e, Conventional karyotypes of control or lamin B2 overexpressing RPE-1 cells long-term-treated with vehicle (DMSO) or reversine, red arrows denote chromosomal abnormalities. f, Principal component analysis (PCA) plot of TP53 KO hTERT-RPE-1 cells based on accessibility from ATAC-seq, n = 3 biological triplicates. g, Principal component analysis (PCA) plot of TP53 KO hTERT-RPE-1 cells based on RNA expression obtained from RNA-seq, n = 3 biological triplicates. h, Normalized copy number (CN) calling results from GATK4 pipeline in RPE-1 cells undergoing long-term DMSO treatment (top) or long-term reversine treatment (bottom). Red line indicates CN loss, green line indicates CN gain, grey line indicates no CN change. i, Pie charts showing genomic region annotations of accessible peaks in RPE-1 p53 KO cells undergoing long-term DMSO treatment vs. long-term reversine treatment (left panels) and reversine vs. reversine + lamin B2 overexpression (right panels). Source data

Extended Data Fig. 9. Chromosome missegregation upregulates cancer pathways and heterogeneity in genomic accessibility.

a, Enrichment plots of genes that are more accessible in MN vs. PN of 4T1 cells against genes that are more accessible in in either long-term reversine-treated or DMSO-treated p53 KO RPE-1 cells. b, Enrichment plots of genes that are more accessible in ruptured MN vs. PN of 4T1 cells against genes that are more accessible in either long-term reversine-treated p53 KO RPE-1 cells or reversine-treated p53 KO RPE-1 cells overexpressing lamin B2. c, Principal component analysis (PCA) plot of TP53 KO hTERT-RPE-1 cells based on a list of tumor suppressor genes and oncogenes, n = 3 biological triplicates. d, KEGG pathway analysis from KEGG Pathways in cancer – Homo sapiens (human) database (map05200). Long-term reversine-treated vs. DMSO-treated p53 KO RPE-1 cells analysis is shown on the left, while long-term reversine-treated lamin B2-overexpressing p53 KO RPE-1 cells vs. long-term reversine-treated p53 KO RPE-1 cells is shown on the right. e, Violin plots representing intraclonal variance across 10kb segments in each of 14 DLD-1 CEN-SELECT clones. Aut. = autosomes, Y = Y chromosome, 1 = chromosome 1, 16 = chromosome 16. Distribution represents 86,666 calculated values for autosomes, 730 calculated values for Y chromosome. ***** p < 0.0001, two-sided Mann-Whitney test. f, Inter-clonal comparison of fold change of ATAC-seq peaks in autosomes, and the Y chromosomes in 14 DLD-1 single clones isolated from CEN-SELECT system to parental control. Aut. = autosomes, Y = Y chromosome, 1 = chromosome 1, 16 = chromosome 16. Bars represent median, statistical significance tested using two-sided Mann-Whitney test. g, Scatter plot comparing ATAC-seq fold change variance to number of breakends per basepair in the Y-chromosome of the clones of DLD-1 CEN-SELECT system cells. h, Scatter plot comparing ATAC-seq fold change variance in regions with breakends to ATAC-seq fold change variance across all regions in the Y-chromosome of the clones of DLD-1 CEN-SELECT system cells. For g and h, statistical test was done using two-sided Pearson’s rank correlation. No adjustments were made for multiple comparisons. i, Density plot showing the comparison of the log2 fold change of H3K27me3 CUT&RUN reads vs. ATAC-seq reads in a given region between DLD-1 CEN-SELECT clones vs. parental, correlation measured with two-sided Spearman’s rank correlation statistic. Source data

Extended Data Fig. 10. Chromosome missegregation and micronucleation promotes epigenetic dysregulation.

a, (Left) Histogram of log2 fold change of ATAC-seq counts between MMCT HCT116 cells’ clones compared to parental for chromosome of interest (COI), chromosome 1 (Chr 1; as a single chromosome control), and all other chromosomes combined (Other). KS = Kolmogorov-Smirnov test’s absolute maximum distance value between COI vs. other chromosomes. (Right) The variance of the log 2 fold change values of ATAC-seq counts of clone vs. parental control in a particular chromosome. Chromosome of interest is colored in red while chromosome 1 is colored in blue, n = 3 biological replicates. b, Analysis of clone 1 of chr 4 missegregation system in RPE-1 cells. Altered region is represented by the chromosome diagram on the top (red = altered, blue = normal). The bar plot represents the variance of log2 fold change of ATAC-seq counts (single cell-derived clones derived after transfection of p53 KO RPE-1 cells with chromosome 4 targeting sgRNA or non-targeting control) in a particular chromosome. Chromosome 1 is shown in blue (Chr 1), the whole chromosome 4 (Ch4) is shown in red, altered region of chromosome 4 is shown in gold (Ch4 ALT), and unaltered chromosome 4 region is labeled as Ch4 NORM. Left inset shows the histogram of log2 fold change of ATAC-seq counts between clone 1 compared to parental for whole chromosome 4 (red), chromosome 1 (blue) or other chromosomes (gray). Right inset shows the histogram of log2 fold change of ATAC-seq counts between clone 1 compared to parental for altered region of chromosome 4 (gold) and the unaltered region of chromosome 4 (gray). KS = Kolmogorov-Smirnov test’s absolute maximum distance value between chromosome 4 values vs. other chromosomes (left inset) or between altered chromosome 4 values vs. unaltered chromosome 4 (right inset). c, Analysis of clone 2 of chr 4 missegregation system in RPE-1 cells. Altered region is represented by the chromosome diagram on the top (red = altered, blue = normal), note that chromosome 18 is also affected due to spontaneous duplication. The bar plot represents the variance of log2 fold change of ATAC-seq counts (single cell-derived clones derived after transfection of p53 KO RPE-1 cells with chromosome 4 targeting sgRNA or non-targeting control) in a particular chromosome. Chromosome 1 is shown in blue (Chr 1), the whole chromosome 4 (Ch 4) is shown in red, altered region of chromosome 4 is shown in gold (Ch4 ALT), unaltered chromosome 4 region is labeled as Ch4 NORM, and chromosome 18 is shown in brown. Left inset shows the histogram of log2 fold change of ATAC-seq counts between clone 2 compared to parental for whole chromosome 4 (red), chromosome 1 (blue) or other chromosomes (gray). Middle inset shows the histogram of log2 fold change of ATAC-seq counts between clone 2 compared to parental for altered region of chromosome 4 (gold) and the unaltered region of chromosome 4 (gray). Right inset shows the histogram of log2 fold change of ATAC-seq counts between clone 2 compared to parental for chromosome 18 (brown), chromosome 1 (blue) or other chromosomes (gray). KS = Kolmogorov-Smirnov test’s absolute maximum distance value between chromosome 4 values vs. other chromosomes (left inset), between altered chromosome 4 values vs. unaltered chromosome 4 (middle inset), or between chromosome 18 vs. other chromosomes (right inset). Source data