Characterization of the DNA polymorphism rs621940 downstream of GFI1B associated with myeloproliferative neoplasms
Myeloproliferative neoplasms (MPN) are characterized by overproduction of mature hematopoietic cells. Clonal driver mutations in JAK2, CALR or MPL have been identified that are causal to the myeloproliferative phenotype. Several germline single nucleotide polymorphisms (SNPs) may predispose to acquiring MPN such as SNPs in TERT, and the JAK2 46/1 haplotype. Genome wide association studies (GWAS) identified multiple predisposing alleles for JAK2 V617F+ clonal hematopoiesis and Philadelphia chromosome-negative myeloproliferative neoplasms (MPN). An undefined locus downstream of Growth Factor Independence 1B (GFI1B) is significantly associated with predisposition to MPN(4). One of the top-hit SNPs (rs621940) located in this region was located in a potential GFI1B enhancer site. Here, we studied the effect of the locus and the downstream polymorphism rs621940 on the expression of GFI1B and the association of the SNP with granulocytic/monocytic and erythroid development.
Biotin labeled double stranded oligo’s harboring the C- or G- allele of rs621940 were incubated with nuclear extracts of MEG-01 cells to determine differential protein binding. Dual luciferase reporter assays were performed using the enhancer region carrying either the C- or G-allele. To determine the effect of the SNP on allele specific GFI1B expression, DNA and RNA was collected from 30 healthy donors. In addition, colony forming units granulocyte-macrophage (CFU-GM) and blast forming units-erythrocyte (BFU-E) were analyzed on ~200 healthy donors. The presence of the SNP was determined using TaqMan SNP Genotyping assay.
The region encompassing the rs621940 C-allele is bound by GFI1B in MEG-01 cells, as determined by ChIP-sequencing analysis. To determine whether there is differential GFI1B binding to the risk-allele, a pull-down was performed using double stranded biotin-labeled oligo’s. Indeed, differential binding of 24 proteins was observed for either the C- or G-allele, among which several hematopoietic transcription factors (NFIA, NFIX, NFIC, KLF3, SOX4), but GFI1B was not among them. Luciferase reporter assays showed that both alleles do not differentially affect GFI1B promoter activity. By removing the rs621940 locus from K562 cells using CRISPR-Cas9, we observed that this did not affect GFI1B expression, nor GFI1B auto-repression. Allele specific expression was not affected by either SNP allele in mononuclear cells from 30 healthy donors. Furthermore, no significant effect was observed on CFU-GM or BFU-E formation in ~200 healthy donors.
Balanced GFI1B expression plays an important role in the development of MPN. This study could not confirm any effect of the top-hit risk allele rs621940 on the expression of GFI1B in mononuclear cells or CFU-GM or BFU-E formation. Based on these results we concluded that GFI1B expression is not affected by the downstream region in the cell line models studied here, that might explain the contribution to MPN development.