Fig. 1: Identification and characterization of fSNP rs1537371.
From: Post-GWAS functional analysis identifies CUX1 as a regulator of p16INK4a and cellular senescence
a, EMSA using NE isolated from human ECs, showing allele-imbalanced gel shifting on 22 of the 24 candidate fSNPs identified by Reel-seq screening of the CDKN2A/B locus using NE isolated from human PBMCs. Data for EMSA represent n = 3 biologically independent experiments. SNPs in red indicate no allele-imbalanced gel shifting. b, Genomic view of the 200-kb CDKN2A/B region showing the following tracks, ordered from top to bottom based on the ENCODE database. (1) SNP track showing locations of the 24 candidate fSNPs; (2–4) three epigenetic tracks for H3K27ac, H3K4me1 and H3K4me3, known as transcriptional factor-binding sites; (5) DNase I hypersensitivity sites (DNase I HS) in human astrocytes; (6) predicted regulatory elements including promoters (red) and enhancers (gray); (7) annotated genes including p14ARF, p16INK4a, p15INK4b and ANRIL. c, Zoomed-in view of the 4-kb genomic region around fSNP rs1537371, showing the same tracks as above plus the negative result from CUX1 ChIP–seq assay in three human cell lines, GM12878, K562 and MCF-7. d,e, Demonstration of fSNP rs1537371 by EMSA (d) and luciferase reporter assay (e). A, risk allele; C, nonrisk allele; T, very rare allele; RLA, relative luciferase activity. Data for EMSA represent n = 3 biologically independent experiments; data for luciferase reporter assays represent n = 6 biologically independent samples. f, Sequences showing mutations around rs1537371 in three independent CRISPR–cas9 clones (nos. 2, 19 and 56), together with wild-type sequence. CON, wild-type control. g, qPCR showing decreased expression of p16INK4a, one of the potential risk genes in the three mutants. Data for qPCR analysis represent n = 3 biologically independent samples, each performed in duplicate. h, Dot plot of fSNP rs1537371 and p16INK4a mRNA levels showing significantly higher expression of p16INK4a in healthy PBMCs carrying homozygous risk allele A/A versus nonrisk allele C/C (P = 0.047, n = 26). P values were calculated using two-tailed Student’s t-test, and all data are presented as mean ± standard error (s.e.). h, Non-normally distributed data related to quantification of p16INK4a expression are presented as median ± interquartile range, and P values were calculated with the nonparametric Mann–Whitney test for pairwise comparisons.
