In 3 of 15 consecutive patients receiving a human leukocyte antigen (HLA)-identical but major ABO incompatible bone marrow transplant (BMT), pure red cell aplasia (PRA) lasting 5 to 8 months was observed. Titers of the incompatible anti-A agglutinin before infusion of the red blood cell (RBC)-depleted BMT was very high in one, and in the usual range in two patients. Decrease of agglutinin titers during the first 4 weeks after BMT were comparable between PRA patients and those of ABO- incompatible BMT recipients with timely RBC recovery. However, in PRA patients, agglutinin titers rose again and remained elevated for 19 to 28 weeks. RBC engraftment and reticulocyte recovery ultimately occurred spontaneously and coincided with the decrease of agglutinin titers below 16. These observations indicate that PRA is antibody-dependent in this setting. Furthermore, it is conceivable that cyclosporine facilitates recipient-derived antibody synthesis after major ABO- incompatible BMT.

Some twenty cases of dispermic chimeras with the karyotype 46,XX/46,XY, discovered because of gonadal dysplasias or a true hermaphroditism, have been reported. This is a report of a phenotypically normal man with 46,XX/46,XY chimerism in whom a prepubertal finding of positive X-chromatin was interpreted as Klinefelter syndrome. The diagnosis was revised 11 years later when the family doctor, who doubted the earlier diagnosis because of the patient's normal-sized testes, sent him to an outpatient clinic. The young man was 23 years old, athletic (74kg, 180cm), with normal body proportions, normal sexual hair distribution, normal libido and potency, normal endocrine parameters, and a normal spermiogram. The karyotype revealed an XX/XY mosaic in a proportion of 1:2. An identical set of maternal markers (Q- and C-banding) was present in male and female cells. Differences were found with respect to two paternal markers. Furthermore, blood, serum, and red cell enzyme groups in five systems showed two phenotypes, again with duality of paternal origin. It is concluded that a positive X-chromatin in prepuperty, especially in the absence of supporting clinical features, must be followed by a karyotype study.

June 13, 1%4 NATURE 1123 body to anti-Yta, the predicted anti-Ytb. This possibility was supported by the frequency of positive reactions. The specificity of the antibody was established by testing six known Yt(a-) samples, kindly supplied by Dr. M. M. Pickles, Dr. P. A. Tippett and Miss M. J. Polley, with an eluate containing only the unknown antibody in Mrs. B.'s serum. All were found to react positively. A family investigation of the original Fy(b-) Yt(b +) donor showed that her father and uncle were, in fact, Yt(a-). Of the 19 random positive samples, 18 reacted with anti-Yta, but one was found to be Yt(a-). Family investigations and random testing are now being undertaken in Zurich and London and will be reported in detail in due course. Already one family has shown independence of the Yt blood group system from Kell, Duffy and Kidd systems. Eaton et al. have already shown independence ofYt from ABO Rh and MNS systems and Allen et al. 3 have described a family showing independence from the Lutheran system.

In 3 of 15 consecutive patients receiving a human leukocyte antigen (HLA)-identical but major ABO incompatible bone marrow transplant (BMT), pure red cell aplasia (PRA) lasting 5 to 8 months was observed. Titers of the incompatible anti-A agglutinin before infusion of the red blood cell (RBC)-depleted BMT was very high in one, and in the usual range in two patients. Decrease of agglutinin titers during the first 4 weeks after BMT were comparable between PRA patients and those of ABO- incompatible BMT recipients with timely RBC recovery. However, in PRA patients, agglutinin titers rose again and remained elevated for 19 to 28 weeks. RBC engraftment and reticulocyte recovery ultimately occurred spontaneously and coincided with the decrease of agglutinin titers below 16. These observations indicate that PRA is antibody-dependent in this setting. Furthermore, it is conceivable that cyclosporine facilitates recipient-derived antibody synthesis after major ABO- incompatible BMT.

M and N are the two common ("normal") alleles at the MN locus of the MNSs blood group system. The antigens M and N that they determine are located within the amino-terminal region of glycophorin A. In the serologically active and glycosylated (*) fragment of glycophorin AN the sequence is Leu-Ser*-Thr*-Thr*-Glu-, and in that of glycophorin AM it is Ser-Ser*-Thr*-Thr*-Gly-. Mg and Mc are very rare ("variant") alleles of M and N; as to the corresponding antigens, Mg is serologically quite distinct from M and N, while Mc is a compound of both. Erythrocytes of genotypes MgN, MgM, Mg~g, and MCM, which were the object ofthe present study, contain normal amounts ofglycophorin A in their membrane. In glycophorin AM" the amino-terminal sequence is related to that of glycophorin AN by substitution of asparagine for threonine in position 4, and it is nonglycosylated: LeuSer-Thr-Asn-Glu-. The corresponding structure of glycophorin AM, is Ser-Ser*-Thr*-Thr*-Glu-it is thus closely related to that of glycophorin AN and AM, by substitution of the amino acids in positions 1 or 5, respectively. All ofthese substitutions can be explained by single base changes. The distinctions in chemical structure not only confirm the location of M and N in this region of glycophorin A, because they are the only differences observed, but also indicate, because they are correlated with the distinctions in antigenic specificity, that M andN are structural genes coding for amino acid sequences. The finding that Mc contains structural features ofboth M and N suggests that these two forms of glycophorin A have evolved from a common ancestral gene by single base substitutions at sites in the genome coding for amino acids in positions 1 and 5 of the sequence. Carbohydrate structures, however, are also necessary for full expression of antigens M and N. Glycosylation during biosynthesis of residues within the polypeptide appears to depend on a particular protein structure.The MN blood group antigens were described over 50 years ago by Landsteiner and Levine (1), who immunized rabbits with human erythrocytes in a deliberate attempt to find antigenic differences independent ofthe ABO blood groups. It has been recognized since that these antigens are situated on the integral membrane protein glycophorin A (2-4). The complex antigenic structure is located within the amino-terminal five amino acids ofthe sequence, three ofthese residues being glycosylated. The amino acid sequence of glycophorin A is different in positions 1 and 5 for molecules isolated from erythrocyte membranes ofindividuals homozygous for M or N, but no differences in carbohydrate structure or sites ofglycosylation (*) are apparent for the two forms (5, §): AN Leu-Ser-Thr-Thr-Glu- Glycophorin A was isolated from 100-400 ml of blood by methods outlined elsewhere (8). Briefly, after preparation of hemoglobin-free membranes the lyophilized ghosts were extracted with lithium diiodosalicylate and the sialoglycoprotein fraction was recovered from the aqueous phase after phenol partitioni...