Diversity of Epigenetic Features of the Inactive X-Chromosome in NK Cells, Dendritic Cells, and Macrophages

Abstract

In females, the long non-coding RNA Xist drives X-chromosome Inactivation (XCI) to equalize X-linked gene dosage between sexes. Unlike other somatic cells, dynamic regulation of Xist RNA and heterochromatin marks on the inactive X (Xi) in female lymphocytes results in biallelic expression of some X-linked genes, including Tlr7, Cxcr3, and Cd40l, implicated in sex-biased autoimmune diseases. We now find that while Xist RNA is dispersed across the nucleus in NK cells and dendritic cells (DCs) and partially co-localizes with H3K27me3 in bone marrow-derived macrophages, it is virtually absent in plasmacytoid DCs (p-DCs). Moreover, H3K27me3 foci are present in only 10-20% of cells and we observed biallelic expression of Tlr7 in p-DCs from wildtype mice and NZB/W F1 mice. Unlike in humans, mouse p-DCs do not exhibit sex differences with interferon alpha production, and interferon signature gene expression in p-DCs is similar between males and females. Despite the absence of Xist RNA from the Xi, female p-DCs maintain dosage compensation of six immunity-related X-linked genes. Thus, immune cells use diverse mechanisms to maintain XCI which could contribute to sex-linked autoimmune diseases.

Keywords: NK cells; X-chromosome inactivation; Xist RNA; interferon alpha; long non-coding RNA; macrophages; plasmacytoid dendritic cells; sex differences.

Figure 1

Xist RNA transcripts are mostly absent from the Xi in female NK cells and DCs. (A) Schematic showing the origin for the immune cells examined here. Hematopoietic stem cells (HSCs); common lymphoid progenitors (CLPs); common myeloid progenitors (CMPs); plasmacytoid dendritic cells (p-DCs); myeloid-derived DCs (m-DCs); lymphoid-derived DCs (L-DCs). (B) Sorting strategy for isolation of m-DCs (MHC-II⁺, CD11c⁺, CD11b⁺, CD8a^lo), L-DCs (MHC-II⁺, CD11c⁺, CD11b⁺, CD8a^hi), and NK cells (NK1.1⁺) using FACS. Spleens from two female mice were pooled for each experiment (repeated twice), and flow results from experiment 1 are shown. (C) Xist RNA FISH analyses of NK cells, m-DCs, L-DCs, using Cy3 labeled oligo probes. (D) Quantification of Xist RNA localization patterns (Types I–IV) for each experiment. The total number of nuclei counted for each cell type is shown above the column. Statistical significance was determined comparing each type of Xist RNA pattern (Types I–IV) for each cell type, using a two-tailed t-test. The comparison between NK cells and m-DCs for Type II patterns was the only significant difference (p < 0.024). NK cells and L-DCs had no significant differences in Xist RNA localization patterns.

Figure 2

Xist RNA and H3K27me3 foci are localized to the Xi in most female BMDMs. (A) Xist RNA FISH for resting BMDMs and in vitro stimulated cells (using 1 μM CpG), collected 3 days after stimulation. (B) Quantification of Xist RNA localization patterns for BMDMs stimulated with either 1 μM CpG or 1 μg/mL LPS. The total number of nuclei counted for each cell type is shown above the column. Statistical significance for comparisons of resting (day 0) vs. stimulated cells was performed for each type of Xist RNA localization pattern (Types I–IV) using a two-tailed t-test, and the only significant difference was for CpG-stimulated Type I cells (p < 0.005). (C) Sequential Xist RNA and H3K27me3 IF for resting and stimulated BMDMs. White arrows indicate H3K27me3 foci. (D) Quantification of co-localization patterns for Xist RNA and H3K27me3 foci. Results from two independent experiments are shown. The total number of nuclei counted for each cell type is shown above the column. Statistical significance for comparisons of resting (day 0) vs. stimulated cells was performed for each type of localization pattern using a two-tailed t-test, and the only significant difference was for CpG-stimulated Type I cells (p < 0.005).

Figure 3

Female plasmacytoid DCs lack Xist RNA at the Xi and exhibit biallelic expression of Tlr7 in some cells. (A) Xist RNA FISH for resting p-DCs and in vitro activated p-DCs after days 1–3 of culture. P-DCs were isolated from spleen and lymph nodes from female mice, in two independent experiments and stimulated with CpG. For the second isolation, cells were stimulated for 3 days before collection for RNA FISH. (B) Quantification of Xist RNA localization patterns for p-DCs showing that all p-DCs are missing Xist RNA on the Xi. The total number of nuclei counted for each cell type is shown above the column. (C) Sequential Xist RNA FISH followed by immunofluorescence (IF) for H3K27me3 enrichment at the Xi. (D) Quantification of co-localization patterns for Xist RNA and H3K27me3 foci. Results from two independent experiments are shown. The total number of nuclei counted for each cell type is shown above the column. (E) Single-molecule RNA FISH for Tlr7 transcripts using wildtype p-DCs from healthy mice. Oligo probes specific for exonic Tlr7 were Cy3-labeled (red), and intronic Tlr7 probes were FITC-labeled (green). White arrows indicate pinpoint signals for nascent Tlr7 expression from the X-chromosome, with signals from both exonic and intronic probes. (F) single-molecule Tlr7 RNA FISH in p-DCs from NZB/W F1 mice with SLE-like disease. Disease development was assessed by proteinuria and DNA autoantibodies prior to p-DC isolation from spleen and lymph nodes. White arrows indicate pinpoint signals for nascent Tlr7 expression from the X-chromosome, with signals from both exonic and intronic probes. (G) Schematic for counting allele-specific expression, and quantification of monoallelic and biallelic Tlr7 expression in wildtype and NZB/W F1 p-DCs. The percentages of total numbers for biallelic and monoallelic expressing cells are shown in parentheses.

Figure 4

p-DCs do not exhibit a sex difference with IFNα production and X-linked genes are dosage compensated in the absence of Xist RNA localization at the Xi. (A) Relative quantity (2^∧ΔCt) of Xist RNA in unstimulated male and female pDCs, and cells stimulated with CPG, R848. Female B cells (naïve and CPG stimulated) were included as positive controls. (B) Relative quantity (2^∧ΔΔCt) of four IFNα signature genes. P-DCs were activated with R848 for 6 h or were unstimulated (NS). The housekeeping gene Rpl13a was used for normalization, and male unstimulated samples (NS) were normalized to 1. (C) Concentration of IFNα protein produced by cultured male and female p-DCs from NZB/W F1 mice measured by ELISA. (D) Relative quantity (2^∧ΔΔCt) of six X-linked immune genes from male and female splenic p-DCs from NZB/W F1 mice. P-DCs were activated with R848 for 6 h or were unstimulated (NS). The housekeeping gene Rpl13a was used for normalization, and unstimulated samples (NS) were normalized to 1.