We report on a 40 year old patient with mild hereditary spherocytosis (RBC: 4.43×1012/dL; Reticulocyte count: 253×109/dL; Hb: 14g/dL), whose red blood cells completely lack protein 4.2. Genetic analysis showed that the patient was a double heterozygote for EPB42 deletions; one allele lacked exon 9 but the sequence remained in frame (protein 4.2 Chartres I) and the other allele contained a di-nucleotide deletion resulting in a premature stop signal (protein 4.2 Chartres II). Homology modelling showed that the hairpin region that forms the proposed band 3 binding site is still present in both mutants. However, the deletion of exon 9 removes a large portion of Domain 2 (core domain) of protein 4.2, potentially removing a band 3 binding groove, and the truncation mutant lacks a portion of the core domain and the whole domains 3 and 4. Therefore, these mutations are likely to destabilize protein 4.2 either directly, or indirectly by disturbing the interaction of protein 4.2 with band 3. Flow cytometry, SDS-PAGE and Western blotting of erythrocyte membranes showed a significant reduction of 70–80 % in CD47 levels, altered Rh associated glycoprotein (RhAG) mobility, reduced GPA/GPB heterodimers, and a 3 fold increase in CD44 levels as reported previously for protein 4.2 null red cells. We stored mature red cells at 4 degrees Celsius over 35 days and found that CD47 continues to be lost in microvesicles as the red cell ages, consistent with a weaker link of CD47 with the cytoskeleton. We investigated band 3 complex stability by performing co-immunoprecipitations and found that lower amounts of band 3 were co-immunoprecipitated using an anti-ankyrin antibody in Chartres red cells compared to wild type, suggesting that the association of band 3 with the cytoskeleton is severely affected. Furthermore, less band 3 was co-immunoprecipitated with an anti-RhAG antibody, consistent with a disturbance of the association of the Rh complex with band 3. We next investigated the stage during erythropoiesis at which the observed changes in band 3 macrocomplex proteins occur. To this end we expanded and differentiated erythroid progenitors from peripheral blood of wild type and the Chartres patient using a three culture system modified from Leberbauer et al. (2005). Synchronous differentiation of a pure erythroid progenitor pool (60% enucleation) demonstrated that protein 4.2 co-immunoprecipitated with band 3 early on in erythroid progenitor differentiation. However, in protein 4.2 Chartres progenitors the mutant forms of protein 4.2 were not expressed at any stage during erythropoiesis, demonstrating that both protein 4.2 mutants are unstable and rapidly degraded. Surprisingly, flow cytometry and western blot analysis revealed that CD47, RhAG, band 3, CD44, and GPA/GPB levels are all similar compared to wild type during erythroid differentiation. Thus, despite the absence of protein 4.2 throughout erythropoiesis, the final changes in the Rh/band3 complex observed in patient’s erythrocytes are not observed. Overall our results suggest that protein 4.2 Chartres is unstable probably due to specific 4.2 mutations that either cause disruption of the band 3 binding sites or an intrinsic instability of these individual mutant proteins. The association of band 3 and ankyrin also appears to be altered in protein 4.2 Chartres suggestive of a weakening of the band 3 cytoskeleton linkage, which could also contribute to the HS phenotype. Importantly, the absence of protein 4.2 not only disturbs ankyrin recruitment to band 3 but also affects association of band 3 with RhAG and disturbs GPA/GPB complexes, which demonstrates the importance of protein 4.2 in the process of band 3 complex formation. Most strikingly, our work demonstrates that the loss of CD47 and the other alterations observed in the band 3/Rh complex in protein 4.2 Chartres must occur late during red blood cell progenitor maturation, presumably after enucleation.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author

Sign in via your Institution