FOSL1-associated condensates modulate superenhancer-mediated selected IL-1 family gene activation in psoriasis

Abstract

Psoriasis is a chronic inflammatory skin disease involving abnormal keratinocyte gene expression, yet the upstream epigenetic mechanisms driving these alterations remain unclear. Here we performed multiomics analyses (chromatin immunoprecipitation sequencing, CUT&Tag, ATAC-seq, and RNA sequencing) on psoriatic lesions, perilesional skin, healthy skin, and M5-stimulated HaCaT cells. Functional validation was performed using CRISPR-mediated superenhancer (SE) editing, epigenetic modulation, circular chromosome conformation capture sequencing, assay for chromatin-bound condensates by exploratory sequencing, coimmunoprecipitation, and fluorescence recovery after photobleaching. We found widespread activation of enhancers and SEs in lesions, including an interleukin 1 (IL-1)–SE regulating IL1A, IL1B, and IL36G transcription. BET inhibition reduced SE activity, repressed IL-1 cytokines, and relieved psoriasis-like inflammation in vivo. Fos-like antigen 1 (FOSL1) was identified as a key transcription factor enriched in active SEs and associated with condensate formation, contributing to the maintenance of EP300 occupancy and histone H3 lysine 27 acetylation deposition. Macrophage-derived IL-1 up-regulated FOSL1 through mitogen-activated protein kinase/ERK signaling, forming a positive feedback loop that amplified inflammation. This study revealed an FOSL1-SE–centered regulatory network driving keratinocyte inflammatory responses, providing previously unknown therapeutic targets for psoriasis.

INTRODUCTION

Psoriasis is a chronic inflammatory skin condition characterized by abnormalities in both epidermal function and immune responses (1). Transcriptomic studies have identified numerous differentially expressed genes (DEGs) in both peripheral immune cells and lesional skin tissue of patients with psoriasis (2). However, the upstream molecular mechanisms that regulate these DEGs remain unclear. Although genome-wide association studies (GWAS) have identified more than 80 genetic loci linked to psoriasis susceptibility, most are in noncoding regions within the human genome (3, 4). The transcriptional regulatory functions of these noncoding genomic regions in psoriasis need to be investigated.

Epigenetic regulation plays a critical role in linking genomic information and environmental factors to shape cellular phenotypes by DNA methylation, histone modifications, and noncoding RNAs (5, 6). Enhancers are noncoding regulatory DNA elements enriched with histone modifications (7). Active enhancers are characterized by high levels of histone H3 lysine 27 acetylation (H3K27ac) and increased chromatin accessibility, which facilitate the recruitment of transcription factors (TFs), coactivators, and RNA polymerase II to activate gene transcription (8). Enhancers activate distal gene transcription via chromatin looping and are often clustered into superenhancers (SEs), which exhibit exceptionally high occupancy of transcriptional coactivators and drive robust expression of genes critical for cell identity and disease (9, 10). SEs exert essential regulatory functions in normal cellular development and differentiation. Lineage-specific SEs control mesenchymal stem cell commitment to osteoblasts or adipocytes by activating core identity genes (e.g., RUNX2 and peroxisome proliferator–activated receptor γ) via assembling TF-centered core regulatory circuitry. GWAS-associated single-nucleotide polymorphisms for bone or adipose disorders are also enriched in these lineage-selective SEs, linking enhancer dysregulation to developmental and pathological phenotypes (11). Dysregulation of SE-mediated gene expression is additionally implicated in malignant transformation. In nasopharyngeal carcinoma, SOX2 was identified as a key gene regulated by SEs, which promotes tumorigenesis via chromatin reorganization and correlates with poor patient survival. A genetic variant in the SOX2 SE correlates with higher SOX2 expression and shorter survival. CRISPR interference (CRISPRi) targeting of SOX2 SEs or BET inhibitor treatment represses SOX2 and impairs tumor growth (12).

The function of enhancers is mediated by sequence-specific TFs that bind conserved motifs, such as activator protein-1 (AP-1) and nuclear factor κB (NF-κB)–binding sites (13). Recent studies reveal that TFs and epigenetic modifiers could undergo liquid-liquid phase separation (LLPS), forming condensates. This leads to regions with a high density of transcriptional elements, which substantial enhances the transcriptional activation of target genes (14, 15). The key TFs governing regulatory elements and the potential role of phase separation in the pathogenesis of psoriasis remain uncertain. Fos-like antigen 1 (FOSL1; also known as FRA-1), a member of the FOS family proteins, dimerizes with JUN family proteins to form the TF complex AP-1. AP-1 binds to different promoters and enhancers to orchestrate multiple cellular and physiological processes in normal tissues (16). Previous studies have demonstrated that FOSL1 activates TRAF3, which in turn triggers the NF-κB signaling pathway, thereby promoting keratinocyte hyperproliferation in psoriasis (17). However, whether the activity of psoriatic enhancers is regulated by FOSL1 remained to be investigated.

In this study, to investigate the role and mechanism of enhancers and SEs in the pathogenesis of psoriasis, we used an integrative multiomics approach to define tissue-specific enhancer landscapes in psoriatic lesions. We identified active enhancers in keratinocytes and uncovered FOSL1 as a key pathogenic TF. Furthermore, we identified and functionally characterized an SE-regulating selected interleukin-1 (IL-1) family genes transcription (IL1-SE) and demonstrated that FOSL1 regulates IL1-SE activity through phase separation. Moreover, we found that IL-1 signaling activated FOSL1 via the ERK pathway, establishing a positive feedback loop that amplified inflammatory gene transcription. These findings reveal an enhancer-centric regulatory mechanism and suggest previously unknown therapeutic targets for psoriasis.

RESULTS

Increased enhancer and SE in psoriatic lesions and proinflammatory cytokines treated keratinocytes

To investigate the characteristics of enhancers in psoriatic lesions, we analyzed public H3K27ac chromatin immunoprecipitation sequencing (ChIP-seq) data (GSE161076) from lesional (PP, n = 4), nonlesional (PN, n = 4), and healthy skin (NN, n = 4) (Fig. 1A). Principal components analysis (PCA) of H3K27ac enrichment across all samples revealed a clear segregation of PP samples from both PN and NN groups (Fig. 1B). The overall pattern of H3K27ac enrichment in the PP group showed a higher signal at the transcription start site (TSS) and a discernible signal at the transcription end site (TES), suggesting increased transcriptional activity in psoriatic lesions (Fig. 1C). Notably, genomic annotation showed that more peaks in intronic and intergenic regions were observed in the PP group compared with the PN and NN groups, which suggested increased activity of enhancers in psoriatic lesions (Fig. 1D).

Fig. 1. Elevated activity of enhancers and SEs in psoriatic lesions.

(A) Schematic overview of samples in the GSE161076. (B) PCA analysis of global H3K27ac signals across psoriatic lesions (PP, n = 4 patients), nonlesional skin (PN, n = 4 patients), and healthy skin (NN, n = 4 patients). (C) Enrichment patterns of H3K27ac around TSS and TES regions. The x axis indicates the position relative to the TSS or TES in kilobases (Kb), where −3.0 represents 3 kb upstream of the TSS and +3.0 represents 3 kb downstream of the TES. The y axis shows the normalized read density, reflecting the signal intensity of H3K27ac. (D) Distribution of H3K27ac binding sites across genomic regions in the NN, PN, and PP groups, including promoters, untranslated regions (UTRs), exons, introns, downstream regions, and distal intergenic regions. (E) Spearman correlations between H3K27ac signal change (Log₂ fold change, PP versus NN) and gene expression change (RNA Log₂ fold change, PP versus NN). Each point represents an enhancer-gene pair. (F) KEGG pathway enrichment analysis for genes associated with the 418 psoriasis-associated enhancers of lesions. The y axis lists enriched KEGG pathways (with associated genes shown below), and the x axis shows enrichment strength as −log₁₀(P value)t. (G) Chromatin accessibility of the 418 psoriasis-associated enhancers across PP, PN, and NN samples. (H) Number of SEs identified in PP, PN, and NN tissues. The x axis shows ranked typical enhancers based on H3K27ac signal intensity, and the y axis represents H3K27ac enrichment. (I) Venn diagram showing the overlap of SEs among PP, PN, and NN groups. PP exhibits the largest number of group-specific SEs, while some SEs are shared across groups. (J) KEGG pathway enrichment analysis of the 502 PP SEs. The y axis lists enriched KEGG pathways, and the x axis shows –log₁₀ (P value), where larger values indicate stronger enrichment.

To identify psoriasis-associated enhancers, we integrated RNA sequencing (RNA-seq) (GSE54456) and ATAC-seq (PRJNA597655) data, focusing on intergenic and intronic H3K27ac peaks. Gene expression changes showed a positive correlation with differential H3K27ac signals. A total of 418 enhancers were identified whose activity changes were consistent with the expression of their linked genes (Fig. 1E and data S1). Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis revealed enrichment of target genes in immune-related pathways, including T cell differentiation, Janus kinase–signal transducers and activators of transcription (JAK-STAT) signaling, cytokine-cytokine receptor interactions, and the IL-17 pathway (Fig. 1F). ATAC-seq confirmed increased chromatin accessibility at these enhancers in lesional skin (Fig. 1G). In addition, the increased number of SEs was identified in lesional skin (Fig. 1H). SE-associated genes were enriched in psoriasis-related pathways such as T helper 17 (T_H17) differentiation, tumor necrosis factor (TNF) and vascular endothelial growth factor (VEGF) signaling, and apoptosis (Fig. 1, I and J), underscoring the potential role of SEs in coordinating pathologic gene expression networks.

To model the inflammatory keratinocyte state in vitro, HaCaT cells were stimulated with a mixture of five proinflammatory cytokines (M5: IL-1α, IL-17A, OSM, TNF-α, and IL-22) for 24 hours and performed integrated RNA-seq, H3K27ac CUT&Tag, and ATAC-seq analyses (fig. S1A). PCA of H3K27ac profiles revealed a clear separation between unstimulated (N) and cytokine-stimulated (M) cells (fig. S1B). Genome-wide H3K27ac signal was markedly elevated upon M5 stimulation (fig. S1C). Similar to psoriatic lesions, H3K27ac peaks were predominantly localized to intronic and intergenic regions in both groups (fig. S1D). Moreover, we identified 333 enhancers whose H3K27ac enrichment positively correlated with changes in the expression of their predicted target genes (fig. S1E and data S2). KEGG pathway analysis of the associated target genes showed significant enrichment for psoriasis-relevant inflammatory pathways, including IL-17 signaling, TNF signaling, NF-κB signaling, and cytokine-cytokine receptor interaction (fig. S1F). ATAC-seq confirmed that these enhancers exhibited increased chromatin accessibility upon M5 stimulation (fig. S1G). In addition, M5 stimulation also increased the number of SEs and the associated 394 genes enriched in TNF, mammalian target of rapamycin (mTOR), VEGF, chemokine, and calcium signaling pathways (fig. S1, H to J).

JQ1 suppresses a psoriasis-relevant transcriptional program by targeting SEs

Given the critical role of SEs in driving pathologic gene expression and their established sensitivity to BET bromodomain inhibitors, we investigated the therapeutic potential of SE disruption in psoriasis (18). We used JQ1, a small-molecule inhibitor that competitively binds to the acetyl-lysine recognition pockets of BET proteins (e.g., BRD4), thereby displacing them from chromatin and disrupting SE-mediated transcription (19–21). First, we assessed the effect of JQ1 on keratinocyte function in vitro. CCK-8 assays confirmed that JQ1 at concentrations up to 100 nM was not cytotoxic to HaCaT cells (fig. S2A). Treatment with 50 and 100 nM JQ1 for 24 hours markedly inhibited keratinocyte proliferation and while promoting apoptosis compared with the dimethyl sulfoxide control (Fig. 2, A and B). Next, M5-stimulated HaCaT cells were treated with 100 nM JQ1, followed by RNA-seq profiling. RNA-seq analysis indicated that JQ1 led to the inhibition of expression for 1312 genes, which significantly exceeded the up-regulated genes (Fig. 3A, fig. S2B, and data S3). Gene Set Enrichment Analysis (GSEA) disclosed that the expression of target genes for cell cycle and cytokine-cytokine receptor interaction was significantly reduced in the JQ1-treated HaCaT cells compared with control cells (Fig. 3, B and C). These genes have been found to be associated with the dysregulation of keratinocyte proliferation and differentiation. Genes down-regulated by JQ1 are primarily enriched in pathways related to the cell cycle, cytokine interactions, TNF, IL-17A, and p53 signaling pathway, which overlap with the pathways enriched in up-regulated DEGs in psoriasis (fig. S2C).

Fig. 2. JQ1 treatment ameliorated psoriatic pathology in vitro and in vivo.

(A) Effect of JQ1 on the proliferation of M5-stimulated HaCaT Cells. Scale bar, 100 μm [n = 3 independent biological replicates, one-way analysis of variance (ANOVA) with Tukey multiple comparisons]. (B) Effect of JQ1 on apoptosis of M5-stimulated HaCaT cells (n = 3 independent biological replicates, one-way ANOVA with Tukey multiple comparisons). (C) Topical JQ1 gel improves skin appearance in IMQ-induced psoriasis-like mice. Representative skin lesions and hematoxylin and eosin staining of mice in the control, IMQ, IMQ + mock, and IMQ + JQ1 groups. Scale bar, 100 μm. (D) PASI scores from day 0 to day 5 for control, IMQ, IMQ + mock, and IMQ + JQ1 groups; statistical comparisons are shown between IMQ + JQ1 and IMQ + mock groups. (n = 6 independent biological replicates, two-tailed t test). (E) Individual components of the PASI score on day 5 for IMQ + mock and IMQ + JQ1 groups (n = 6 independent biological replicates, two-tailed t test). (F) Quantification of acanthosis in control, IMQ, IMQ + mock, and IMQ + JQ1 groups (n = 6 independent biological replicates, one-way ANOVA with Tukey multiple comparisons). (G) Immunohistochemical staining of the keratinocyte proliferation marker Ki-67 in mice and quantification of Ki-67–positive cells (n = 6 independent biological replicates, one-way ANOVA with Tukey multiple comparisons). Scale bar, 50 μm. All error bars represent SD, ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05.

Fig. 3. Validation of the SE regulating selected IL-1 family gene transcription (IL1-SE).

(A) Heatmap of gene expression in M5-stimulated HaCaT cells treated with JQ1 or dimethyl sulfoxide (DMSO). The color gradient represents changes in gene expression levels. (B) GSEA analysis of the effects of JQ1 on the cell cycle (hsa04110) and cytokine-cytokine receptor interaction (hsa04060) pathways. The left panels show the enrichment score (ES) curves from GSEA. NES indicates the normalized enrichment score, and FDR represents the false discovery rate. (C) Relative mRNA expression levels of relevant genes based on RNA-seq data of the cell cycle (hsa04110) and cytokine-cytokine receptor interaction (hsa04060) pathways, with DMSO as the control group and JQ1 as the treatment group (n = 3 independent biological replicates, two-tailed t test). (D) Venn diagram showing the overlap among M5-induced up-regulated genes, M5-induced up-regulated H3K27ac peaks, and genes down-regulated by JQ1. (E) Enriched signaling pathways for the 82 overlapping genes. (F) Genes associated with SEs, including members of the S100, IL-1, and IL-10 families. Genes within the same genomic cluster are indicated by the same color, while black bars represent genes in different genomic regions. (G) Genomic features of the IL-1 family gene cluster and its annotated SE. (H) mRNA expression of IL-1 family genes (IL1A, IL1B, IL36G, IL37, IL36RN, IL1RN, and IL36B) in IL1-SE KO HaCaT cells before and after M5 stimulation (n = 3 independent biological replicates, two-way ANOVA with Sidak test for multiple comparisons). (I) Secretion levels of IL-1α, IL-1β, and IL-36γ in the supernatant of IL1-SE KO HaCaT cells measured by ELISA before and after M5 stimulation. (n = 3 independent biological replicates, two-way ANOVA with Sidak test for multiple comparisons) All error bars represent SD, ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05.

Pharmacological disruption of SEs ameliorates psoriatic pathology in vitro and in vivo

Consistent with our in vitro observations, H3K27ac CUT&Tag and RNA-seq analyses of epidermis tissue from IMQ-induced mice revealed an increased number of SEs under disease conditions (fig. S3, A and B). Therefore, we formulated a topical gel containing JQ1 at doses of 100 μg/cm² and applied it to the skin of IMQ-induced psoriasis-like mice daily for five consecutive days. JQ1 treatment markedly ameliorated disease severity, as evidenced by significantly lower psoriasis area and severity index (PASI) scores compared to the vehicle-treated (IMQ + Mock) group (Fig. 2, C and E). Histological analysis revealed that JQ1 reduced epidermal hyperplasia (acanthosis) (Fig. 2F) and decreased keratinocyte proliferation, as indicated by reduced Ki67 staining in the epidermis (Fig. 2G). Moreover, we also found that topical JQ1 application led to the coordinated expression down-regulation of key SE-associated gene clusters in vivo, including proinflammatory members of the Il36 and Mmp families (fig. S3, C to E). To further evaluate the pharmacological activity of JQ1 in another psoriasis-like mouse model, we established an IL-23–induced psoriasis-like dermatitis model. Similar to the findings observed in the IMQ model, topical administration of JQ1 markedly alleviated IL-23–induced skin inflammation, as evidenced by reduced ear thickness (fig. S4, A to C). JQ1 treatment significantly reduced epidermal hyperplasia (fig. S4D) and decreased the proportion of Ki67-positive areas (fig. S4E), indicating effective suppression of psoriasis-like skin inflammation. Together, these findings indicated that SEs play an important role in the pathogenesis of psoriasis via regulating proinflammatory and cell cycle genes expression in keratinocyte.

A keratinocyte-associated SE controls selected IL-1 family genes transcription

To reveal psoriasis-associated genes regulated by SEs, we integrated the identified enhancers and SEs with RNA-seq data in M5-stimulated keratinocytes with JQ1 treatment. We found that 82 up-regulated DEGs following M5 stimulation were down-regulated upon JQ1 treatment, with corresponding enhancer-regulated regions identified (Fig. 3, D and E), suggesting that these genes, mainly including S100 family, IL1 family, and IL10 family, may be regulated by enhancers or SEs (Fig. 3F).

We further analyzed the SEs located in these gene clusters using the SEdb2.0 database (http://licpathway.net:8081/sedb/), an IL1-SE was identified, overlapping with keratinocyte-associated SEs. Furthermore, increased H3K27ac enrichment and chromatin accessibility were observed at the IL1-SE locus in psoriatic lesional skin compared with healthy and nonlesional skin, supporting activation of IL1-SE in psoriasis in vivo (Fig. 3G). Although a putative S100A-associated enhancer region was observed following M5 stimulation, it did not overlap with annotated keratinocyte SEs (fig. S5A). IL10-SEs ranked very low in activity (fig. S5B). To determine whether these chromatin activation patterns are conserved in primary human epidermal keratinocytes (HEKs), we next analyzed HEKs under the same conditions. Genome browser visualization of H3K27ac signals revealed increased enrichment at TSSs and 333 psoriasis-associated enhancers following M5 stimulation in HEKs (fig. S6, A and B). H3K27ac profiling in HEKs also identified an IL1-SE upon M5 stimulation. This region partially overlapped with the corresponding SE in HaCaT cells (fig. S6C). These results support the physiological relevance of IL1-SE activation beyond immortalized keratinocytes.

Moreover, analysis of public Hi-C data from human keratinocytes (HiC Data Browser, https://3dgenome.fsm.northwestern.edu/) revealed that IL1-SE and IL-1 family genes IL1A, IL1B, IL36A, IL36B, IL36G, IL37, and IL36RN were all located within the same topologically associating domain (fig. S7A), indicating a coordinated control mechanism within this domain. Following M5 stimulation, the IL1-SE ranking improved from 1009th to 685th, suggesting an increased activity of IL1-SE in psoriatic keratinocytes (fig. S7B).

To reveal the function of IL1-SE, CRISPR-Cas9 technology was used to target and excise DNA fragments within the IL1-SE region (fig. S7C). Disruption of the IL1-SE led to reduced expression levels of IL1A, IL1B, and IL36G (Fig. 3H). Enzyme-linked immunosorbent assay (ELISA) of culture supernatants demonstrated significant reductions in IL-1α, IL-1β, and IL-36γ protein levels upon IL1-SE deletion (Fig. 3I). To examine whether IL1-SE physically contacts its target genes, we performed circular chromosome conformation capture sequencing (4C-seq) in HaCaT keratinocytes. The interaction frequencies between IL1-SE and the promoters of IL1A, IL1B, and IL36G were significantly higher in keratinocytes than in Jurkat T cells, confirming cell type–specific chromatin looping (fig. S7D). Together, these results indicate that the transcription of psoriasis-associated gene family such IL-1 family can be controlled coordinately by single SE in keratinocyte during the development of psoriasis.

FOSL1 is a key TF regulating psoriasis-associated enhancer activity

To identify key TFs driving enhancer activity in psoriasis, we performed motif enrichment analysis on differential H3K27ac peaks from both psoriatic lesions and M5-stimulated HaCaT cells. The results revealed that the TF FRA1 (encoded by FOSL1) was significantly enriched across multiple comparison groups, particularly in lesional versus healthy skin (PP versus NN) and lesional versus perilesional skin (PP versus PN) (Fig. 4A). Further analysis demonstrated that in M5-stimulated HaCaT cells, FOSL1 was enriched in 41.17% of regions gaining H3K27ac signal but showed no significant enrichment in regions losing H3K27ac (Fig. 4B). Corresponding ATAC-seq footprint analysis confirmed increased chromatin accessibility specifically at these FRA1/FOSL1 motif sites (Fig. 4, C and D). These analyses suggested that FOSL1 plays a key role in the positive regulation of psoriasis-associated gene expression.

Fig. 4. Identification of TFs binding in enhancer and up-regulation of FOSL1 expression in psoriatic lesions.

(A) Motif analysis of differential H3K27ac peaks identified in psoriatic lesions (PP) versus healthy skin (NN) and PP versus nonlesional skin (PN) comparisons. (B) Motif analysis of up-regulated (gain) and down-regulated (loss) H3K27ac peaks in M5-stimulated HaCaT cells. (C) Footprint analysis of FOSL1 in psoriatic lesions (PP) and healthy skin (NN). (D) Footprint analysis of FOSL1 in M5-stimulated (M) and unstimulated (N) HaCaT cells. (E) Relative mRNA expression levels of FOSL1 in NN, PN, and PP samples (GSE13355: one-way ANOVA with Tukey multiple comparisons test; GSE54456: two-tailed t test; GSE30999: paired ratio two-tailed t test). (F) Spearman correlations between FOSL1 expression and PASI scores (n = 149 patients). (G) Changes in FOSL1 transcription levels in psoriatic lesions after treatment with adalimumab and brodalumab (GSE85034: n = 30 patients, one-way ANOVA with Tukey multiple comparisons test; GSE53552: n = 24 patients, paired ratio two-tailed t test). (H) Protein expression of FOSL1 in NN and PP. (n = 6 patients, two-tailed t test). All error bars represent SD, ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05.

To identify the change of FOSL1 expression in psoriasis, we first analyzed three independent GEO datasets including GSE13355 from Nair’s study (22), GSE30999 from Suárez-Fariñas’s study (23), and GSE54456 from Li’s study (24) and consistently observed significantly elevated FOSL1 mRNA levels in PP skin compared with both NN and PN samples (Fig. 4E). In addition, FOSL1 mRNA levels were positively correlated with the PASI score (25) (Fig. 4F). Notably, a reduction in FOSL1 mRNA levels was found in psoriatic skin following treatment with both anti–TNF-α (adalimumab) (25) and anti–IL-17 (brodalumab) (26) therapies, suggesting a potential overlap in the molecular response to these treatments (Fig. 4G). We also confirmed up-regulation of FOSL1 protein level in psoriasis lesions (Fig. 4H). Analysis of the single-cell RNA-seq dataset GSE173706 (27) revealed the specifically elevated FOSL1 expression in keratinocytes from patients with psoriasis compared with those from healthy individuals (fig. S8, A to C). In contrast, other AP-1 family TFs such as JUNB, FOSL2, FOS, and ATF3 did not display keratinocyte-predominant expression (fig. S8D).

We further performed CUT&Tag, ATAC-seq, and RNA-seq in HaCaT cells with FOSL1 knockdown. First, a global analysis of FOSL1-binding sites relative to annotated TSS showed that most FOSL1 peaks were found at intragenic and intergenic regions, positioned far from promoters (fig. S9, A and B). This indicates that FOSL1 may primarily bind to enhancer regions rather than promoter regions. As expected, DNA motif analysis revealed a significant enrichment of binding motifs for AP-1 family members within FOSL1 peak regions, with the FOSL1 motif ranking at the top of the list (fig. S9C). Then, to investigate the relationship between FOSL1 occupancy and chromatin accessibility, we integrated FOSL1 CUT&tag and ATAC-seq data. Chromatin-accessible regions bound by FOSL1 exhibited significantly higher chromatin accessibility compared to those without FOSL1 bound (Fig. 5, A and B). Approximately 75% of FOSL1 peaks colocalized with H3K27ac signals (Fig. 5C). In HaCaT cells with FOSL1-knockdown, both H3K27ac levels and chromatin accessibility were reduced at 333 psoriasis-associated enhancers (Fig. 5D). Moreover, FOSL1 loss decreased the number of detectable SEs from 1163 to 805 (Fig. 5E). Notably, 92.3% of SEs, compared with 26% of typical enhancers, contained FOSL1-binding sites (Fig. 5F), and H3K27ac intensity showed strong correlation with FOSL1 occupancy across SEs (Fig. 5G). GSEA further confirmed that FOSL1 knockdown significantly attenuated the transcriptional programs driven by M5-induced enhancers (Fig. 5H). Together, the above analysis indicates the potential role of FOSL1 in regulating the activities of enhancers and SEs and gene transcription in psoriasis lesions.

Fig. 5. FOSL1-coordinated regulation of gene expression via enhancers and SE.

(A) Proportion of FOSL1-binding sites within ATAC-seq–enriched peaks. (B) ATAC-seq signal density profiles at genomic regions bound by FOSL1 and at regions without FOSL1 binding. (C) Genomic distribution and overlap of FOSL1-binding sites with H3K27ac-modified regions. (D) Effects of FOSL1 knockdown on chromatin accessibility and H3K27ac histone modification at 333 psoriasis-associated enhancers in HaCaT cells. (E) SE landscapes in HaCaT cells transfected with siNC or siFOSL1, defined by H3K27ac enrichment. (F) Percentage of enhancers and SE bound by FOSL1. (G) Spearman correlations between coverage of FOSL1 and H3K27ac signals across SE regions. (H) GSEA analysis of genes associated with H3K27ac up-regulated peaks and the 333 enhancers after FOSL1 knockdown. (I) Genomic organization of the IL-1 family gene-regulating SE (IL1-SE), showing five constituent enhancer elements (E1 to E5) defined by H3K27ac enrichment, chromatin accessibility, and FOSL1 binding. (J and K) Relative mRNA expression levels of IL1A, IL1B, and IL36G in HaCaT cells following CRISPRi-mediated repression (J) and CRISPRa-mediated activation (K) of individual enhancer elements (E1 to E5) within the IL1-SE region, with each gene compared to its corresponding negative control (NC) (n = 3 independent biological replicates, one-way ANOVA with Tukey multiple comparisons test). (L) Schematic illustration of dual-luciferase reporter constructs containing individual or combined enhancer elements and FOSL1 motif mutations. (M and N) Dual-luciferase reporter assays showing FOSL1-dependent transcriptional activity of IL1-SE enhancer elements in Rluci-293T (M) and HaCaT (N) cells, with FOSL1-overexpressing groups compared to the negative control (NC) group. Firefly luciferase activity was normalized to Renilla luciferase activity (n = 4 independent biological replicates, one-way ANOVA with Tukey multiple comparisons test). All error bars represent SD, ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05.

FOSL1 cooperates with IL1-SE to regulate selected IL-1 family gene expression

We next explored whether FOSL1 regulates selected IL-1 family genes expression via IL1-SE. First, analysis of bulk RNA-seq (GSE85034) (25) from psoriatic lesions revealed a significant positive correlation between FOSL1 expression and IL1A, IL1B, and IL36G (fig. S9D). Consistently, FOSL1 depletion suppressed M5-induced expression of IL1A, IL1B, and IL36G, while FOSL1 overexpression enhanced their expression (fig. S9, E and F). Similar results were observed in HEKs. M5 stimulation markedly induced FOSL1 protein expression (fig. S9G). Functionally, FOSL1 knockdown reduced, whereas FOSL1 overexpression increased, the expression of IL1A, IL1B, and IL36G in primary keratinocytes (fig. S9, H and I).

Furthermore, on the basis of colocalization of FOSL1 and H3K27ac signals, we partitioned IL1-SE into five typical enhancers (E1, E2, E3, E4, and E5) (Fig. 5I). To validate the functional contribution of each individual enhancer, we used CRISPRi and CRISPR activation (CRISPRa) to selectively repress or activate E1 to E5, respectively. CRISPRi-mediated silencing of single enhancer resulted in a significant reduction in IL1A, IL1B, and IL36G mRNA expression, with a weaker effect on IL36G observed in the E1 and E5 groups (Fig. 5J). Conversely, CRISPRa-mediated activation of each enhancer up-regulated the transcription of these selected IL-1 family genes, with E3 and E4 showing stronger transcriptional activation effects (Fig. 5K).

To validate whether FOSL1 could directly drive the activity of these enhancers, we cloned each enhancer sequence into pGL3-promoter vectors (Fig. 5L). Overexpression of FOSL1 significantly increased the luciferase activity of all five individual enhancers in both RLuci-293 T and HaCaT cells. Furthermore, a reporter construct containing the five enhancers in tandem exhibited stronger activity than any single enhancer alone. Mutation of FOSL1-binding motifs within the tandem enhancer construct abolished this activation, confirming that FOSL1 acts through direct binding to these enhancers (Fig. 5, M and N). These results confirm that FOSL1 binding in IL1-SE play an important role in regulating selected IL-1 family gene expression.

FOSL1-associated condensates modulate enhancer activity and EP300 occupancy in keratinocytes

To demonstrate the mechanism of FOSL1 regulating enhancer activity in keratinocyte, we detected whether FOSL1 undergo LLPS condensates to form nuclear condensates (15). 1,6-Hexanediol (1,6-Hex) is a chemical known to disrupt weak hydrophobic interactions underlying LLPS. Assay for chromatin-bound condensates by exploratory sequencing (ACC-seq) was designed to identify condensate binding regions by detecting DNA accessibility differences under native, fix, and 1,6-Hex + fix conditions using Tn5 transposase (Fig. 6A). The cluster_2 signal enrichment heatmap showed that DNA accessibility decreased in fix samples and recovered in 1,6-Hex + fix samples, while cluster_1 represented the consistent open regions (Fig. 6B). Motif analysis of cluster_2 suggested that c-MYC, PAX8, ACSL2, PAX5, and FOSL1 play a role in condensate formation (Fig. 6C). Besides, differential DNA accessibility in the FOSL1-bound region in the genome exhibited a “deep U-shape” pattern, known as the 1,6-Hex–released pattern, observed under native, fix, and 1,6-Hex + fix conditions (Fig. 6D). These ACC-seq results collectively indicated that FOSL1 associated with chromatin-bound phase-separated condensates in keratinocytes, and its chromatin occupancy was dependent on the LLPS-mediated hydrophobic interactions sensitive to 1,6-Hex. Moreover, analysis of the disordered regions in the FOSL1 protein using the PONDR (28) and IUPred2A (29) tools identified a wide low-complexity region within the FOSL1 coding sequence. This region may promote phase separation and the formation of liquid-like condensates (fig. S10A) (30).

Fig. 6. FOSL1-associated condensates modulate EP300 occupancy and disease-associated enhancer activity.

(A) ACC-seq experiment schedule. (B) Enrichment heatmap of chromatin accessibility signals in two clusters, showing that cluster 2 exhibits reduced accessibility upon fixation and recovery with 1,6-Hex treatment, while cluster 1 maintains stable accessibility across conditions. (C) TF motifs enriched in cluster_2. (D) Relative chromatin accessibility of FOSL1-bound regions under different treatment conditions. (E) FRAP images and fluorescence recovery curves of EGFP-FOSL1 nuclear puncta in 293T cells. (F) Co-IP assay demonstrating interaction between FOSL1 and EP300. IB, Immunoblot. Left, IP with FOSL1 antibody; right, IP with EP300 antibody. Input, whole-cell lysate; IgG, negative control. (G) Colocalization of EGFP-FOSL1 and mCherry-EP300 nuclear puncta in 293T cells cotransfected with EGFP-FOSL1 and mCherry-EP300 vectors, observed by confocal microscopy. Nuclei stained with Hoechst. Scale bar, 5 μm. (H) Enrichment patterns of EP300 CUT&Tag signals around 333 differentially active enhancers of HaCaT. The x axis indicates the genomic distance relative to enhancer center in kilobases (Kb), where −3.0 represents 3 kb upstream and +3.0 represents 3 kb downstream of the enhancer center. The y axis shows the normalized read density reflecting signal intensity in siNC and siFOSL1 groups. (I) Enrichment patterns of EP300 CUT&Tag signals around global FOSL1-binding regions. (J) Relative mRNA expression levels of IL1A, IL1B, and IL36G upon M5 stimulation with or without EP300 inhibitor C646 intervention (n = 3 independent biological replicates, two-way ANOVA with Sidak test for multiple comparisons). All error bars represent SD, ****P < 0.0001; ***P < 0.001.

Furthermore, we constructed an enhanced green fluorescent protein (EGFP)–tagged FOSL1 vector and transfected into 293T cells. We observed that ectopic expression of EGFP-tagged FOSL1 in 293T cells resulted in the formation of discrete nuclear puncta, indicative of condensate assembly. As shown in fig. S10B, treatment with 1,6-Hex led to a near-complete dissolution of EGFP-FOSL1 nuclear puncta, suggesting that FOSL1 condensate formation depends on hydrophobic intermolecular interactions. Besides, disruption of condensate formation significantly reduced the mRNA expression of IL1A, IL1B, and IL36G (fig. S10C). Liquid-like condensates are characterized by dynamic behaviors, including droplet fusion and rapid molecular exchange. To assess the dynamic properties of FOSL1 condensates, fluorescence recovery after photobleaching (FRAP) assays were performed. Following photobleaching, the fluorescence intensity within EGFP-FOSL1 condensates recovered rapidly to approximately 55% of the initial signal (Fig. 6, E), demonstrating active molecular exchange within the condensates.

Our earlier findings demonstrated that FOSL1-bound regions extensively colocalize with H3K27ac marks and are highly enriched at enhancers. Since H3K27ac deposition is catalyzed by the histone acetyltransferase EP300, we hypothesized that FOSL1 might contribute to enhancer activation by maintaining EP300 occupancy at these loci. Coimmunoprecipitation (Co-IP) and Western blot assays in HaCaT cells confirmed a specific interaction between FOSL1 and EP300, as EP300 (∼300 kDa) was detected following FOSL1 immunoprecipitation and FOSL1 (∼40 kDa) was found in EP300 pull-downs, whereas no signal was observed in immunoglobulin G (IgG) controls (Fig. 6F). Intrinsic disorder analysis using PONDR and IUPred2 predicted that more than 70% of EP300 consists of intrinsically disordered regions, indicating a high potential for LLPS formation (fig. S10D). Immunofluorescence revealed nuclear colocalization of FOSL1 and EP300 with highly overlapping fluorescence intensity profiles (fig. S10E). Coexpression of EGFP-FOSL1 and mCherry-EP300 in 293T cells demonstrated pronounced nuclear colocalization within punctate structures, supporting their co-assembly into LLPS condensates (Fig. 6G).

EP300 occupancy was globally reduced across M5-induced SE regions, as well as at FOSL1 binding regions, following FOSL1 depletion in HaCaT cells (Fig. 6, H and I). In addition, EP300 enrichment at the IL1-SE locus was modestly decreased after FOSL1 knockdown (fig. S10F). These findings suggest that FOSL1 contributes to maintaining EP300 occupancy at active SEs and FOSL1-bound regulatory regions. Furthermore, treatment with 10 μM C646, a selective EP300 inhibitor, suppressed the mRNA expression of IL1A, IL1B, and IL36G in M5-stimulated HaCaT cells (Fig. 6J and fig. S10G), supporting a functional cooperation between FOSL1 and EP300 in regulating psoriasis-associated inflammatory gene expression.

Myeloid-derived IL-1 cytokines induce FOSL1 expression in psoriatic keratinocytes through IL-1R–ERK signaling

To identify upstream signaling pathways regulating FOSL1 expression in keratinocytes, we examined pathways previously implicated in psoriasis pathogenesis, with a particular focus on the mitogen-activated protein kinase (MAPK) cascade (31). HaCaT cells were stimulated with the M5 cytokine cocktail in the presence of selective MAPK inhibitors. Treatment with the ERK1/2 inhibitor SCH772984 over a broad concentration range did not affect cell viability, indicating minimal cytotoxicity under the experimental conditions (fig. S11A). Pharmacological inhibition of ERK signaling using SCH772984 significantly reduced M5-induced FOSL1 expression at both the mRNA and protein levels in a dose-dependent manner (fig. S11, B and C). Consistently, SCH772984 treatment effectively suppressed ERK1/2 phosphorylation without altering total ERK levels, confirming efficient blockade of ERK signaling (fig. S11C). In contrast, ERK inhibition under M5 stimulation did not uniformly suppress selected IL-1 family genes expression, with partial reduction of IL1A and IL1B but minimal effect on IL36G, suggesting involvement of additional cytokine-dependent signaling pathways (fig. S11, D to F). However, p38 blockade by adezmapimod did not affect M5-induced FOSL1 expression at either mRNA or protein levels, despite effectively suppressing p38 phosphorylation, indicating that FOSL1 induction relies on ERK rather than p38 MAPK signaling (fig. S11, G to I).

Given that IL-1 family cytokines play prominent roles in psoriasis and are associated with ERK pathway, we next examined whether IL-1α, IL-1β, and IL36γ directly regulate FOSL1. Stimulation of HaCaT cells with individual cytokine showed that IL-1α and IL-1β, but not IL-36γ, induced FOSL1 expression (Fig. 7, A and B). Furthermore, IL-1α and IL-1β stimulation combined with ERK inhibition markedly suppressed FOSL1 protein expression (Fig. 7, C and D), along with strong down-regulation of IL1A, IL1B, and IL36G expression (Fig. 7, E and F). Myeloid cells are a primary source of IL-1 cytokines in psoriatic skin (fig. S11J) (32). We further tested whether macrophage-derived factors regulate FOSL1 in keratinocytes. HaCaT cells exposed to macrophage-conditioned medium showed marked up-regulation of FOSL1, IL1A, IL1B, and IL36G expression. Addition of the IL-1 receptor antagonist substantially reduced expression of FOSL1 and selected IL-1 family genes (Fig. 7, G and H), suggesting that macrophage-derived factors acting through the IL-1R–ERK–FOSL1 signaling induce the inflammatory response in keratinocytes.

Fig. 7. IL-1 signaling regulates FOSL1 expression via ERK pathway activation in keratinocytes.

(A) Western blot analysis of FOSL1 protein levels in HaCaT cells stimulated with IL-1α, IL-1β, or IL-36γ for 24 hours. α-Tubulin was used as a loading control. (B) Relative mRNA expression levels of FOSL1, IL1A, IL1B, and IL36G mRNA levels in HaCaT cells stimulated with IL-1α, IL-1β, or IL-36γ for 24 hours. Data are means ± SD (n = 3 independent biological replicates, one-way ANOVA with Tukey multiple comparisons test). (C and D) Western blot analysis of FOSL1, phosphorylated ERK1/2 (p-ERK1/2), and total ERK1/2 in HaCaT cells pretreated with the ERK inhibitor SCH772984 (10 μM), followed by stimulation with IL-1α (C) or IL-1β (D) for 24 hours. α-Tubulin served as a loading control. (E and F) Relative mRNA expression levels of FOSL1, IL1A, IL1B, and IL36G mRNA levels in HaCaT cells pretreated with SCH772984 (10 μM), followed by stimulation with IL-1α (E) or IL-1β (F) for 24 hours. Data are means ± SD (n = 3 independent biological replicates, two-way ANOVA with Sidak test for multiple comparisons). (G) Western blot analysis of FOSL1 protein levels in HaCaT cells treated with macrophage-conditioned medium (CM-T) in the presence or absence of the IL-1 receptor antagonist (Raleukin, 10 μg/ml) for 24 hour. α-Tubulin served as a loading control. (H) Relative mRNA expression levels of FOSL1, IL1A, IL1B, and IL36G mRNA levels in HaCaT cells treated with CM-T ± Raleukin (10 μg/ml) for 24 hour. Data are means ± SD (n = 3 independent biological replicates, one-way ANOVA with Tukey multiple comparisons test). All error bars represent SD, ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05.

In addition, under macrophage-conditioned medium stimulation, FOSL1 knockdown significantly suppressed the expression of IL1A, IL1B, and IL36G, indicating that FOSL1 is required for macrophage-induced activation of selected IL-1 family genes (fig. S11K). Together with the ability of IL-1 cytokines to induce FOSL1 expression, these findings support the existence of a positive inflammatory feedback circuit that may contribute to the amplification and maintenance of psoriatic inflammation.

DISCUSSION

Enhancers, marked by H3K27ac, serve as key epigenetic regulators that coordinate gene expression (6). Altered enhancer activity affects gene expression and contributes to disease (10). Previous studies showed that most enhancers regulate genes affecting epidermal homeostasis or immune function in psoriasis and atopic dermatitis (6, 33). In this study, integrated H3K27ac ChIP-seq, ATAC-seq, and RNA-seq analyses revealed widespread enhancer activation in psoriatic lesions, predominantly in gene-distal regulatory regions. We identified 418 psoriasis-associated enhancers enriched for inflammatory pathways, regulating genes in JAK-STAT, cytokine-cytokine receptor, and IL-17 signaling. These results support a role for enhancers as important regulatory elements associated with pathogenic transcription and broaden our understanding of epigenetic mechanism in inflammation skin diseases.

In addition to typical enhancers, we identified SE-driven transcriptional programs as key amplifiers of inflammation. SEs serve as high-capacity transcriptional hubs, occupancy of active histone marks, mediator complex, and BRD4, and regulate the expression of genes controlling cell identity, immune responses, or disease development (34). For example, SE located between human leukocyte antigen(HLA)–DRB1 and HLA-DQA1 can simultaneously regulate the expression of both HLA-DR and HLA-DQ. In individuals carrying high-risk haplotypes, immune cells produce more proinflammatory factors upon pathogen stimulation, promoting the development of vitiligo (35, 36). SE located between CD274 and CD273 (PD-L1L2-SE) can simultaneously drive high expression of PD-L1 and PD-L2. Deletion of PD-L1L2-SE significantly reduces the immune evasion ability of tumor cells, indicating its critical role in regulating immune checkpoint molecule expression and tumor immune escape (37). In alcoholic hepatitis, 4C-seq revealed that an SE in liver sinusoidal endothelial cells induced the expression of multiple CXCL chemokines (CXCL1, CXCL2, CXCL3, CXCL6, and CXCL8), thereby promoting neutrophil infiltration in the liver (38). Similar to previous reports showing that a single SE selectively regulates a subset of chemokine genes, our data revealed that the IL1-SE preferentially regulated IL1A, IL1B, and IL36G, while exerting minimal effects on other IL-1 family members. This selective regulation may be attributable to preferential chromatin interactions between the IL1-SE and the promoters of IL1A, IL1B, and IL36G, whereas other IL-1 family genes may be spatially insulated by local chromatin architecture. Although a putative S100A-associated enhancer region was observed following M5 stimulation, the lack of overlap with annotated keratinocyte SEs suggests that regulation of the S100 gene cluster may differ from the IL1-SE–driven mechanism identified for selected IL-1 family genes. Notably, we further observed that different enhancer elements within the IL1-SE did not contribute equally to gene regulation. The differential responsiveness of selected IL-1 family genes to individual enhancer perturbation suggests functional heterogeneity within the IL1-SE region. Among these enhancers, E3 and E4 exhibited stronger H3K27ac enrichment and consistently showed more pronounced effects in both CRISPRi and CRISPRa assays, indicating that they may function as core regulatory enhancer elements within the IL1-SE. In contrast, IL36G displayed relatively weaker responses to perturbation of certain enhancers, which may be partly attributable to differences in chromatin spatial organization and enhancer-promoter interaction distance compared with IL1A and IL1B. In addition, partial functional compensation between different enhancer elements may also explain why perturbation of individual enhancers resulted in variable regulatory effects. As key mediators of psoriasis pathogenesis, IL-1 family cytokines play pivotal roles in driving keratinocyte dysfunction and immune activation in lesional skin of patients with psoriasis (39–41). Recent advances in enhancer-targeted genome editing have enabled precise therapeutic interventions. CRISPR-Cas9–mediated disruption of the BCL11A enhancer in CD34⁺ hematopoietic stem cells reactivates γ-globin expression and fetal hemoglobin production, a strategy that has received Food and Drug Administration approval for the treatment of sickle cell disease and transfusion-dependent β-thalassemia in patients aged 12 years and older (42). Our findings suggest that IL1-SE may represent a feasible upstream target for CRISPR-based epigenetic or genome-editing interventions in psoriasis.

Small-molecule inhibitors targeting the BET family (e.g., JQ1) competitively block the binding of BET proteins to acetylated histones, thereby suppressing the transcriptional activation of genes downstream of SEs (43). Specifically, JQ1 has demonstrated remarkable efficacy in various tumor models (20). Beyond oncology applications, BET inhibitors also exhibit therapeutic potential for inflammatory diseases and allergy. For instance, JQ1 effectively alleviates inflammation-driven colonic injury in endotoxemia by inhibiting NF-κB phosphorylation and activation of the NLRP3/ASC/caspase-1 inflammasome complex, which in turn reduces the levels of proinflammatory cytokines (IL-6, IL-1β, and IL-18) and blocks pyroptosis (44). Disruption of IL9-SEs by JQ1 abolished T_H9 cell induction in vitro and inhibited T_H9 cell–mediated allergic airway inflammation in vivo (45). Notably, JQ1 administration markedly repressed SE-associated genes and alleviated psoriasis-like pathology in our experiments, by inhibiting keratinocyte hyperproliferation and inducing apoptosis. Although BET inhibition by JQ1 effectively suppressed psoriasis-associated inflammatory gene expression and disease phenotypes, the direct occupancy of BRD4 at the identified enhancer and SE regions was not examined in the present study. Future studies will be required to determine the genome-wide binding landscape of BRD4 and its functional interaction with FOSL1 at psoriasis-associated regulatory elements. While JQ1 showed efficacy in ameliorating psoriasis-like skin inflammation, its pharmacokinetic properties and potential side effects require further investigation. Optimization through medicinal chemistry, improving bioavailability and half-life, and development of next-generation selective BET inhibitors could enhance translational potential.

FOSL1 is a member of the FOS gene family and forms heterodimers with JUN family proteins to constitute the AP-1 complex, which widely participated in fundamental biological processes, including cell proliferation, differentiation, and apoptosis (46). FOSL1 as a key oncogenic TF that drives tumor proliferation, invasion, metastasis, and phenotypic plasticity across multiple cancer types through diverse transcriptional programs (47). GEO dataset analysis showed that FOSL1 is markedly up-regulated at transcriptional and protein levels in psoriatic lesions, correlating with disease severity. Previous studies have shown that DNA hypomethylation and histone hyperacetylation contribute to many pathogenic gene transcriptional activation and psoriasis pathogenesis (48). However, the mechanism of aberrant epigenetic modifications remain still unclear. In this study, our data suggest that FOSL1 forms LLPS-associated condensates and interacts with EP300, which may contribute to a transcriptionally permissive environment that supports enhancer activity. Although the present study mechanistically focused on the IL1-SE because of its prominent inflammatory relevance, genome-wide analyses showed that FOSL1 depletion broadly reduced chromatin accessibility and H3K27ac enrichment across 333 differentially active enhancers, suggesting that FOSL1 broadly contributes to enhancer-mediated inflammatory transcriptional programs rather than functioning exclusively at the IL1-SE locus. Notably, our study also identified multiple psoriasis-associated SEs located near inflammatory gene clusters, including the S100A and IL10 families. Although these loci were not functionally investigated in the present study, they represent promising candidates for future studies to determine whether FOSL1 similarly regulates their enhancer activity and associated gene expression. Notably, while FOSL1 was required for enhancer-associated inflammatory gene expression, its overexpression alone failed to induce downstream target genes in unstimulated keratinocytes, although it further enhanced their expression under M5 stimulation. These findings indicate that FOSL1 is necessary but not sufficient to establish enhancer or SE activity de novo and likely functions within an inflammation-induced regulatory context that provides the chromatin accessibility and cofactor support required for enhancer activation.

Recent studies have revealed that TFs could exploit LLPS to regulate SE activity and drive disease progression. Through its ZF3 domain, SP1 engaged in LLPS, facilitating SE formation and RGS20 activation to drive lung adenocarcinoma progression (49). Similarly, FOXP1 underwent LLPS at SEs to form transcriptional condensates that activate SESN3, suppress mTORC1 signaling, and thereby retarded endothelial cell senescence and atherosclerosis (50). In addition, LLPS is also increasingly recognized as a potential strategy for therapeutic intervention in diseases (51). For example, the oncogenic TF FOXM1 sustains SE activity via LLPS to drive tumor metastasis, and AMPK agonists disrupt FOXM1 condensates by phosphorylating its IDRs, thereby suppressing tumor growth and enhancing antitumor immunity (52). Building on these findings, future studies should focus on (i) defining the specific domains within FOSL1 and EP300 essential for their interaction and condensate formation via mutagenesis and domain-deletion analyses and (ii) identifying small-molecule inhibitors that inhibit FOSL1 phase separation to evaluate their therapeutic potential in psoriasis.

Aberrant cross-talk between macrophages and keratinocytes shapes psoriasis’ pathogenic immune microenvironment (53). Our data indicate that myeloid-derived selected IL-1 family cytokines up-regulate FOSL1 expression in keratinocytes through activation of the ERK pathway rather than the p38 MAPK signaling. This is consistent with previous reports that IL-1α and IL-1β activate ERK to regulate downstream proinflammatory transcriptional programs in keratinocytes (54). Given that FOSL1 is a core AP-1 component with established ERK responsiveness, ERK activation may represent a more effective signaling axis for FOSL1 induction than p38 (55). Consistently, although IL-1α, IL-1β, and IL-36γ belong to the same cytokine family, they signal through distinct receptors and may therefore exert different effects on FOSL1 expression. IL-1α and IL-1β primarily activate IL-1R1–mediated signaling, which is closely associated with MAPK/AP-1 pathway activation and may more effectively induce FOSL1 expression in keratinocytes. In contrast, IL-36γ signals through the IL-36 receptor complex and may exhibit different downstream signaling properties, which could explain its inability to significantly up-regulate FOSL1 despite overlapping inflammatory functions with IL-1 cytokines. Elevated FOSL1 was associated with increased enhancer activity and expression of selected IL-1 family genes, suggesting a potential mechanism that may amplify inflammatory signaling. This establishes a self-amplifying inflammatory loop linking immune cues to chromatin remodeling in psoriatic lesions.

While our study underscores the pivotal role of IL-1 signaling in psoriatic pathogenesis, we acknowledge that IL-17 and IL-23 inhibitors currently represent the gold standard for clinical efficacy, achieving high rates of skin clearance (PASI 90/100) in most patients with psoriasis (56, 57). However, we posit that the IL-1 pathway remains a clinically meaningful and nonredundant therapeutic target for several strategic reasons. First, IL-1 acts as an upstream master regulator in the inflammatory cascade. Unlike IL-17 or IL-23, which operate further downstream in the adaptive immune response, IL-1 (particularly IL-1α and IL-1β) bridges innate and adaptive immunity. It directly activates keratinocytes to produce antimicrobial peptides and chemokines and, critically, drives the differentiation of T_H17 cells and the production of IL-17 by γδ T cells (58, 59). Therefore, targeting IL-1 may suppress the inflammatory axis at its source, potentially offering a broader immunomodulatory effect that is not fully replicated by downstream blockade alone. Second, the translational implication of our findings suggests that IL-1–directed therapy may offer distinct value in selected patient subgroups. Specifically, patients with neutrophil-rich phenotypes (e.g., pustular psoriasis) or those who exhibit primary nonresponse or secondary loss of response to IL-17/23 inhibitors may benefit from IL-1 inhibition. Furthermore, given the established link between IL-1 signaling and systemic comorbidities such as cardiovascular disease and metabolic syndrome, IL-1 antagonists could provide additional extracutaneous benefits, addressing the systemic inflammatory burden often seen in severe psoriasis (60, 61). Last, rather than viewing IL-1 inhibition as a direct replacement for current biologics, we propose its potential utility in combination strategies. Low-dose IL-1 blockade combined with IL-17 or IL-23 inhibitors might synergistically enhance efficacy, reduce the required dosage of expensive biologics, or prevent the emergence of resistance mechanisms. Future clinical trials should focus on these niche populations and combination regimens to fully elucidate the translational potential of IL-1 targeting in precision medicine for psoriasis.

This study has several limitations. First, although we have identified a lot of psoriasis-associated enhancers based on the integrated multiomics analyses, the function of these enhancers and SEs still need to be validated in more disease samples and in vitro experiments. Second, as the conservation of enhancers is relatively poor across species, it is difficult to manipulate specific enhancers such as IL1-SE in human and mice cells. Third, the keratinocyte-associated FOSL1 gene knockout studies in mice model need to be performed to validate its role in the development of psoriasis in the future. Last, this study focused on enhancer and SE in lesional keratinocytes under inflammatory conditions but did not address the cellular origin of this pathogenic transcriptional program. Whether basal stem or progenitor keratinocytes contribute to its initiation, persistence, or inflammatory memory remains unclear. Future lineage-tracing and single-cell multiomics studies will be required to clarify the cellular sources and hierarchy underlying enhancer-mediated inflammatory programs in psoriasis.

Overall, our findings revealed that the activated enhancers and SEs in psoriasis and uncovered an SE-driven mechanism in psoriasis, with FOSL1 integrating inflammatory signals and activating IL1-SE via phase separation (Fig. 8). These upstream regulatory hubs are associated with pathogenic gene expression and may represent promising targets for therapeutic intervention. Our findings provided a foundation for developing enhancer-directed or genome-editing therapies, offering potential for more durable and precise treatment of psoriasis.

Fig. 8. Graphical summary of the main findings of the manuscript.

This study reveals that psoriasis is associated with aberrant activation of keratinocyte enhancers and SEs. The TF FOSL1 is associated with condensate formation and EP300 occupancy at an IL-1 family SE, potentially contributing to its activation. IL-1–ERK signaling reciprocally induces FOSL1, establishing a feed-forward enhancer circuit that amplifies epidermal inflammation [Created in BioRender. Qiu, Y. (2026) https://BioRender.com/j28rnpp].

MATERIALS AND METHODS

Experimental animals

Female BALB/c mice (6 to 8 weeks old) were purchased from Beijing SPF Biotechnology Co. Ltd. (Beijing, China). All animals were housed under specific pathogen–free conditions at Hospital for Skin Diseases, Institute of Dermatology, Chinese Academy of Medical Sciences and Peking Union Medical College. Mice were maintained under standardized temperature, humidity, and 12-hour light/dark cycles with ad libitum access to food and water. All animal experiments were conducted in accordance with the German Animal Welfare Act and approved by the Institutional Animal Ethics Committee of the Hospital for Skin Diseases, Institute of Dermatology, Chinese Academy of Medical Sciences and Peking Union Medical College (approval no. 2024-DW-041). Age- and weight-matched mice were used in all experiments to ensure experimental consistency and reproducibility. For the IMQ-induced psoriasis-like model, mice received a daily topical application of IMQ cream on the shaved dorsal skin for five consecutive days. Six hours after each IMQ application, a gel containing JQ1 (100 μg/cm²) or vehicle control was topically applied to the dorsal skin. For the IL-23–induced psoriasis-like model, recombinant mouse IL-23 (mIL-23; abclonal, RP02928S) was dissolved in phosphate-buffered saline (PBS), and 1 μg of rmIL-23 in 25 μl of PBS was intradermally injected into one ear daily for three consecutive days. The contralateral ear received an equal volume (25 μl) of PBS as a control. Six hours after each IL-23 injection, a gel containing JQ1 (100 μg/cm²) or vehicle control was topically applied to the injected ear.

Cell lines

Human immortalized keratinocyte cell line HaCaT, human T lymphoblast cell line Jurkat, and human monocyte cell line THP-1 were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and authenticated by short tandem repeat (STR) profiling. HEKs (catalog no. FH-H090, FuHeng Biology) were maintained in customized HEK complete medium (catalog no. PY-H090, FuHeng Biology). The Rluci-293T stable cell line (human embryonic kidney 293T cells expressing Renilla luciferase) was provided by X. He (The Second Xiangya Hospital, Central South University, China). HaCaT and Rluci-293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, #11965118), whereas Jurkat and THP-1cells were maintained in RPMI 1640 medium (Gibco, #11875093). All media were supplemented with 10% fetal bovine serum (FBS, Gibco, #16000044) and 1% penicillin-streptomycin (Gibco, #15140122). Cells were maintained at 37°C in a humidified incubator with 5% CO₂.

Human skin samples

Human skin paraffin-embedded sections were obtained from patients diagnosed with psoriasis and from healthy controls recruited at the Hospital for Skin Diseases, Institute of Dermatology, Chinese Academy of Medical Sciences and Peking Union Medical College. Six psoriatic lesional samples were collected from clinically and histopathologically confirmed patients, and six normal skin samples were collected from all of them. All participants were ≥18 years of age, clinically evaluated, and provided written informed consent (table S1). The study was approved by the Institutional Ethics Committee of the Hospital for Skin Diseases, CAMS (approval no. 2024-KY-016).

Data acquisition and preprocessing

Public datasets of H3K27ac ChIP-seq [GSE161076 (62)], ATAC-seq (PRJNA597655) (63), and RNA-seq (GSE54456) (64) from psoriatic lesions, perilesional skin, and healthy controls were retrieved from the GEO database. Raw reads were downloaded from the National Center for Biotechnology Information Sequence Read Archive (SRA) and converted to FASTQ format using SRA-tools (https://github.com/ncbi/sra-tools). Adapter trimming and quality filtering were performed with Trimmomatic, followed by quality control using FastQC. High-quality reads were aligned to the human reference genome (hg38) using Bowtie2 for ChIP-seq and ATAC-seq while HISAT2 for RNA-seq. Duplicate reads were removed by Picard MarkDuplicates, and reads with mapping quality <Q30 were excluded. Quality control metrics were evaluated according to ENCODE-recommended standards. Samples with TSS enrichment scores >5 and fraction of reads in peaks scores >0.2 were considered high quality. In addition, libraries with >20 million uniquely mapped nonduplicate reads were retained for downstream analyses.

CUT&Tag library preparation and sequencing

CUT&Tag libraries were generated using the Hyperactive Universal CUT&Tag Assay Kit for Illumina Pro (Vazyme, TD904) according to the manufacturer’s instructions. Briefly, 1 × 10⁵ HaCaT cells were collected for each sample and lysed to isolate nuclei. Nuclei were immobilized on ConA-coated magnetic beads and incubated overnight at 4°C with antibodies against H3K27ac (abcam, ab4729), FOSL1 (CST, 5281S), or EP300 (CST, 86377) (all diluted 1:50), followed by incubation with a secondary antibody and pA/G-Tn5 transposome. After targeted tagmentation and DNA extraction, sequencing libraries were amplified by polymerase chain reaction (PCR) using indexed primers. Library quality was assessed using an Agilent 2100 Bioanalyzer, and paired-end sequencing was performed on an Illumina platform.

ATAC-seq library preparation and sequencing

ATAC-seq libraries were prepared using the Hyperactive ATAC-Seq Library Prep Kit for Illumina (Vazyme, TD711) following the manufacturer’s protocol. Briefly, 1 × 10⁵ HaCaT cells were harvested and lysed to obtain nuclei. Chromatin accessibility was profiled by Tn5 transposase-mediated tagmentation at 37°C for 30 min. Following DNA purification, libraries were amplified by PCR using indexed primers and purified using magnetic beads. Library quality was evaluated with an Agilent 2100 Bioanalyzer, and paired-end sequencing was performed on an Illumina platform.

ChIP-seq and ATAC-seq analysis

Peak calling was conducted using MACS2 (65) with a q value cutoff of 0.05. For H3K27ac ChIP-seq data, broad peak calling was performed using the “--broad” option with default broad cutoff settings. Differential peak analysis was performed with DiffBind (FDR < 0.05, |log2FC| > 1). Peak annotation and genomic distribution analysis were carried out with ChIPseeker (66, 67). H3K27ac enrichment profiles were visualized using deepTools (68) and Integrative Genomics Viewer (69).

Identification of enhancers and SEs

Typical enhancers were defined as distal (>2 kb from TSS) H3K27ac peaks, whereas SEs were identified using the ROSE (Rank Ordering of Super Enhancers) (70). Enhancers within 12.5 kb were stitched together and ranked based on total H3K27ac signal intensity, while regions within ±2 kb of TSSs were excluded from stitching. The inflection point of the ranked curve was used to distinguish typical from SEs.

RNA-seq and processing

Total RNA was extracted from HaCaT cells using the TRIzol reagent (Invitrogen, #15596018CN) according to the manufacturer’s protocol. RNA purity and concentration were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific), and RNA integrity was assessed with the Agilent 2100 Bioanalyzer (Agilent Technologies). RNA-seq libraries were constructed using the VVAHTS Universal V8 RNA-seq Library Prep Kit for Illumina (Vazyme, NR605) following the manufacturer’s instructions. Transcriptome sequencing and bioinformatic analyses were performed by OE Biotech Co. Ltd. (Shanghai, China).

Sequencing was carried out on an Illumina NovaSeq 6000 platform to generate 150–base pair (bp) paired-end reads. Raw FASTQ files were processed using fastp for adapter trimming and quality filtering, and low-quality reads were removed to obtain clean reads. Clean reads were mapped to the human reference genome (GRCh38) using HISAT2. Gene-level read counts were quantified using HTSeq-count, and normalized expression levels were calculated as fragments per kilobase of exon model per million mapped reads.

Multiomics integration analysis

RNA-seq reads were quantified using featureCounts and normalized by DESeq2 (71). DEGs were identified with thresholds of adjusted P < 0.05 and |log2FC| > 1. Enhancer–gene pairs were established based on proximity and correlated changes in H3K27ac signal and gene expression levels. Functional enrichment analysis for Gene Ontology, KEGG, Reactome, and WikiPathways was performed using clusterProfiler in R (v3.2.0), and results were visualized as bar, bubble, and chord plots. GSEA was conducted using the GSEA software (Broad Institute, v4.3) to identify significantly enriched pathways. Unless otherwise specified, default parameters were used for all bioinformatic analyses and software tools.

Cell culture and cytokine stimulation

The human immortalized keratinocyte cell line HaCaT was obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and authenticated by STR profiling. Cells were maintained in DMEM (Gibco, 10564011) supplemented with 10% FBS (Gibco, A5256701) and 1% penicillin-streptomycin (Gibco, 15140122) at 37°C in a humidified 5% CO₂ incubator. Cells were treated with a cytokine M5 cocktail including IL-17α (novoprotein, C774), IL-22 (novoprotein, CU85), TNF-α (novoprotein, C008), IL-1α (novoprotein, C070), and oncostatin M (novoprotein, C099), each at 10 ng/ml for 24 hours to mimic psoriatic inflammatory conditions. Untreated HaCaT cells served as controls.

Coculture of THP-1 and HaCaT cells

To determine whether HaCaT is affected by IL-1α and IL-1β from macrophages, THP-1 and HaCaT conditioned media culture model was established. THP-1 was induced to differentiate into macrophages using phorbol 12-myristate 13-acetate (Sigma-Aldrich, P8139) at 10 ng/ml for 24 hours, and then M5 was added to stimulate macrophages. Last, the supernatant was collected and transferred to HaCaT cell culture, and Raleukin (20 μg/ml; MCE, HY-108841) was added.

Cell viability assay (CCK-8)

Cells were treated with increasing concentrations of JQ1 (0 to 400 nM, MCE, HY-13030), SCH772984 (0 to 40 μM, MCE, HY-50846), adezmapimod (0 to 100 μM, MCE, HY-10256), and C646 (0 to 10 μM, MCE, HY-13823) for 24 hours. After treatment, 10 μl of CCK-8 reagent (Vazyme, A311-01) was added to each well and incubated for 2 hours. Absorbance was measured at 450 nm using a microplate reader, and cell viability was calculated as a percentage relative to the untreated control.

Cell proliferation assay (EdU incorporation)

Cell proliferation was assessed using the BeyoClick EdU-488 Cell Proliferation Kit (Beyotime, C0071S). HaCaT cells were seeded in 12-well plates and treated with M5 cytokines (10 ng/ml each) with or without 100 nM JQ1 for 24 hours. Cells were then incubated with 10 μM EdU for 2 hours, fixed with 4% paraformaldehyde, and permeabilized with 0.3% Triton X-100. After washing, click reaction buffer containing Azide-488 was added for 30 min in the dark, followed by Hoechst 33342 nuclear staining. EdU-positive cells (green) and total nuclei (blue) were visualized by fluorescence microscopy.

Apoptosis assay (annexin V–FITC/PI staining)

Apoptosis was analyzed using the annexin V–fluorescein isothiocyanate (FITC) Apoptosis Detection Kit (Beyotime, C1062S). HaCaT cells were treated with M5 and JQ1 (0 to 400 nM) for 24 hours, digested with trypsin without EDTA, and resuspended in binding buffer. Cells were stained with 5 μl of annexin V–FITC and 10 μl propidium iodide (PI) for 15 min in the dark at room temperature and analyzed by flow cytometry within 1 hour.

RNA extraction and RT-qPCR

Total RNA was extracted using TRIzol reagent (Invitrogen, 15596018CN) and reverse-transcribed using the PrimeScript RT Reagent Kit (Takara, RR037A). Gene expression levels were determined using TB Green Premix Ex Taq (Takara, RR420A) on a QuantStudio 6 Real-Time PCR system. Primer sequences for IL-1A, IL-1B, IL-36G, IL-36B, IL-36R, IL-1RN, FOSL1, and β-actin (internal control) were designed and validated (table S2). Relative mRNA levels were calculated by the 2^-ΔΔCt method.

ELISA assay

Supernatants from HaCaT cell cultures were collected and subjected to ELISA detection of IL-1α, IL-1β, and IL-36γ using commercial kits (Elabscience, E-EL-H0088, E-EL-H0149, Cusabio, CSB-EL011620HU). Optical density was measured at 450 nm, and cytokine concentrations were calculated from standard curves.

CRISPR-Cas9–mediated SE knockout

Two single guide RNAs (sgRNAs) targeting the 5′ and 3′ boundaries of the SE were designed using CHOPCHOP and IDT online tools and synthesized by IDT (table S3). sgRNAs were complexed with recombinant Cas9-GFP protein (Alt-R S.p. Cas9 V3, IDT, 10008100) to form ribonucleoproteins and electroporated into HaCaT cells using a Lonza 4D-Nucleofector (program EM-138). GFP-positive cells were sorted by flow cytometry and expanded as single clones in 96-well plates. Genomic deletion of the target SE region was validated by PCR and Sanger sequencing (table S4).

Circular chromosome conformation capture

4C-seq was performed to detect promoter-enhancer interactions at the IL1A-IL36G locus. Cross-linked chromatin was digested with DpnII (NEB, R0543S) and NlaIII (NEB, R0125V), ligated, and amplified using viewpoint-specific primers with Illumina adaptors. Sequencing and data processing were performed on the Illumina NovaSeq platform. Primer sequences for 4C-seq library construction were list in table S5.

Protein extraction and quantification

Total cellular proteins were extracted using radioimmunoprecipitation assay lysis buffer supplemented with protease and phosphatase inhibitors (Beyotime, P1046). Cell pellets were lysed on ice for 30 min, and supernatants were collected after centrifugation at 12,000g for 10 min at 4°C. Protein concentrations were determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, #23227) according to the manufacturer’s instructions. Absorbance was measured at 562 nm, and protein concentrations were calculated from standard bovine serum albumin (BSA) curves.

Western blot analysis

Equal amounts of total protein were denatured in 1× SDS loading buffer (Beyotime, P0015A), separated by 8 to 12% SDS–polyacrylamide gel electrophoresis, and transferred onto polyvinylidene difluoride membranes (Millipore, USA). Membranes were blocked with 5% nonfat milk and incubated overnight at 4°C with primary antibodies against FOSL1 (1:1000, CST, 5281S), EP300 (1:1000, CST, 86377), glyceraldehyde-3-phosphate dehydrogenase (GAPDH, 1:5000, Proteintech, 10494-1-AP), or β-actin (1:2000, Proteintech, 66009-1-Ig). After washing, membranes were incubated with horseradish peroxidase (HRP)–conjugated secondary antibodies (CST, 7074) for 1 hour at room temperature, and bands were visualized using an ECL detection system (Bio-Rad, ChemiDoc).

Coimmunoprecipitation

Co-IP was performed using Protein A/G magnetic beads (MCE, HY-K0202). Cell lysates were precleared with beads and incubated overnight at 4°C with FOSL1 (1:50, CST, 5281S) or EP300 (1:50, CST, 86377) or control IgG (1 μg, CST, 2729). Immunocomplexes were captured with Protein A/G beads, washed, and eluted by boiling in 2× loading buffer. Samples were analyzed by Western blot to detect protein-protein interactions.

Chromatin immunoprecipitation–qPCR

ChIP assays were performed using the EZ-Magna ChIP A Kit (Merck Millipore, 17-10086) following the manufacturer’s protocol. Cells were cross-linked with 1% formaldehyde, sonicated to 200- to 500-bp fragments, and immunoprecipitated with antibodies against FOSL1 (1:50, CST, 5281S) RNA Pol II (Merck Millipore, 05-623) or IgG control. Immunoprecipitated DNA was purified and analyzed by quantitative PCR (qPCR) using TB Green Premix Ex Taq (Takara, RR420A). Enrichment was calculated as a percentage of input DNA using the 2^-ΔCt method. Primer sequences for ChIP-qPCR were list in table S6.

Immunofluorescence

Cells grown on coverslips were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, and blocked with 5% BSA. Samples were incubated with primary antibodies against FOSL1 (1:50, CST, 5281S) or EP300 (1:500, CST, 86377) overnight at 4°C, followed by Alexa Fluor–conjugated secondary antibodies (1:500, Thermo Fisher Scientific, A-11034). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (Beyotime, C1005), and fluorescence was visualized using a Zeiss LSM710 confocal microscope.

CRISPRi assay

To repress enhancer activity, a stable HaCaT cell line expressing dCas9-KRAB was established using lentiviral transduction (GeneChem). Cells were infected at a multiplicity of infection (MOI) of 50 and cultured for 48 hour before fluorescence-activated cell sorting of EGFP-positive cells (>90% purity). Five sgRNAs (sgE1 to sgE5) targeting the IL1-SE region were designed using CRISPOR (table S7). Transfection was performed using Lipofectamine 3000 (Thermo Fisher Scientific, L3000001) following the manufacturer’s protocol. Gene expression was measured 48 hours posttransfection by reverse transcription (RT)–qPCR, normalized to GAPDH using the 2^-ΔΔCt method.

CRISPRa assay

To activate enhancer regions, HaCaT cells were transduced with lentivirus expressing dCas9-VP64 (GeneChem) at an MOI of 50. Cells were cultured for 48 hours postinfection and selected with puromycin (2 μg/ml; Beyotime, ST551) for 7 days, followed by maintenance in puromycin (1 μg/ml). For CRISPRa, sgRNAs were designed to contain MS2 aptamer sequences at the 3′ end, enabling recruitment of the MS2-p65-HSF1 co-activator complex. sgRNA cassettes were cloned into the U6-sgRNA-SV40-MS2-P65-HSF1-T2A-Neo lentiviral vector. Transduction of HaCaT-dCas9-VP64 cells was performed at MOI = 50, followed by G418 selection (1000 μg/ml, Sangon, B540723-0005). Gene expression was measured 48 hours posttransfection by RT-qPCR, normalized to GAPDH using the 2^-ΔΔCt method.

siRNA transfection

HaCaT cells were transfected with siRNAs targeting FOSL1 or a negative control (RiboBio) using Lipofectamine 3000. Cells were harvested 48 hours posttransfection for RNA-seq, ATAC-seq, and CUT&Tag analysis.

Dual-luciferase reporter assay

SE fragments (E1 to E5) from the SE locus (table S8) were cloned upstream of the pGL3-Promoter vector (Promega). Site-directed mutagenesis was introduced to disrupt FOSL1-binding motifs. HaCaT or 293T cells were cotransfected with pRL-TK (internal control), and pcDNA3.1-FOSL1 or control vector using Lipofectamine 3000. Firefly and Renilla luciferase activities were measured using the Dual-Luciferase Reporter Assay System (Promega, e1910). Promoter activity was expressed as the Firefly/Renilla ratio.

Assay for chromatin-bound condensates by exploratory sequencing

ACC-seq was performed as previously described (72) to identify chromatin condensate occupancy. HaCaT cells were subjected to three experimental conditions: native, fixation (fix), and 1,6-Hex pretreatment followed by fixation (1,6-Hex + fix), all processed in parallel to minimize technical variation. For the fixation condition, cells were fixed with 1% formaldehyde at room temperature for 10 min, and the reaction was quenched with 125 mM glycine for 5 min. For the 1,6-Hex + fix condition, cells were pretreated with 5% (w/v) 1,6-Hex (Sigma-Aldrich, #240117) at 37°C for 5 min to disrupt weak hydrophobic interactions associated with phase-separated condensates, before fixation. Cells were then subjected to the same formaldehyde fixation and glycine quenching steps as described above. Native samples were processed in parallel without hexanediol treatment or fixation. Following treatment, cells were immediately subjected to Tn5 transposition and library preparation using the Hyperactive ATAC-Seq Library Prep Kit for Illumina (Vazyme, TD701), according to the manufacturer’s instructions. Libraries were then sequenced and analyzed using a standardized ATAC-seq bioinformatic pipeline, including read preprocessing, alignment to the reference genome, and downstream comparative analysis across conditions to identify chromatin regions associated with condensate-dependent occupancy.

Fluorescence recovery after photobleaching

FRAP assays were performed using a Zeiss LSM710 confocal microscope (63× oil objective) to evaluate FOSL1 condensate dynamics in live cells. Fluorescent foci were bleached with 488-nm laser light (80% power), and recovery was recorded at 1-s intervals for 60 s. Fluorescence intensity over time was quantified using ZEN software to determine the mobile fraction and recovery kinetics, indicating liquid-like phase behavior.

Preparation of mouse epidermal keratinocytes

Epidermal keratinocytes were isolated from dorsal skin of 6- to 8-week-old C57BL/6 mice. For dorsal skin, a 2 cm–by–2 cm area was excised from the back of each mouse. Subcutaneous fat was carefully removed using forceps and scissors. Skin samples were transferred to 1.5-ml tubes containing 1-ml D-Hanks buffer (biosharp, BL559A). Cleaned dorsal skin samples were cut diagonally into four triangular pieces to facilitate efficient enzymatic digestion. Skin pieces were then incubated in 1 ml of Dispase II (Sigma-Aldrich, D4693) working solution in a 12-well plate at 4°C overnight. After digestion, tissues were transferred to cold D-Hanks buffer, and the epidermis was gently separated from the dermis using curved forceps.

For preparation of single-cell suspensions, epidermal sheets were incubated in prewarmed 0.05% trypsin–EDTA (Gibco, 25200-072) at 37°C for 10 min with gentle agitation, followed by an additional 10-min incubation. Enzymatic digestion was terminated by adding trypsin inhibitor solution. Epidermal cells were released by gentle pipetting and passed through a 40-μm cell strainer to remove tissue debris. Cells were washed with cold D-Hanks buffer and centrifuged at 400g for 5 min at 4°C. Cell pellets were resuspended in D-Hanks buffer, and cell viability was assessed by trypan blue exclusion. Only cell suspensions with viability ≥80% were used for CUT&Tag or subsequent downstream assays.

Dead cell removal

Dead cells were removed using the Dead Cell Removal Kit (Miltenyi Biotec, 130-090-101) following the manufacturer’s instructions. Briefly, 1 × 10⁷ total cells were resuspended in 100 μl of Dead Cell Removal MicroBeads and incubated for 15 min at room temperature in the dark. Labeled cells were loaded onto an MS column placed in a MACS separator. Viable unlabeled cells were collected by washing the column four times with binding buffer and used for subsequent experiments.

Tissue fixation and paraffin embedding

Mouse skin samples were fixed in 4% paraformaldehyde (Biosharp, BL539A) for 24 to 48 hours at room temperature, dehydrated through a graded ethanol series (50, 75, 80, 90, 95, and 100%), cleared in xylene (Sinopharm, #10023418), and infiltrated with molten paraffin at 60°C for 1 hour. Paraffin-embedded tissues were sectioned at 5-μm thickness using a rotary microtome. Sections were floated on a 42°C water bath, mounted on glass slides, and dried at 50°C for at least 30 min before staining.

Hematoxylin and eosin staining

Paraffin sections were deparaffinized in xylene and rehydrated through a graded ethanol series to distilled water. Slides were stained with hematoxylin (Servicebio, G1004) for 5 min, rinsed, differentiated, and blued before counterstaining with eosin (Servicebio, G1001) for 5 min. Sections were dehydrated, cleared with xylene, and mounted with neutral resin. Images were captured using a bright-field microscope (Leica Microsystems).

Immunohistochemistry

Tissue sections were deparaffinized, rehydrated, and subjected to antigen retrieval in pH 9.0 EDTA buffer (Servicebio, G1218) using a pressure cooker for 7 min. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide (Beyotime, P0100A) for 15 min, and sections were incubated in 5% BSA for 10 min. Slides were incubated overnight at 4°C with anti–Ki-67 antibody (1:500, Servicebio, GB111499), followed by HRP-conjugated secondary antibody for 1 hour at room temperature. Visualization was performed with DAB substrate (Servicebio, G1212), and nuclei were counterstained with hematoxylin. Slides were dehydrated, cleared, and mounted. Images were obtained using a Leica DM6B upright microscope.

Statistical analysis

Data are presented as means ± SD. Statistical significance was determined through tests denoted in the figure legends, performed by SPSS version 26.0 and GraphPad Prism version 9.0. Before parametric statistical analyses, data were assessed for normality and homogeneity of variance to verify the assumptions required for the applied tests. Statistical analyses were conducted according to experimental design and group comparisons. Unpaired two-tailed Student’s t tests were used for comparisons between two groups. One-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test was applied for comparisons among three or more groups with a single independent variable. Two-way ANOVA with Sidak’s multiple comparisons test was used for experiments involving two independent variables. Spearman’s correlation analysis was performed to assess correlations between continuous variables. P < 0.05 was considered statistically significant.

Supplementary Materials

The PDF file includes:

Figs. S1 to S11

Tables S1 to S8

Legend for supplementary data

Other Supplementary Material for this manuscript includes the following:

Supplementary Data