Functional testing of thousands of osteoarthritis-associated variants for regulatory activity

Abstract

To date, genome-wide association studies have implicated at least 35 loci in osteoarthritis but, due to linkage disequilibrium, the specific variants underlying these associations and the mechanisms by which they contribute to disease risk have yet to be pinpointed. Here, we functionally test 1,605 single nucleotide variants associated with osteoarthritis for regulatory activity using a massively parallel reporter assay. We identify six single nucleotide polymorphisms (SNPs) with differential regulatory activity between the major and minor alleles. We show that the most significant SNP, rs4730222, exhibits differential nuclear protein binding in electrophoretic mobility shift assays and drives increased expression of an alternative isoform of HBP1 in a heterozygote chondrosarcoma cell line, in a CRISPR-edited osteosarcoma cell line, and in chondrocytes derived from osteoarthritis patients. This study provides a framework for prioritization of GWAS variants and highlights a role of HBP1 and Wnt signaling in osteoarthritis pathogenesis.

Fig. 1

Schematic of massively parallel reporter assay. a For each GWAS-lead SNP, we identified all SNPs in LD with r² > 0.8. Colored lines indicate SNPs in the same LD block. b For all SNPs, we extracted 196 nt of genomic sequence centered at the SNP, and separately synthesized the minor (hollow circle) and major alleles, flanked by common adaptor sequences (cyan and purple). c, d We amplified our library from the array via PCR with primers directed at the common adaptors, in the process appending five nt degenerate barcodes (black lines) and additional sequences homologous to the vector (cyan). We cloned our barcoded library of all major and minor alleles into the STARR-seq vector. Each putative regulatory region is cloned into the 3′-UTR of a reporter gene (cyan) with a minimal promoter (dark blue). e We transfected our library into Saos-2 cells via electroporation. Forty-eight hours post transfection, we extracted RNA and DNA. f We determined the abundance of each allele–barcode combination in the mRNA and DNA population through sequencing. For each allele, we calculated one activity score as the average log₂(RNA/DNA) across all independent measurements

Fig. 2

Pairwise correlations between replicates and identification of alleles with differential regulatory activity. a Pairwise correlation of RNA/DNA ratios between three replicates. The lower left triangle shows pairwise scatter plots. The diagonal provides replicate names and the respective histogram of the RNA/DNA ratios for that replicate. The upper triangle provides Pearson (rho) and Spearman (r) correlation coefficients. b Scatter plot of the major versus minor allele for each SNP. Six SNPs with differential regulatory activity (indicated by colors in the legend) were identified using a Mann–Whitney U Test with an FDR of 10%. Source data are provided as a Source Data file

Fig. 3

Functional validation of rs4730222. a Gene model of HBP1 ensemble isoform ENST00000468410 (major) and ENST00000497535 (alternative). UCSC genome browser zoomed in to the two transcriptional start sites. Rs4730222, within the 5′-UTR of the alternative transcript, is indicated by a vertical gray line traversing the annotation tracks. The four black tracks are H3K27ac performed in human embryonic limb bud at E33, E41, E44, and E47, respectively. The green track is H3K27ac data from chondrocytes derived from cultured bone marrow mesenchymal stem cells. Layered H3K27ac is H3K27ac ChIP-seq (a marker for active enhancers and active promoters) layered from GM12878, H1-hESC, HSMM, HUVEC, K562, NHEK, and NHLF cells. Layered H3K4me3 is H3K4me3 ChIP-seq (marker for active promoters) layered from the same seven ENCODE cell lines. b Allelic expression imbalance in SW1353, a chondrosarcoma cell line heterozygote for rs4730222. Black bars represent the fraction of the minor allele in DNA, and gray bars indicate the fraction of the minor allele in cDNA. n = 3 independent experiments. c Allelic expression imbalance in Saos-2 cells with the minor allele of rs4730222 introduced through CRISPR-mediated HDR. Bars are the same as in Fig. 3B. n = 4 independent experiments. Source data are provided as a Source Data file