HbA₂: at the borderline of the KLF

In this issue of Blood, Perseu et al provide new insights into our understanding of the genetic basis of elevated hemoglobin A₂.¹  This is a major step forward for physicians interpreting hemoglobin electrophoreses of patients with borderline hemoglobin A₂, normal or slightly reduced mean corpuscular volume (MCV), and normal mean corpuscular hemoglobin (MCH).

These patients are a diagnostic dilemma. Does the borderline HbA₂ represent an outlier in the normal population? Is there an inherited pathologic determinant increasing HbA₂? Detailed studies have identified genetic variants in a minority of these patients, including α-globin chain triplication, β-globin gene promoter mutations, and β- or δ-globin gene variants. Should additional diagnostic evaluation, such as family studies and β-globin locus sequencing, be performed?

Perseu and colleagues studied 145 patients with borderline HbA₂, normal or slightly reduced MCV, and normal MCH. They excluded associated variants including mutations of the β-globin or δ-globin gene, the β-globin promoter, and triplicated α-globin genes. Nucleotide sequence analysis of the KLF1 gene in these patients identified mutations in 52 (36%). Variants included nonsense mutations, in/del mutations, and 3 different missense mutations in the second zinc finger of KLF1. There was no influence of XmnI, BCL11A, or HBSIL-MYB HbF-associated polymorphisms on HbF levels in affected patients.

KLF1 (or EKLF: Erythroid Kruppel-Like Factor) is a zinc-finger transcription factor that plays critical roles in erythropoiesis including modifying chromatin architecture, regulating β-like globin gene switching, and activating or repressing gene transcription.^2,3  The association of KLF1 with β-thalassemia is well known. Mutations in the CACCC box of the β-globin gene promoter, a binding site for KLF1, disrupt this interaction and perturb β-globin gene expression. Mutations of KLF1 were first described as the genetic basis of the rare (In)Lu (inhibitor of Lutheran antigen expression) blood group.⁴  Variants included nonsense mutations, in/del mutations, missense mutations in conserved amino acids in zinc finger 1 or 2, and mutation of a GATA-1 binding site in the KLF1 gene promoter. Expression profiling of cultured patient erythroblasts identified a large group of genes with altered mRNA expression including BCAM, which carries the Lutheran antigens, and CD44, which carries the Indian antigens. No other phenotypic abnormalities were described.

Reports of KLF1 mutations associated with nondeletional hereditary persistence of fetal hemoglobin (HPFH) and congenital dyserythropoietic anemia soon followed.^5,6  The HPFH family carried a nonsense mutation, K288X, associated with KLF1 haploinsufficiency and decreased BCL11A in erythroid cells. KLF1 is a direct activator of BCL11A, which represses γ-globin gene expression. KLF1 deficiency releases this repression, leading to elevated HbF.^7,8  CDA patients had a mutation in a conserved residue, E325K, in the second zinc finger of KLF1. Erythroid cells from affected CDA patients exhibited failure of terminal erythroid differentiation, with elevated fetal and embryonic globins, and absent CD44 and AQP1 expression.

A report from Sardinia revealed the rich complexity of gene regulation by KLF1 and its alteration by different mutations.⁹  Erythrocytes from patients heterozygous for 1 of 2 KLF1 mutations, a nonsense mutation S270X or a missense mutation in a conserved residue K332Q, unexpectedly had only minimally elevated HbF, whereas erythrocytes from compound heterozygotes for these mutations had significantly elevated HbF and increased zinc protoporphyrin.

The pleiotrophic effects of KLF1 mutations can be attributed in part to quantitative and qualitative effects on KLF1 structure and function. KLF1 haploinsufficiency appears to be most commonly associated with the In(Lu) phenotype and with mild elevation in HbF. Missense mutations exert their influence by their location and the change they impose on KLF1 protein structure and function. Missense mutations may alter protein-protein or protein-DNA interactions. Study of KLF1 mutants in the second zinc finger revealed different mutations led to differing alterations in KLF1-DNA binding affinity.¹⁰  Clinical severity did not correlate with alterations in binding affinities in all cases. Some variants may alter composition of KLF1-associated protein complexes, affecting transcriptional regulators, chromatin-associated factors, and other important regulatory proteins. Missense mutations could also alter KLF1 by impairing trafficking, altering sites of posttranslational modification, or producing a mutant protein acting as a dominant negative.

Detailed expression and genomics studies indicate that there are many KLF1 targets in erythroid cells, including regulators of cell metabolism, structural membrane proteins, heme synthesis enzymes, and proteins responding to oxidative stress. Different mutations may perturb interactions based on variability in the quality of associated KLF1-DNA binding sites and their interactions. Other potential sources of clinical variability in KLF1 mutant patients include inheritance of other genetic modifiers. In some but not all kindreds with mutant KLF1, variability in HbF has been attributed, in part, to genetic variability at the BCL11A locus. Genetic variability may reside in factors regulated by KLF1, or in factors that are dysregulated when there is altered KLF1 structure or function.

The influence of KLF1 on HbA₂ has not been examined in detail. Cultured erythroid cells from patients with elevated HbA₂ showed an increase in the expression of the absolute amount of δ-globin mRNA relative to β-globin mRNA as differentiation progressed, suggesting a delay in the switch from δ-globin to β-globin.¹  KLF1 or KLF1-regulated factors such as BCL11A could influence δ-globin gene interactions with the LCR and/or with local chromatin-associated proteins. The δ-globin gene could be a target of KLF1, although its promoter CACCC box is degenerate and no other KLF1 binding motifs are nearby.

The mysteries of KLF1 are now beginning to be revealed. When evaluating a patient with HPFH, atypical CDA, borderline HbA₂, In(Lu) blood group, or elevated zinc protoporphyrin, one should consider the presence of a KLF1 mutation. And now, the astute clinician will entertain defects of KLF1 when evaluating patients with otherwise unexplained erythrocyte disorders.