Fig. 3: Identification of a super-enhancer for CXCL chemokines.

a 4 C was performed on LSEC cells with and without TNFα stimulation. Interactions with the CXCL1 promoter were plotted using read counts. A genome region of about 75 kb contained two peaks of CXCL1 interaction under TNFα stimulation (red box). Viewpoint (VP) was labeled with a blackline indicating the location of the reference sequence. b H3K27ac and H3K4me1 ChIP-seq signals of normal (blue) and AH (red) livers were plotted. The black dot indicates the location of the NF-κB site (E1) targeted for subsequent analyses. Scale bar represents 50 kb. c ChIP-qPCR assays for BRD4 and NF-κB binding at the aforementioned locus (black dot). Sequence enrichment was normalized to input. Data represented as mean ± SD. A two-sided, paired t-test was performed on percent input values (n = 8, biologically independent samples from three independent experiments). d, e ROSE algorithm of putative super-enhancer analysis from LSECs without (d) or with (e) TNFα treatment. The region contained in the orange dashed box contained sequences with top H3K27ac enrichment and are considered to be putative super-enhancers. f 3 C experiments were performed on LSECs to detect an interaction of the predicted CXCL super-enhancer with promoters of various CXCLs without (thick line) and with TNFα (thin line). The aforementioned NF-κB binding site (black dot) within the CXCL super-enhancer (dash lines) was used as a reference sequence. Interaction frequencies were plotted after normalized to that of RASSF6, a nearby noninflammatory gene used as control. Multiple other sequences were selected between target CXCL promoters as additional controls. X-axis maps relevant gene sequences as distance (in kb) from RASSF6. Two-way paired ANOVA analysis was performed on the relative interaction frequencies at target sites followed by post hoc Sidak’s multiple comparison correction (n = 4, biologically independent samples from three independent experiments). Data represented as mean ± SD. Treatment with TNFα was found to be a significant variable (p = 0.0034). CXCL1, 2, 3, 6, and 8 loci were found to have increased interaction frequencies with p values as labeled.