Blood group antigens reveal their maker

For the first time, heterozygous mutations within human EKLF/KLF1 have been identified and shown to alter the expression of blood group antigens.

Subtle variations in certain groups of genes may not lead to a dramatic clinical presentation, but when enriched within human subpopulations, they can serve to distinguish that group of individuals from others. Such is the case for cell surface antigens in red blood cells. For example, the Lutheran (Lu) antigen, which consists of 2 alternatively spliced components (Lutheran and B-CAM) derived from the same gene, is widely expressed in many cell types, but particular geographic regions (eg, southeastern England) contain a higher prevalence of individuals who do not express this blood group antigen. This group is phenotypically categorized as In(Lu)(inhibitor of the Lutheran antigen). Two intriguing properties noted early on were that the deficiency was tissue specific, as it occurred only in the red blood cells of such individuals, and that it behaved genetically as an independently segregating and dominant regulator unlikely to be a repressor. These observations led to a prediction that In(Lu) encodes a red cell–restricted DNA-binding protein that directly or indirectly affects the expression of the Lu antigen.¹  In this issue of Blood, Singleton and colleagues now show this prediction to have been on the mark, as they demonstrate that transcription factor EKLF (KLF1) is the InLu gene.

These authors began by making carefully quantified and extensive array analyses to determine genes whose expression was either down-regulated or up-regulated within In(Lu) human blood cells, compared with those with wild-type Lu⁺. Among those expressed at lower levels in In(Lu) cells were a range of membrane and iron metabolism genes in addition to the α-and β-globins (although not many were affected as dramatically as the Lu gene). This strongly suggested that a “master” erythroid regulator might be affected in the In(Lu) population. The authors then interrogated the exons of 8 likely suspects, including GATA1, FOG1, and MYB, but none of these genes were altered in sequence. However, in 21 of 24 samples (compared with none of 37 controls) the EKLF/KLF1 gene was mutated, with the changes primarily leading to truncation or substitution mutants of the protein (although one also fell within an important Gata site within the EKLF promoter). All were heterozygous; in no case were any homozygous null mutations found, consistent with the lethality that results from EKLF's genetic ablation in the mouse.

Genetic mutations within human EKLF/KLF1, much less any association with a red cell blood group phenotype, had not been previously discovered, even though EKLF was first identified in 1993 and its critical role in erythroid gene expression, most notably in activation of the adult β-globin, has been extensively studied since then.²  As Singleton et al point out, their array data are nicely congruent with recent murine array data that directly compared EKLF-null and wild-type red cell gene-expression patterns. This also explains the classic observations that erythrocyte antigens unlinked to the Lu antigen are also down-regulated in In(Lu) blood samples, and that these red cells exhibit mild morphologic abnormalities.

A number of interesting issues arise from this exciting study, of which I will mention only two. At a molecular level, what is unique about the mechanism of EKLF transcriptional regulation such that genes like In(Lu) are so much more dramatically affected by EKLF haploinsufficiency than others? Is it simply a timing issue as postulated by the authors? At a biological level, might there be any advantage to having a significant decrease in In(Lu) antigen expression? For example, this antigen associates with laminins within the subendothelial extracellular matrix.^3,4  It is tempting to speculate that lower levels of In(Lu) protein leading to less “stickiness” of the red cell might be clinically advantageous in some cases to maintain an unimpeded circulation.