Krüppeling erythropoiesis: an unexpected broad spectrum of human red blood cell disorders due to KLF1 variants

Abstract

Until recently our approach to analyzing human genetic diseases has been to accurately phenotype patients and sequence the genes known to be associated with those phenotypes; for example, in thalassemia, the globin loci are analyzed. Sequencing has become increasingly accessible, and thus a larger panel of genes can be analyzed and whole exome and/or whole genome sequencing can be used when no variants are found in the candidate genes. By using such approaches in patients with unexplained anemias, we have discovered that a broad range of hitherto unrelated human red cell disorders are caused by variants in KLF1, a master regulator of erythropoiesis, which were previously considered to be extremely rare causes of human genetic disease.

Figure 1

KLF1 target genes and associated clinical phenotypes. KLF1 is a master regulator of ∼700 genes in human erythroid cells involved in a wide variety of molecular processes (blue circles). Deregulated expression of a subset or all of these genes, depending on the KLF1 variant, leads to a diverse array of erythroid phenotypes (gray circles). HbA, adult hemoglobin (α₂β₂); HbA2, adult hemoglobin 2 (α₂δ₂); PK, pyruvate kinase; ZnPP, zinc protoporphyrin.

Figure 2

Functional domains of KLF1 and variants reported in the literature. The KLF1 protein (362 amino acids) contains two N-terminal transactivation domains (TAD1 and TAD2) which are required for it to work as a transcriptional activator. The 3 zinc fingers (ZF1, ZF2, and ZF3) located at the C terminus form the DNA-binding domain, which enables KLF1 to bind to specific sites in the genome, typically CACCC boxes and related GC-rich elements. Residues conserved in all human KLF factors are indicated. The cysteine and histidine residues involved in zinc coordination are highlighted in blue; residues contacting DNA are highlighted in yellow. The arrows point to residues making base-specific contacts thus recognizing KLF1-binding sites in the genome. Variants are color-coded: class 1, green; class 2, blue; class 3, red; class 4, black. See supplemental Table 1 for details and references.

Figure 3

Impact of class 2 or 3 KLF1 variants on clinical severity of hemoglobinopathies. (A) Three major hematologic indices— mean cell volume (MCV), HbA₂, and HbF—in controls and α- or β-thalassemia trait individuals with (wt/var) or without (wt/wt) class 2 or 3 KLF1 variants. Values are mean ± standard deviation, derived from Liu et al and Yu et al. (B) Effects of class 2 or 3 KLF1 variants on globin chain balance in healthy controls (normal) and α- or β-thalassemia patients. Significant recovery of globin chain balance is observed only when β-thalassemia is co-inherited with a class 2 or 3 KLF1 variant. α-like denotes ζ- and α-globin chains, and β-like denotes ε-, γ-, δ-, and β-globin chains. wt, wild type; var, variant.

Figure 4

Flags indicating carrier status of class 2 or 3 KLF1 variants. Automated complete blood count (CBC): mean cell volume (MCV) and mean corpuscular hemoglobin (MCH) are low or at the low end of the normal range. Reticulocyte count (RTC) is high or at the high end of the normal range. ZnPP measured by hematofluorometer is high or at the high end of the normal range, with normal iron stores and absence of lead poisoning. HbA₂ and HbF are measured by high-performance liquid chromatography (HPLC) and are moderately increased or at the high end of the normal range. Lutheran antigens are measured by serology and are decreased but not completely absent; Indian antigen (CD44) is measured by flow cytometry and is reduced with normal expression of glycophorin A (CD235a). Weight of the arrows indicates the likelihood that a class 2 or 3 KLF1 variant is present; this increases when several flags are observed simultaneously.