Peter D. Issitt

Erythrocyte autoantibodies in paediatric patients with sickle cell disease receiving transfusion therapy: frequency, characteristics and significance

The formation of erythrocyte autoantibodies following transfusion therapy has been described in case reports and small series. To determine the frequency, serological characteristics, and clinical significance of this phenomenon in paediatric patients with sickle cell disease, we analysed the patient database at the Duke University Pediatric Hematology Clinic. We identified children who received multiple erythrocyte transfusions, then reviewed clinical records to identify children who developed erythrocyte autoantibodies in association with transfusions. Among 184 paediatric patients who received multiple erythrocyte transfusions, 14 children (7.6%) developed warm (IgG) erythrocyte autoantibodies. Median transfusion exposure at the time of autoantibody formation was 24 erythrocyte units, range 3–341 units. The autoantibody reacted as a panagglutinin in 11 cases but had anti‐e specificity in three patients. Surface complement also was detected in five patients. Clinically significant haemolysis was documented in four patients, each of whom had both surface IgG and C3 detected. The development of erythrocyte autoantibodies was associated with the presence of erythrocyte alloantibodies. Formation of warm erythrocyte autoantibodies in association with transfusions is not rare in paediatric patients with sickle cell disease. Clinicians should be aware of this complication and recognize that the presence of surface C3 is often associated with significant haemolysis.

Lack of clinical significance of “enzyme-only” red cell alloantibodies

In a retrospective study on samples from 10,000 recently transfused patients, 35 samples were found to contain an antibody that reacted with ficin‐treated red cells but was not demonstrable by low‐ionic‐ strength saline solution and indirect antiglobulin test (LISS‐IAT). In those 35 patients, the specificity of the antibody was such that each patient would have been transfused with antigen‐negative blood had the antibody reacted in LISS‐IAT. Tests on red cells from the units already transfused showed that 19 patients had among them received, by chance, 32 antigen‐positive and 74 antigen‐negative units. The remaining 16 patients had among them received 57 units that were, again by chance, all antigen negative. One patient given antigen‐positive blood suffered a delayed transfusion reaction; in two others the antibodies became LISS‐IAT active after transfusion. However, similar changes to the LISS‐ IAT‐active state were seen with two antibodies of patients given only antigen‐negative blood. Also found in the 10,000 patients were 28 clinically insignificant antibodies, 77 sera in which the antibody was too weak to identify, and 216 autoantibodies that reacted only with ficin‐treated red cells. These data support a belief, generally held in the United States but not necessarily elsewhere, that the use of protease‐treated red cells for routine pretransfusion tests creates far more work than the accrued benefits justify.

Atypical presentation of acute phase, antibody-induced haemolytic anaemia in an infant

Summary. We describe a case of ‘warm’antibody‐induced haemolytic anaemia (WAIHA) in which marked depression of red cell Rh antigen expression resulted in the patient presenting with severe anaemia but a negative direct antiglobulin test (DAT). The serum contained potent IgG Rh antibodies. Unlike two previously reported cases (Koscielak, 1980; Veer et al, 1981) in which the diagnosis of WAIHA was established before the DAT became negative, this patient presented with negative serological findings during his first episode of anaemia. As a result, the serum antibodies appeared to be allo‐ not autoimmune in nature and to be unrelated to the patients anaemia. Confirmation of the autoimmune nature of the Rh antibodies was not possible until nearly 2 years after the first episode of anaemia.

JMH variants: serologic, clinical, and biochemical analyses in two cases

BACKGROUND: JMH is a high‐frequency red cell blood group antigen that resides on a 76‐ to 80‐kDa glycosylphosphatidylinositol‐linked protein also known as CDw108. Antibodies with JMH specificity are often autoimmune and are usually, if not always, clinically benign. Some individuals with JMH‐variant antigen produce alloantibodies to JMH, but little evidence concerning their clinical significance is available. This article reports on two patients who express a JMH‐variant antigen and produced alloanti‐JMH.

Rh haplotypes that make e but not hrB usually make VS.

Background and objectives: The Rh phenotypes hrB– and VS+ are both rare in Whites but more common in Blacks. The high‐incidence antigen hrB is present on most red cells that are e+. The presence of VS on red cells is associated with an aberrant expression of e, often called eS. Materials and methods: Using conventional serologic methods, including a monoclonal anti‐hrB‐like antibody, we studied 65 e+ samples that were apparently hrB–. Results: Of the 65, we found that 59 (91%) were VS+. Recent findings have indicated that in VS+ persons a change from leucine to valine occurs at amino acid 245 of the RHCE‐encoded polypeptide. While this residue is predicted to lie within the red cell membrane bilayer, the change presumably affects alanine 226 (that is present when e is expressed) in such a way that eS is seen. Conclusions: Our findings suggest that the change from e to eS may result in nonexpression or marked depression of expression of hrB that is, perhaps, an epitope of e. While the molecular basis of the hrB– phenotype is not known, it is unlikely that the leucine‐to‐valine change at residue 245, resulting in the aberrant form of e, explains all hrB– samples. First, hrB– VS+ and hrB– VS– samples must differ. Second, some hrB– VS+ samples are C+, some are C–. Presumably diverse molecular bases are involved in hrB– phenotypes.

D, weak D (Du), and partial D: the molecular story unfolds

Molecular cloning of Rh cDNA and subsequent studies of both genomic DNA and cDNA have revealed much about the structures of the Rh genes and the proteins they encode. In 1991, Colin et a1.l reported that individuals with D+ red cells have two Rh genes, RHD and RHCE, while most of those with Dred cells have only one gene, RHCE. RHCE is the name of the gene whose alleles include RHCe, RHcE, RHce, and R H C E most individuals with Dred cells are homozygous forRHce. The RHD and RHCE genes are closely aligned on chromosome one, and each comprises 10 exons. In Dpersons, a lack of D antigen almost always represents deletion of RHD, although rare Dindividuals have one or more detectable but nonfunctional RHD genes. From this information, a better understanding of the weak D (D) and partial D phenotypes has emerged. It has long been known that D may be weakly expressed on red cells; the phenotype involved was named D by Stratton2 in 1946. Most individuals of this phenotype are unable to make alloimmune anti-D, which suggests that their red cells carry all epitopes of D. In contrast, a small subset of persons with D+ red cells are able to make anti-D that reacts with all normal D+ red cells but not with their own or with those of other individuals of the same unusual D+ phen~type.~ Various terms have been used to denote this situation: the most descriptive is that introduced by Tippett: who described these persons as having partial D on their red cells. In partial D phenotypes, one or more epitopes of D are missing, and such persons can produce alloimmune anti-D against the epitope(s) of D that their red cells lack. However, the partial D phenotypes may, but do not necessarily, involve recognizably weakened expression of D. The D phenotype-or, as it is better de~cribed,~ the weak D phenotype-involves a quantitative variation of D. Studies with both polyclonal and monoclonal anti-D have been used to determine the number of D antigen sites on red cells. Normal D+ red cells of the R,r and R,R, phenotypes bear about 10,OOO and 30,000 D antigen sites

Heterogeneity of Anti-U

In a recent report, Mentor and Richards [l] described an anti-U, made by an individual with S-, s-, Ured cells, that reacted in direct tests with S-, s-, U+ (variant) samples. They claimed that the literature does not contain reports of previous anti-U that reacted in this manner; in fact, such antibodies have been documented and well known since 1954. In the sixth edition of the textbook of Race and Sanger [2] and in the third edition of mine [3] it is pointed out that heterogeneity of anti-U was apparent right from the time of discovery of the antibody. In three separate studies on red cells from black donors, the incidences of U-negatives were 1 in 605 [4], 3 in 997 [5] and 7 in 3,081 [6] for percentages of 0.17,0.3 and 0.28, respectively (combined incidence 11 in 4,683 or 0.23%). In three other studies the incidences were 12 in 987 [7], 7 in 500 [8] and 4 in 322 [9], for percentages of 1.21,1.4 and 1.28, respectively (combined incidence 23 in 1,809 or 1.27%). Clearly, two different incidences for the U-negative phenotype had been suggested, each from three large studies in close agreement with each other but not with the other three. Later studies [2,3] explained these findings when it was shown that circa 0.25% of American blacks have S-, sUred cells, while about another 1% have cells of the S-, s-, U+ (variant) phenotype. Tests [7-91 with the anti-U that suggested that the incidence of U-negatives is 1.27% clearly classed both S-, s-, Uand S-, s-, U+ (variant) red cells as Unegative. Those [4-61 that established the incidence as 0.23% clearly classed the U+ (variant) samples as U+. In other words, all the anti-U used in those studies [4-61, reacted in direct tests with U variant samples. These findings have been confirmed many times since then [for references see 31. They also illustrate very clearly why anti-U described by Mentor and Richards [l] reacted with 80% of “U-negative” donor units; the reactive bloods were from S-, sU+ (variant) donors and had been misclassified as “Unegative” in tests with anti-U that do not react directly with such cells. The anti-U described by Mentor and Richards [l] was merely another example of the type of anti-U, used in the earlier studies [4-6] that does. The figure cited is an exact fit. Since 0.23% of bloods from American blacks are S-, s--, U-, while another 1% are S-, s-, U+(variant), tests with the type of anti-U used, against red cells of both phenotypes, would be expected to show an incompatibility rate of 81%. In considering the use of adsorption-elution tests in recognition of the U-negative phenotype, it follows from the data given above that if the type of anti-U that is nonreactive with 1.27% of bloods from blacks is used, some four of every five nonreactive samples found will be U+ (variant) and not U-negative. It is in this setting that adsorptionelution tests are necessary to differentiate the two phenotypes. If the type of anti-U that is nonreactive with only 0.23% of samples from blacks is used, those that are U+ (variant) will be recognized as such in the initial direct tests. The fact that it is difficult to categorize anti-U into one type or the other is the reason that it has been suggested [3] that adsorption-elution studies be performed before any sample is classified as Unegative. Had this been done on the units tested by Mentor and Richards [l], the 80% found incompatible with their patient’s serum would already have been known to be of the U+ (variant) phenotype and the patient’s antibody would have been seen to be no different from anti-U of the type first described 35 years earlier [ 5 ] .

The clinical significance of anti-“Mi(a)”

To the Editor: I read with interest the recent TRANSFUSION article by Broadberry and Lin, but I find that the case reports have done nothing to illustrate the potential clinical significance of antiMi that was mentioned by the authors. It is simply incorrect to infer a causal relationship whenever neonatal jaundice and fetal red cells coated with maternal anti-Mia are found together in the same patient. The same goes for anti-Mi associated with symptomatic dyspnea and chills. In Hong Kong, for example, the incidence of Mi.111 phenotype is 6.28 percent,2 that of anti-Mi among hospital patients is 0.28 percent (presumably similar to the incidence in Taiwan), and that of neonatal jaundice is 50 per~ent-7.~; an infants chance of having all three is 6.28 percent x 0.28 percent x 50 percent, or approximately 8 in 100,000. Each year, there are 80,000 births in Hong Kong, and one would expect to see six or seven cases of neonatal jaundice associated with fetal red cells coated with maternal anti-Mi-if one looks for them. In Taiwan, the number of such cases should be many times more, some of which will incidentally require exchange transfusion. I suspect that some people, after reading such case reports, may take home the message that anti-Mi is a health hazard. Perhaps Chinese all over the world will demand a customized panel of screening cells whenever they are hospitalized. In my opinion, the priorities in the study of anti-Mi are a) evaluation of the in vitro indicators of clinical significance including IgG subtyping, complement activation, monocyte monolayer assay, and chemiluminescence test; b) determination of red cell survival in vivo; and c) carrying out of population surveys on newborns with red cells coated with maternal anti-Mi, to see if they have a higher incidence of neonatal jaundice. If indeed some examples of anti-Mi are proved to be clinically significant, one still has to determine the extent of the problem. If only a small minority of examples of anti-Mi are capable of causing hemolysis and if one must perform tens of thousands of extra antibody screening tests using customized cells just to prevent one single case of neonatal jaundice (which can be easily diagnosed and treated), is it worth the time, money, and effort? CHI S. FENG, FCAP, FRCPA Hong Kong Government Department of Health Institute of Pathology Lek Yuen Health Centre Shatin, hT Hong Kong

RECENT ADVANCES IN THE RH BLOOD GROUP SYSTEM

Recent exciting advances in the biochemistry and molecular genetics of the Rh blood group system have understandably overshadowed concurrent new serological findings. This does not, of course, lessen the value of new findings about the serology of Rh. Indeed, initial serological studies often direct molecular biologists to samples that, when studied by the powerful new techniques, will yield the most information. Accordingly, this review will begin with some brief descriptions of newer serological findings in the system. The most recent publication’ from the ISBT Working Party on Terminology for Red Cell Surface Antigens, lists antigens up to RH51. However, RH25(LW) and RH38(Duclos) have been excluded from the system and the numbers declared obsolete because the encoding genes are not at the RHlocus. The antigens RH13(RhA), RH14(RhB), RH15(RhC), RH16(RhD) and RH24(ET) have also been declared obsolete because definitive antibodies are no longer available for characterization of the antigens. The most recently numbered Rh antigens and some potential newcomers are described below. RH48(JAL). A low incidence antigen thus far seen as a product of haplotypes that encode weak C and e in Whites and weak c and e in Anti-RH48 caused HDN severe enough to require exchange transfusion in an infant in one family and perhaps caused the disease in a second. RH49(STEM). Another low incidence antigen. The authors4 reporting RH49 found it to be present on some but not all RH:1 9(hrs-) and RH:-31 (hP-) bloods. It was suggested that linkage disequilibrium might explain this finding. RHSO(FPTT). Also a low incidence antigen; initially seen5 to be encoded by two different haplotypes that made weak C and weak e, by one that made normal C and weak e, and by those that encode RH33. Later studies6 showed that RH:33 red cells carry RH50 and that RH50 is associated with the partial D phenotype, DFR. Although the RH:33 red cells of persons with the RHar haplotype are RH:50, the partial D of War differs from the partial D phenotype, DFR.’ RH51(MAR). An antigen of very high incidence present on all red cells of “normal” Rh phenotypes except those from individuals homozygous for CW or Cx or heterozygous for CW and Cx.8 Rh deletion (e.g. D-/D-) and Rhnu,, red cells are RH:51. Thus, at the serological level, RH51 behaves as an antithetical antigen to both RH8(CW) and RH9(CX). However, as described in table 3, the Rh polypeptide that carries RH8 and the one that carries RH9 have amino acid residue changes at different po~itions.~ Thus in spite of the antithetical nature of RH51, RH8 and RH9 at the serological level, the description of the encoding genes as alleles depends on one’s definition of that term.

The Rh blood group system: from clumps to clones

This issue of TRANSFUSION contains an elegant and concise review of the genetic basis of the Rh blood group system and an original report describing a carefully conducted study that identified a new blood group antigen that might belong in the Rh system.2 Between them, these two articles illustrate that progress in understanding this highly complex blood group system can still be made at both ends of a widely divergent spectrum of laboratory investigations. The Rh system is the most complex of all known red cell polymorphisms. A recent review3 listed 46 Rh antigens; two 0thers~3~ were added to the system while the review was in press. In the first 40 years after discovery of the Rh system, numerous accomplished biochemists, many of whom achieved considerable success in other areas of immunohematology, attempted to identify the red cell membrane components on which the Rh antigens are carried and to characterize the antigens themselves. In spite of the talents of the investigators, these early studies were rewarded with remarkably few advances. Indeed, it was not until 1982, when simultaneous but independent seminal studies were reported from Edinburgh6 and Helsinki, that the era of biochemical secrecy surrounding the Rh system began to approach an end. The remarkable progress made since 1982 is described by Professor Mollison in his review. Thanks in large part to the research teams in Paris* (sometimes working in collaboration with researchers in Baltimore) and in Bristol? much is now known about the Rh system in terms of the red cell membrane components that contribute to Rh antigen structure and the genes that encode their production. The age-old riddle concerning the usual antithetical relationships between C and c, and between E and e, and the lack of an antithetical partner to D has been answered. Rh-positive people have two Rh genes, one encoding the Ccand Ee-bearing protein or (more likely) proteins and a second encoding the D-bearing protein, while Rh-negative persons have only one Rh gene, the first of the two described above.I0 Methods that led to the recognition of these facts and additional details concerning the current state of knowledge are concisely but comprehensively described by Mollison. One (considerable) difficulty in further characterization of the Rh antigens and the membrane components on which they are canied is that no red cell membrane isolate with the ability to inhibit an Rh antibody has yet been made. Further, the transfer of Rh genes to cell lines that do not normally express Rh antigens has not yet resulted in production of an antigenically active Rh transfectant. These difficulties must surely relate to the fact that, in situ, nonglycosylated Rh polypeptides encoded by genes on chromosome I , Rh glycoproteins encoded from a locus on chromosome 6, the LW glycoprotein encoded by genes on chromosome 19, glycophorin B encoded from a locus on chromosome 4, and probably other membrane components, such as CD47 encoded by a gene on chromosome 3 and the component that carries Fy5, interact to establish Rh antigen specificity. The relative contributions of these components to Rh antigen structure have not yet been completely established, although previous genetic studies,I2 clearly demonstrate that the nonglycosylated Rh polypeptides, encoded from lp36.1-1~34.3,~~ are the major players. At the other end of the spectrum, in terms of laboratory methods used, Coghlan et a1.2 describe a new lowincidence antigen, LOCR. This antigen is clearly associated with Rh at the phenotypic level. In two unrelated families and in one additional unrelated proposita, presence of the LOCR antigen was associated with depression of expression of c. In a fourth unrelated proposita, the LOCR+ red cells had a remarkably depressed expression of e in an Ee heterozygote. The study by Coghlan et a1.2 provides a dramatic example of how carefully performed serologic studies can contribute far more information than more routine phenotyping methods. Tests with well-characterized Rh (or other) antibodies (that have been used over long periods of time in numerous comparative studies), performed by a capillary tube method in which the time required for visible agglutination to occur is measured, provide results that can be interpreted at a quantitative as well as a qualitative level. Expertise in the application of this method has been developed over many years in the Rh laboratory in WinnipegI4-l6; few of the rest of us have acquired the necessary skills. One is reminded that this method, as used in Winnipeg, detected the slight weakening of D on category VII red cells that was overlooked in most laboratories using slide and rapid tube reagents for routine D typing. In spite of the clear association between LOCR and Rh at the phenotypic level, Coghlan et a1. demonstrate admirable restraint and point out that, as yet, the evidence needed to call LOCR an Rh system antigen is incomplete. Similar admirable restraint was shown by Bizot et a1.18 when the low-incidence antigen FPTT was first described. That antigen was initially seen to be associated with Rh:33 red cells and with some phenotypes in which expression of C and/or e was depressed. It was not until the completion of careful family studies to establish the inheritance pattern of FPTT that the lod scores became