Abstract
Most spontaneous somatic mutations in humans would be difficult to detect, because they are either silent or lethal, and for autosomal genes, there is a second functional copy that would complement the mutant phenotype. For the detection of spontaneous inactivating mutations, there is a great advantage to studying sentinel genes on the X-chromosome, which is present in one copy in males and one functional copy in females. Using flow cytometry, we previously demonstrated that the PIG-A gene (Xp22.1) can be used as a sentinel for spontaneous mutations in granulocytes and lymphoid cells. However, when this is applied specifically to red cells, in vivo selection against PIG-A mutants is likely to result in an underestimate. We therefore performed an in silico search for other X-linked genes that affect the red cell membrane, and XK (Xp21.1) emerged as a prime candidate. XK encodes a 444 amino acid surface protein, spans the membrane ten times, and is covalently linked to the Kell protein. Inherited inactivating mutations of XK greatly reduce the expression of Kell antigens on the red cell surface, a part of the phenotype known as the McLeod syndrome. In this condition, the hemoglobin level is normal, suggesting that there would not be significant selection against spontaneously arising XK mutants. Of note, a broad spectrum of mutations, including large contiguous chromosomal deletions are known to inactivate XK. To identify rare red cells with a McLeod-like phenotype in normal individuals, we first stained red cells using a mouse antibody specific for a non-polymorphic Kell antigen (K14), followed by a rabbit-anti-mouse immunoglobulin antibody conjugated to PE, and then a FITC-conjugated antibody specific for a non-polymorphic glycophorin A antigen. Cells were gated based on FSC/SSC and expression of glycophorin A. Control red cells from a patient with the McLeod syndrome demonstrated decreased Kell expression, and as expected from random X-chromosome inactivation, a female obligate carrier had two populations, in about a 1:1 ratio, suggesting lack of significant selection against the mutants in vivo. By analyzing a large number of cells by flow cytometry, in 8 adult normal donors, we identified a population of spontaneously arising red cells with a McLeod-like phenotype, at a median frequency of 39 per million (range 26 to 61 per million). This value is somewhat higher than reports by others regarding spontaneous allele loss of M and N red cell antigens in compound heterozygotes (approximately 7–10 per million). In 5 adults, the frequency of red cells with the PIG-A mutant phenotype ranged from 0.8 to 6.3 per million, which we believe is lower due to the decreased half-life of red cells with the PNH phenotype. It was possible to greatly enrich for the McLeod-like red cells in samples from normal donors by bead and column depletion of Kell-positive cells. In some cases, by bead enrichment or sorting of cells from normal donors, we were able to identify cells with membrane projections, which have been reported to occur in a subset of red cells in patients with the McLeod syndrome. We hypothesized that McLeod-like cells in adults are due to somatic mutations in red cell progenitors that accumulate with age. Indeed, in 5 cord blood samples there was a considerably lower frequency of McLeod-like red cells-- a median of 9 per million, (range 6.9 to 16.3 per million, p = 0.004). We believe that this represents the first assay for spontaneous inactivating mutations that can be applied to red cells from any individual and that this will be an ideal tool for studying the relationship between spontaneous mutations, aging, and cancer.
Re-evaluation of the structural integrity of red-cell glycoproteins during aging in vivo and nutrient deprivation
