Genetic heterogeneity of KLF1 deficiency and the pleiotropy of haemoglobin phenotypes

Abstract

Kruppel like factor 1 (KLF1) also known as the ‘Master regulator of erythropoiesis’ is located on the short arm of chromosome 19. KLF1 activates a diverse set of genes that have an important role in the regulation of key pathways such as erythropoiesis, cell membrane and cytoskeleton, apoptosis, heme synthesis & transport, cell cycling, iron procurement and globin chain production. To date over 65 molecular variants have been recorded. Their haematological phenotypes range from the clinically benign In(Lu) type of Lu(a-b-) blood group, variability in the HbA2 levels, congenital dyserythropoietic anaemia (CDA) and in most extreme cases hydrops foetalis secondary to profound anaemia. In 2010 sequencing revealed a nonsense mutation in KLF1 in a large Maltese family with hereditary persistence of foetal haemoglobin (HPFH). The p.Lys288Ter mutation was found to ablate the DNA binding domain of the key erythroid transcription factor. Through the Thalassemia and Molecular Genetics clinic, we identified 5 other families with the same KLF1 p.Lys288Ter truncation mutation but with normal or slightly increased HbF levels and borderline HbA2. The phenotypic variability of HbF was not readily explained by the co-inheritance of know HbF- modulating variants in the HBB, HBS1L-MYB and BCL11A loci. To investigate the phenotypic variability of KLF1 deficiency, we took advantage of primary erythroid cultures of individuals carrying the KLF1 p.Lys288Ter mutation with variable HbF and unrelated Maltese individuals wildtype for this mutation. A combined analysis of whole genome sequencing (WGS), gene expression (RNA-seq), chromatin accessibility assays (ATAC-seq) and promoter activity tests was performed to explore the molecular basis of the heterogeneity in HbF levels displayed by KLF1 p.Lys288Ter individuals. WGS identified two KLF1 promoter mutations, one in cis (-1133C>A) and one in trans (-251C>G) to the truncation mutation. A significant decrease in the luciferase expression was noted in constructs with the -1133C>A and -251C>G mutation when compared to construct with the wildtype promoter. After carrying out EMSA, for the -1133C>A mutation a super-shift in the wildtype type oligo was noticed. This super shift was absent in the mutant oligo. Using regulomeDB it was shown that YY1 transcription factor binds to the wildtype sequence, but it does not bind to the mutant sequence. After combining, the results from the promoter assays, the RNA-sequencing and chromatin accessibility studies, the best explanation from HbF variability in KLF1 deficient subjects is the allelic variation of the wildtype KLF1. A retrospective search in the Malta biobank for undiagnosed subjects with borderline HbA2 and β-thalassaemia heterozygotes was carried out. Four-hundred and twenty-six subjects were collected together with 200 controls. One hundred and thirty-two subjects were pseudothalassemia while 294 were β-thalassemia heterozygotes. More than 55% of pseudothalassemia, had a mutation in the KLF1 gene while KLF1 mutations were present in 40% of the β-thalassemia mutation. Most of the KLF1 variants were closer to the 5’ end of the gene spanning from the promoter region to approximately 60% of exon 2. Three SNPs were found at the 3’ UTR. The four promoter mutations showed a statistically significant difference in the luciferase expression of constructs containing the mutant KLF1 promoter when compared to the wildtype. Furthermore, from a clinical point of view, the data presented, highlighted the importance of KLF1 sequencing in patients with haematological features resembling β-thalassaemia for differential diagnosis of microcytosis and for genetic counselling.