B cell stimulation changes the structure and higher-order organization of the inactive X chromosome

Abstract

X chromosome inactivation (XCI) equalizes X-linked gene expression between sexes. B cells exhibit dynamic XCI, with Xist RNA/heterochromatic marks absent on the inactive X (Xi) in naive B cells but returning following mitogenic stimulation. The impact of dynamic XCI on Xi structure and maintenance was previously unknown. Here, we find dosage compensation of the Xi with state-specific XCI escape genes in naive and in vitro-activated B cells. Allele-specific OligoPaints indicate similar Xi and active X (Xa) territories in B cells that are less compact than in fibroblasts. Allele-specific Hi-C reveals a lack of TAD-like structures on the Xi of naive B cells and stimulation-induced alterations in TAD-like boundary strength independent of gene expression. Notably, Xist deletion in B cells changes TAD boundaries and large-scale Xi compaction. Altogether, our results uncover B cell-specific Xi plasticity, which could underlie sex-biased biological mechanisms.

Keywords: B cell stimulation; B cells; CP: Immunology; CP: Molecular biology; TAD remodeling; X chromosome inactivation; XCI escape genes; XCI maintenance; Xist RNA; allele-specific Hi-C; chromosome compartments; chromosome structure; inactive X chromosome; topological associated domains.

Figure 1.. The Xi in B cells is dosage compensated and has XCI escape genes in female B cells

(A) Schematic of the F1 mus × cast mouse with skewed XCI. Female M. musculus (Musculus) heterozygous for Xist deletion (XistΔ) is mated to a WT male M. castaneus (Castaneus) mouse. The F1 generation expresses Xist exclusively from the Castaneus allele; thus, the WT Xi is paternally inherited (Castaneus) in every cell of this mouse. Splenic CD23⁺ B cells from female F1 mus × cast mice are stimulated in vitro with CpG for 12 or 24 h and used for the experiments in this study. Cartoon representation of Xist RNA FISH patterns at each time point are previously reported. (B) Allele-specific RNA-seq analyses showing total reads per kilobase per million mapped (RPKM) of X-linked reads mapping to either the Xi or Xa genomes in B cells across 0, 12, and 24 h. Bars represent mean ± SD. Statistical testing was performed using one-way ANOVA with Tukey’s correction for multiple comparisons. **p < 0.005, ****p < 0.0001; ns, not significant. (C) X chromosome plots generated by chromoMap, showing the genomic location of Xi-specific expression in 0 h (blue), 12 h stimulated (green), and 24 h stimulated (coral) B cells for one replicate sample. Values are displayed as the sum of RPKM in 1.7 Mb bins. The Xist locus is indicated by dotted lines. Colored regions represent annotated and expressed regions of the X chromosome, and black bars represent silenced and un-annotated regions. (D) Venn diagram showing the distribution of the 44 XCI escape genes in female B cells across time points. (E) X chromosome maps showing the location of each XCI escape gene in B cells and female fibroblasts. Xist and Dxz4 regions are indicated with red lines. Genes are listed by location; colors indicate the role in immune processes (purple) and chromatin/transcription (green). Underlined and italicized genes indicate unique B cell XCI escape genes expressed at each time point.

Figure 2.. The Xi has attenuated A/B compartments in B cells and is less compact compared to primary fibroblasts

(A) Allele-specific Hi-C heatmaps at 200 kb resolution of each X chromosome from B cells at 0 and 24 h post stimulation. A green arrow denotes the Dxz4 boundary region that separates two mega-domains on Xi. Compartment tracks depicting the A (red) and B (blue) compartments are shown below each heatmap. Scale is ±5E–2. (B) Top: Hi-C heatmaps binned at 30 kb resolution, showing the Dxz4 boundary region (chrX: 70,970,797–74,970,797) on the Xi in B cells at 0 and 24 h. Bottom: quantification of the local insulation scores using sliding windows for the Dxz4 region (chrX: 72,960,000–73,020,000) on the Xi in B cells at 0 h (blue) and 24 h (coral). (C) Haplotype OligoPaints (HOPs) DNA FISH imaging for distinguishing the Xi and Xa in B cells and primary fibroblasts from female F1 mus × cast mice and diagram of the probe design for allele-specific resolution of each X chromosome. Representative DNA FISH images (single channel and merged channels) show dual-color probe labeling of the Xi (cyan) and the Xa (pink) in B cells at the 0 and 24 h time points. (D) Allele-specific surface area measurements of each X chromosome territory, calculated as the Xa/Xi ratio. Violin plots show median with quartiles; ****p < 0.0001. (E) Normalized allele-specific surface area measurements of each X chromosome territory, normalized by total nuclear size. Boxplots show median with quartiles and min-to-max whiskers; **p = 0.0089, ****p < 0.0001. (F) Allele-specific sphericity measurements of each X chromosome territory, calculated as the Xa/Xi ratio. Violin plots show median with quartiles; ****p < 0.0001. (G) Raw allele-specific measurements of sphericity for each chromosome territory. Boxplots show median with quartiles and min-to-max whiskers; ****p < 0.0001. Statistical significance for surface area and sphericity measurements was determined using Kruskal-Wallis tests. Hi-C experiments had 2 replicate female mice/time point. Imaging experiments had 3 replicate female mice for each cell type, and mice were matched for both time points (B cells).

Figure 3.. B cell stimulation increases TAD structures across the Xi

(A) Xi-specific Hi-C heatmaps at 30 kb resolution of a region (chrX: 65,674,654–72,500,000) in B cells across the 0, 12, and 24 h time points. Green arrows denote increased TAD interactions as stimulation progresses. Insulation scores (green line) are shown below each heatmap, along with locations of repressed genes on the Xi (C) and XCI escape genes (S). (B) Xi-specific Hi-C heatmaps at 30 kb resolution of a region (chrX: 158,373,375–166,000,000) in B cells across the 0, 12, and 24 h time points. Insulation scores (green line) are shown below each heatmap, along with locations of repressed genes on the Xi (C). (C) Xi-specific Hi-C heatmaps at 30 kb resolution of a 4 Mb region (chrX: 98,655,712–102,655,712) encompassing the Xist gene (black arrowhead) in B cells across the 0, 12, and 24 h time points. Insulation score is shown below each heatmap. (D) Insulation score (median score over time points) distribution for promoters of all B cell escape genes (red, n = 44) relative to genes subject to XCI silencing (black, n = 248) for the Xa and Xi. p value was determined using a one-sided Kolmogorov-Smirnov test and is indicated on the graphs. (E) Detection of TADs at the Nsdhl region on the Xi, using Hi-C and two-color DNA FISH imaging. Left: Hi-C heatmaps (30 kb) of the Nsdhl region on the Xi in B cells across the 0, 12, and 24 h time points. The TAD probe location (green and purple) spanning the boundary and insulation score are shown below each heatmap. Right: DNA FISH imaging of the TAD at the Nsdhl regions on the Xa and Xi (white arrowhead). Shown are representative images of single nuclei at each time point; magnified images show probe-labeled TADs for each X chromosome. (F) Quantification of bi-color probe overlap for the Nsdhl TAD region on the Xi. Measurements of center-to-center distances between each probe color and boxplot of percent overlap for the two colors. Percent overlap is normalized to the volume of the green probe. Violin plots show median with quartiles. Bars represent mean ± SD. (G) Imaging of the Xiap region, lacking TADs and spanning no boundary, on the Xi using Hi-C and two-color DNA FISH. Left: Hi-C heatmaps (30 kb) of the Xiap region on the Xi in B cells across the 0, 12, and 24 h time points. The two-color probe location (green and purple) and insulation score are shown below each heatmap. Right: DNA FISH imaging of the TAD at the Xiap region on the Xa and Xi (white arrowhead). Shown are representative images of single nuclei at each time point; magnified images show probe-labeled regions for each X chromosome. (H) Quantification of bi-color probe overlap for the Xiap region. Shown are measurements of center-to-center distances between each probe color and boxplot of percent overlap for the two colors. Percent overlap is normalized to the volume of the green probe. Violin plots show median with quartiles. Bars represent mean ± SD. For DNA FISH imaging experiments, n = 3 replicate female F1 mus × cast mice for the 0 and 24 h time points and n = 2 female mice for the 12 h time point. Scale bars: 5 μm. Statistics were performed using a Kruskal-Wallis test; ***p < 0.0005.

Figure 4.. Xist deletion in B cells increases inter-TAD interactions across the Xi

(A) Schematic of the Xist locus and loxP sites in Xist2lox mice for Xist deletion in mature CD23⁺ B cells using Mb1-Cre recombinase to generate Xist^cKO/cKO (Xist^cKO) mice after multiple mating steps. See STAR Methods for further details. (B) Representative Xist RNA FISH images of in vitro-activated B cells (24 h) from Xist2lox and Xist^cKO mice showing loss of Xist RNA signal and quantification of nuclei with robust Xist RNA clouds for Xist2lox, Xist^cKO/+, and Xist^cKO mice (n = 3 for each genotype). Significance was determined using unpaired t test; **p < 0.001. (C–F) Detection of TADs across the X chromosomes, using allele-specific Hi-C and two-color DNA FISH imaging, for B cells at the 0 and 24 h time points. DNA FISH imaging of TADs using 2-color probes (green and magenta) measured either TAD-TAD boundaries or the TAD-3′ boundary region, centered (±2 Mb) on the (C) Stk26 region, (D) Nsdhl region, (E) Ppef1 region, and (F) Xiap region (lacking TADs or Xi boundary, control). Images show representative nuclei for Xist 2lox (F1 mus 3 cast) and Xist cKO B cells at each time point, and white arrowheads indicate the Xi in Xist 2lox samples (Xi cannot be distinguished from Xa in Xist cKO samples). Scale bars for TAD imaging: 5 μm. Quantification of 2-color probe overlap (center-to-center distances and probe color overlap) for each region is shown below. Percentage of probe color overlap is normalized to the volume of the green probe within each pair. Violin plots show median with quartiles. Bars represent mean ± SD. Statistical significance was determined using a Kruskal-Wallis test; *p < 0.05, **p < 0.005, ***p = 0.0006, ****p < 0.0001. n = 3 replicate female mice for each genotype; n = 2 replicate female F1 mice for the Stk26 region at 24 h.