Hereditary spherocytosis (HS) is one of the most common hereditary haemolytic anaemias. HS red cells from both autosound dominant and recessive variants are spectrin-deficient, which correlates with the severity of the disease. Some patients with recessive HS have a mutation in the spectrin alpha-2 domain (S.L.M. et al., unpublished observations), and a few dominant HS patients have an unstable beta-spectrin that is easily oxidized, which damages the protein 4.1 binding site and weakens spectrin-actin interactions. In most patients, however, the cause of spectrin deficiency is unknown. The alpha- and beta-spectrin loci are on chromosomes 1 and 14 respectively. The only other genetic locus for HS is SPH2, on the short arm of chromosome 8 (8p11). This does not correspond to any of the known loci of genes for red cell membrane proteins including protein 4.1 (1p36.2-p34), the anion exchange protein (AE1, band 3; 17q21-qter), glycophorin C (2q14-q21), and beta-actin (7pter-q22). Human erythrocyte ankyrin, which links beta-spectrin to the anion exchange protein, has recently been cloned. We now show that the ankyrin gene maps to chromosome 8p11.2, and that one copy is missing from DNA of two unrelated children with severe HS and heterozygous deletions of chromosome 8 (del(8)(p11-p21.1)). Affected red cells are also ankyrin-deficient. The data suggest that defects or deficiency or ankyrin are responsible for HS at the SPH2 locus.

No abstract

Summary. Nine patients with either b-thalassaemia/haemoglobin E (7) or homozygous b-thalassaemia (2) not requiring regular transfusions were treated with the oral iron chelator, deferiprone 25-50 mg/kg/d for between 17 and 86 weeks (mean 49 weeks). There were significant decreases in serum ferritin (initial mean ± standard deviation 2168 ± 1142, final 418 ± 247 lg/l; t-test for paired samples, P ¼ 0AE005), hepatic iron (initial 20AE3 ± 6AE26, final 11AE7 ± 4AE83 mg/g/dry weight; P ¼ < 0AE02), red cell membrane iron (initial 76AE2 ± 3AE64, final 7AE2 ± 0AE56 mmol/mg protein; P ¼ < 0AE0005) and serum non-transferrin bound iron (initial 9AE0 ± 0AE56, final 5AE9 ± 0AE89 lmol/l; P ¼ < 0AE0005). There was also a significant rise in serum erythropoietin (initial 240 ± 195AE1, final 433AE2 ± 269AE2 U/l; P ¼ 0AE034). The haemoglobin level rose in three patients and transfusion requirements were reduced substantially in four patients. Serum thiobarbituric acid reactive substance (TBARS) also fell in six of eight patients. Patients generally improved clinically, with weight gain observed. Side-effects were mild and included gastrointestinal symptoms (6) and arthralgia (1), not requiring withdrawal of the drug. One patient died at 17 weeks of therapy as a result of an intercurrent infection. His neutrophil count was normal. We conclude that deferiprone is an effective, well-tolerated iron chelator for patients with thalassaemia intermedia. Further studies are needed to determine the optimum dose and length of treatment needed to reduce iron burden to a safe level in these patients.

We identified 22 women with thrombocytopenia of < 100,000/microliters found incidentally during pregnancy and prospectively monitored their platelet count and clinical outcome for a minimum of 6 months postpartum. During the study period, four women became pregnant twice, accounting for a total of 26 pregnancies. The lowest platelet count during pregnancy was 65,600/microliters +/- 19,400 (mean +/- SD), and at delivery 84,500/microliters +/- 32,300 (P < 0.02). The thrombocytopenia was virtually asymptomatic in all patients during the pregnancy and delivery, whether vaginal or surgical. Neonatal platelet counts (n = 18) were normal (270,700/microliters +/- 69,900), and none of the newborns (n = 24) had a bleeding diathesis. Normalization of the platelet count (i.e., > 150,000/microliters) was documented in 18 patients within 1 month postpartum, in five within 3 months postpartum, and in two as late as 5 months after delivery. One woman did not recover from the thrombocytopenia and eventually developed other stigmata of an autoimmune disease. Long-term follow-up showed recurrence of thrombocytopenia in four patients: three in the context of a subsequent pregnancy and one who developed idiopathic thrombocytopenic purpura. Retrospective analysis of blood counts obtained from 12 previous pregnancies demonstrated thrombocytopenia of a similar degree to the index pregnancy. We conclude that gestational thrombocytopenia of < 100,000/microliters is clinically a benign phenomenon that can recur in subsequent pregnancies and is not accompanied by neonatal thrombocytopenia. In some cases, however, pregnancy-associated thrombocytopenia may be a manifestation of an autoimmune disease with its attendant implications for the neonate. Since the differential diagnosis between the two conditions may be difficult to establish when first encountered during pregnancy, a conservative approach emphasizing careful surveillance and guarded reassurance is justified as long as the platelet counts are > 50,000/microliters.

Abnormal deposits of free iron are found on the cytoplasmic surface of red blood cell (RBC) membranes in ␤ -thalassemia. To test the hypothesis that this is of importance to RBC pathobiology, we administered the iron chelator deferiprone (L1) intraperitoneally to ␤ -thalassemic mice for 4 wk and then studied RBC survival and membrane characteristics. L1 therapy decreased membrane free iron by 50% ( P ϭ 0.04) and concomitantly improved oxidation of membrane proteins ( P ϭ 0.007), the proportion of RBC gilded with immunoglobulin ( P ϭ 0.001), RBC potassium content ( P Ͻ 0.001), and mean corpuscular volume ( P Ͻ 0.001). Osmotic gradient ektacytometry confirmed a trend toward improvement of RBC hydration status. As determined by clearance of RBC biotinylated in vivo, RBC survival also was significantly improved in L1-treated mice compared with controls ( P ϭ 0.007). Thus, in vivo therapy with L1 removes pathologic free iron deposits from RBC membranes in murine thalassemia, and causes improvement in membrane function and RBC survival. This result provides in vivo confirmation that abnormal membrane free iron deposits contribute to the pathobiology of thalassemic RBC. ( J. Clin. Invest.

To test the hypothesis that transfusion of blood donated by individuals with glucose-6-phosphate dehydrogenease (G6PD) deficiency may result in a hemolytic reaction, we conducted a prospective longitudinal study in which 10 patients transfused with 1 unit of G6PD-deficient and 1 unit of normal red blood cells (RBC) were compared with 10 patients transfused with 2 units of age-matched normal RBC. We found that 24h after transfusion serum bilirubin (µmol/l) in the recipients of G6PD-deficient RBC was significantly higher than in the recipients of normal RBC (36±14 vs. 18±5, respectively, p>0.004). A parallel increase was found in the serum lactate dehydrogenase (LDH; IU/l) between the two groups (378±151 vs. 264±56, p<0.001). The difference in serum bilirubin (26±10 vs. 15±5, p<0.03) was still noted 48 h after transfusion, with only a marginal difference (p<0.08) in LDH. We conclude that an immediate posttransfusional hemolytic reaction can occur in recipients of G6PD- deficient RBC and therefore suggest that the differential diagnosis of posttransfusional hemolysis, particularly in populations where G6PD deficiency is prevalent, includes transfusion of erythrocytes from G6PD-deficient donors.

Red blood cell (RBC) membranes from patients with the thalassemic and sickle hemoglobinopathies carry abnormal deposits of iron presumed to mediate a variety of oxidative-induced membrane dysfunctions. We hypothesized that the oral iron chelator deferiprone (L1), which has an enhanced capacity to permeate cell membranes, might be useful in chelating these pathologic iron deposits from intact RBCs. We tested this hypothesis in vitro by incubating L1 with RBCs from 15 patients with thalassemia intermedia and 6 patients with sickle cell anemia. We found that removal of RBC membrane free iron by L1 increased both as a function of time of incubation and L1 concentration. Thus, increasing the time of incubation of thalassemic RBCs with 0.5 mmol/L L1 from 0.5 to 6 hours, enhanced removal of their membrane free iron from 18% +/- 9% to 96% +/- 4%. Dose-response studies showed that incubating thalassemic RBC for 2 hours with L1 concentrations ranging from 0.125 to 0.5 mmol/L resulted in removal of membrane free iron from 28% +/- 15% to 68% +/- 11%. Parallel studies with sickle RBCs showed a similar pattern in time and dose responses. Deferoxamine (DFO), on the other hand, was ineffective in chelating membrane free iron from either thalassemic or sickle RBCs regardless of dose (maximum, 0.333 mmol/L) or time of incubation (maximum, 24 hours). In vivo efficacy of L1 was shown in six thalassemic patients whose RBC membrane free iron decreased by 50% +/- 29% following a 2-week course of L1 at a daily dose of 25 mg/kg. As the dose of L1 was increased to 50 mg/kg/d (n = 5), and then to 75 mg/kg/d (n = 4), 67% +/- 14% and 79% +/- 11%, respectively, of their RBC membrane free iron was removed. L1 therapy-- both in vitro and in vivo--also significantly attenuated the malondialdehyde response of thalassemic RBC membranes to in vitro stimulation with peroxide. Remarkably, the heme content of RBC membranes from L1-treated thalassemic patients decreased by 28% +/- 10% during the 3-month study period. These results indicate that L1 can remove pathologic deposits of chelatable iron from thalassemic and sickle RBC membranes, a therapeutic potential not shared by DFO. Furthermore, membrane defects possibly mediated by catalytic iron, such as lipid peroxidation and hemichrome formation, may also be alleviated, at least in part, by L1.

The protein composition of ghosts, inside-out vesicles (IOV), and membrane skeletons (MS) of erythrocytes (RBC) from splenectomized (spx) and nonsplenectomized (non-spx) patients with beta-thalassemia major and beta-thalassemia intermedia was determined. Ghosts from spx thalassemia intermedia patients had a significant increase in their globin content (which was mostly heme reactive) and contained extra polypeptides in the protein 4.2 to 5 and 6-globin areas. The Triton- extracted MS from all of the thalassemic patients showed two major abnormalities: they retained up to twice the amount of protein 3 when compared with controls; they had a significant increase in their globin content, the concentration of which was independent of their protein 3 content. Analysis of the IOV revealed no differences between those prepared from normal controls and those of the patients. MS from spx thalassemia intermedia patients were grossly abnormal when examined by scanning electron microscopy and they exhibited aggregates of material that on transmission electron microscopy suggested the presence of globin precipitates. We propose that, although the integral protein composition, as reflected in the IOV, from severely affected beta- thalassemics is intact, their MS assembly is deranged. The altered skeletal structure of thalassemic RBC could result from attachment of denatured globin to the skeleton components. These abnormalities may contribute to the premature cell death seen in severe beta-thalassemia.