Modulating immune cell fate and inflammation through CRISPR-mediated DNA methylation editing

Abstract

Immune cell differentiation and activation are associated with widespread DNA methylation changes; however, the causal relationship between these changes and their impact in shaping cell fate decisions still needs to be fully elucidated. Here, we conducted a genome-wide analysis to investigate the relationship between DNA methylation and gene expression at gene regulatory regions in human immune cells. By using CRISPR-dCas9-TET1 and -DNMT3A epigenome editing tools, we successfully established a cause-and-effect relationship between the DNA methylation levels of the promoter of the interleukin-1 receptor antagonist (IL1RN) gene and its expression. We observed that modifying the DNA methylation status of the IL1RN promoter is sufficient to alter human myeloid cell fate and change the cellular response to inflammatory and pathogenic stimuli. Collectively, our findings demonstrate the potential of targeting specific DNA methylation events to directly modulate immune and inflammatory responses, providing a proof of principle for intervening in a broad range of inflammation-related diseases.

Fig. 1.. Integrative epigenome analyses uncover IL1RN promoter as a top DNAm-expression correlated event during myeloid cell commitment.

(A) Schematic overview of samples and methodology used in the epigenome profiling. iMac, induced macrophages. (B) Multidimensional Scaling analysis showing DNAm dynamics at 1-kb bins genome-wide during transdifferentiation. A gray arrow indicates the hypothetical trajectory. (C) Clustering of genome-wide DNAm dynamics (by WGBS) at 1-kb bins at GREs. Only 1-kb bins showing at least 10% DNAm changes are depicted. (D) Quantification of DNAm levels (by WGBS) and chromatin accessibility (by ATAC-seq) at the clusters in (C). Plots represent the mean (line) and interquartile range (shaded region). (E) Scatter plot comparing DNAm levels at iMac’s chromatin accessible regions between iMacs and B cells (only ΔDNAm>10% are shown). Blue dots highlight ATAC⁺ regions of interest. (F) Plot showing the correlation between the step changes (iMac versus B cells) in DNAm and gene expression at iMac’s chromatin accessible regions (ATAC⁺ peaks). Light red dots, promoter regions losing at least 10% of DNAm in iMacs. The large yellow dot highlights the correlation between DNAm and expression at the IL1RN promoter. (G) Genome browser snapshot showing signal for chromatin accessibility (by ATAC-seq) in iMacs, and gene expression (by RNA-seq) during transdifferentiation at the IL1RN locus. The blue-shaded region represents the DNAm dynamic promoter of the IL1RN short isoform (ENST00000409930.4). h, hours.

Fig. 2.. IL1RN promoter shows strong methylation–expression correlation during myeloid cell commitment in human primary cells.

(A) Schematic overview of samples and methodology used in the epigenome profiling of human primary blood cells collected from the Blueprint Consortium. (B) DNAm levels (by WGBS) in primary human B cells and primary human macrophages (Macs) at the clusters in (Fig. 1C). Each dot represents an individual sample. Unpaired two-tailed Student’s t test (**P < 0.01 n = 10 versus n = 6). (C) Plot showing the correlation between the step changes in DNAm at promoter ATAC⁺ regions and gene expression (primary Macs versus B cells). The colored dots are labeled according to their belonging to clusters in (Fig. 1C). The large yellow dot highlights the IL1RN promoter region in primary blood cells. ns, not significant. d, days.

Fig. 3.. CRISPR-mediated epigenome editing at the IL1RN promoter efficiently modulates DNAm levels and gene expression.

(A) Schematic of the dCas9-TET1 IL1RN promoter epigenome editing. The four sgRNAs targeting the IL1RN TSS are shown. #1 to #4 indicate the CpGs analyzed in (C). (B) ChIP-qPCR showing dCas9-TET1 enrichment at the IL1RN promoter. A 5-kb downstream region from the TSS is shown as a negative control region. Unpaired two-tailed Student’s t test, n = 3, mean SEM, (****P < 0.0001). (C) DNAm levels (pyrosequencing) at four CpGs upstream of IL1RN TSS [as in (A)] in CTRL and edited B cells. One-way analysis of variance (ANOVA) with Dunnett’s post hoc correction, n = 3, means ± SEM, (*P < 0.05; **P < 0.01, ***P<<0.001,****P < 0.0001). (D) Top, IL1RN expression (RT-qPCR) in CTRL and edited B cells. Unpaired two-tailed Student’s t test, n = 3, means ± SEM, (***P < 0.001). Bottom, IL1RN protein levels in CTRL and edited B cells. Fold change normalized to β-actin (CTRL versus sgIL1RN). (E) Schematic of the dCas9-DNMT3A IL1RN promoter epigenome editing experiment as in (A), showing two Infinium MethylationEPIC v2.0 probes. (F) Genome browser snapshot showing dCas9-DNMT3A binding at the IL1RN promoter. The location of the four sgRNAs targeting the IL1RN promoter and the two array probes is depicted. (G) Scatter plot showing differential dCas9-DNMT3A enrichment in sgIL1RN B cells. Large dots indicate top-bound promoters. (H) Scatter plot showing differentially hypermethylated CpGs in sgIL1RN day-3 cells. Blue dots indicate significantly hypermethylated CpGs (Δβ ≥ 0.3, FDR < 0.05, n = 13). Yellow dots highlight IL1RN probes shown in (E). (I) MA plot showing DEGs in sgIL1RN day-3 cells. (J) Upset plot depicting overlap of the dCas9-DNMT3A–binding sites, hypermethylated CpGs, and DEGs. (K) DNAm dynamics (array data) of two significantly hypermethylated IL1RN promoter CpGs. Unpaired two-tailed Student’s t test, n = 4, means ± SEM, (****P < 0.0001). (L) IL1RN dynamics (RNA-seq) in CTRL and edited cells. Unpaired two-tailed Student’s t test, n = 2, means ± SEM, (***P < 0.001).

Fig. 4.. IL1RN depletion alters myeloid cell fate and phagocytic function.

(A) Clustering of the DEGs during transdifferentiation between dCas9-DNMT3A CTRL and IL1RN edited cells. (n = 2 biologically independent replicates, FDR < 0.05). TC1 to TC4, transdifferentiation clusters. Representative genes for each cluster are annotated. (B) Gene Ontology Biological Processes (GO:BP) enrichment analysis for the genes associated with clusters in (A). The top five most significant terms for each cluster are plotted. Significant terms (q value < 0.05) are highlighted with a black stroke. (C) Mean fluorescence intensity (MFI) of macrophage-lineage surface markers in CTRL and edited iMacs. (D) Representative histograms showing signal intensity for selected cell surface markers in CTRL and edited iMacs. (E) Phagocytic capacity evaluation [by fluorescence-activated cell sorting (FACS)] for CTRL and edited iMacs. Left, histogram showing the uptake of blue fluorescent beads. Right, blue beads MFI quantification. Unpaired two-tailed Student’s t test, n = 3, means ± SEM, (***P < 0.001). (F) Up, schematic of the IL1RN shRNAs experiments during transdifferentiation. Bottom, MFI quantification of myeloid markers in shLuc and shIL1RN iMacs. Two-way ANOVA with Dunnett’s post hoc test, n = 3, means ± SEM, (****P < 0.0001). (G) Phagocytic capacity evaluation (by FACS) for shLuc and shIL1RN iMacs. Left, histogram showing the uptake of blue fluorescent beads. Right, blue beads MFI quantification. Two-way ANOVA with Dunnett’s post hoc test, n = 3, means ± SEM, (****P < 0.0001). (H) Left, schematic of the colony-forming assay performed in shLuc and shIL1RN bone marrow–derived human CD34⁺ cells; CFU, colony forming unit. Right, total number of CFUs obtained from shLuc and shIL1RN_1/_2 CD34⁺ cells. n = 3 biologically independent samples per group, means ± SEM. (I) Comparison of different types of CFUs obtained from shLuc and shIL1RN_1/_2 CD34⁺ cells. Two-way ANOVA with Dunnett’s post hoc test, n = 3, means ± SEM,(*P < 0.05; **P < 0.01; ****P < 0.0001).

Fig. 5.. IL1RN methylation editing disrupts NF-κB/IFN activation by IL-1β and dampens macrophage responses to immune challenges.

(A) Schematic of the IL-1β treatment experiment in CTRL and sgIL1RN edited iMacs. (B) MA plot showing DEGs in CTRL and sgIL1RN iMacs treated with IL-1β for 3 hours. (C) Lollipop plot depicting the DoRothEA TF activity predicted in CTRL and edited iMacs treated with IL-1β for 3 hours. (D) Quantification of p65 nuclear versus cytoplasmic localization signal in CTRL and edited iMacs treated with IL-1β for 3 hours. Unpaired two-tailed Student’s t test, n = 20 cells per group, means ± SEM, (****P < 0.0001). (E) GO:BP enrichment analysis of the most significantly down-regulated genes (log₂FC > 0.5, FDR < 0.05) at 3 hours of IL-1β treatment. Significant terms are shown in orange. (F) GSEA of IFN signature in edited iMacs treated with IL-1β for 3 hours. NES, normalized enrichment score. (G) Expression profiling of the 3-hour Up and 3-hour Down gene signatures in (B) across the MoMac-Verse scRNA-seq dataset (40). Top, average gene signatures scaled expression across the identities. Bottom, Uniform Manifold Approximation and Projection (UMAP) plots showing the average expression of each gene signatures. Cell identities with the highest signature expression are bolded and shown in the UMAPs. (H) Top, schematic representation of the cytokine release assay in CTRL and sgIL1RN iMacs, in response to various inflammatory and pathogenic stimuli. Bottom, heatmap of the cytokines concentrations in response to different stimuli in the iMacs. White rectangles represent cytokine concentrations below the detection threshold of the assay. Unpaired two-tailed Student’s t test with Benjamini-Hochberg, n = 6 biologically independent samples per group, means ± SEM, (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001).