Answers are grounded in this paper and cite their sources.
In this issue of TRANSFUSION, Young and colleagues1 describe two patients with autoimmune hemolytic anemia (AIHA) that was suspected to have been associated with previous blood transfusions. In a retrospective study, they found that 5 percent of 2618 patients with a positive DAT had demonstrable RBC autoantibodies and 34 percent of these had a history of the positive DAT developing after transfusions and the subsequent development of alloantibodies. This study draws attention to a phenomenon that has been in the literature for many years but, as the authors point out, seems to have received little attention. It is a phenomenon that I think is very important and touches on some basic immunologic principles that might have broader implications than has been realized. In 1980, Dr Petz and I thought this phenomenon to be of such interest that we devoted a special section on this subject in the first edition of our book, Acquired Immune Hemolytic Anemias;2 the recently published second edition3 contains an enlarged section with many pertinent references. As early as 1918, Rous and Robertson4 showed that rabbits, who had received small transfusions of blood from other rabbits, developed strong cold autoagglutinins. In 1965, Ovary and Speigelman5 confirmed and extended these findings. These authors found that IgG warm autoantibodies sometimes were formed in addition to the IgM cold autoagglutinins. In the same year Zmijewski6 found that chimpanzees developed autoagglutinins after immunization with human RBCs or WBCs. The chimpanzees also developed strongly positive DATs, owing to RBC-bound IgG and complement. Liu and Evans7 described positive DATs in 3 of 12 rabbits, after infections of allogeneic blood. There are several publications2,3,8 relating to the induction of RBC autoantibodies in mice immunized with rat RBCs. The mice developed autoagglutinins (reactive to 37°C) and positive IATs; they also became DAT-positive. Some mice developed AIHA. In 1943, Dameshek and Levine9 reported a case in which successive transfusions resulted in extreme, almost fatal, hemolysis. They suggested that an irreversible autohemolytic process had been established by the multiple transfusions and that this was then intensified by a severe hemolytic transfusion reaction. Polesky and Bove10 described a fatal hemolytic transfusion reaction in a woman with marrow hypoplasia associated with acute leukemia. RBC survival studies, which were in progress at the time, made it possible to demonstrate that a severe autohemolytic crisis was part of the reaction. Fudenberg and coworkers11 described two patients who produced alloantibodies as a result of transfusion and then deliberate immunization to produce a source of typing sera. Both patients developed strongly positive DATs, with autoantibodies as well as alloantibodies present in the serum. They suggested that immunization could arise either de novo at the initial sensitization or as a result of initial immunization to a blood factor in the Rh system, with subsequent widening of the “spectrum of specificity” as further stimulation occurred. Cook12 immunized 34 Rh– male volunteers with Rh+ blood. In addition to the development of the expected anti-D in 68 percent of the volunteers, 11 (32%) men developed an antibody that reacted with Rh– enzyme-treated RBCs; 2 developed a positive DAT that persisted for about 4 months. No evidence of hemolytic anemia was present, although RBC survival studies were not performed. Lalezari and associates13 described a patient who had a partial D phenotype; she had received transfusion of several units of Rh+ blood, and 10 years later received 3 additional units. She developed a high-titer anti-D, and this anamnestic response was associated with the development of an autoantibody. The DAT, which had been negative before transfusion, became positive and remained so for 6 months. An eluate from the patient's RBCs revealed a strong anti-D and a weak “panagglutinin” by the Polybrene technique. Worlledge14 has stated that persistent alloimmunization may lead eventually to autoimmunization and that one sees patients who have had repeated transfusions and who have developed multiple alloantibodies, in whom positive DATs appear even though compatible RBCs have been given. The antibody eluted from the RBCs in such cases often reacted with all RBCs except Rhnull RBCs. The most obvious association of autoantibody formation following transfusion is seen in sickle cell disease (SCD) and thalassemia. Chaplin and Zarkowsky15 de-scribed 4 patients with SCD who developed AIHA; a fifth patient developed a positive DAT without AIHA. In the patients with AIHA, anemia was severe, with reticulocyte counts of 60 to 88 percent. All 5 patients had made alloantibodies, as a result of transfusions, before they made autoantibodies. All patients responded to steroid therapy. After discharge, the DATs became negative. Sosler and coworkers16 described 2 multiply transfused alloimmunized patients with SCD who developed severe AIHA. One patient had made seven, and the other five, alloantibodies before making autoantibodies. Both patients responded to steroids. Argiolu and colleagues17 reported 4 patients with thalassemia major in whom they diagnosed AIHA on the basis of an increase in transfusion requirement associated with the presence of RBC autoantibodies. All of the patients were treated with IV IgG therapy, and 3 responded as indicated by normalization of the blood consumption. Chan and associates18 reported 2 patients who had received transfusions and who later developed severe intravascular hemolysis owing to autoantibodies. Zumberg and coworkers19 described 4 patients who developed autoantibodies in temporal association with alloantibody formation. Two patients developed AIHA. Castellino and associates20 reviewed the medical records of 184 pediatric patients with SCD who had received multiple RBC transfusions. They found that 14 children (7.6%) developed IgG RBC autoantibodies. Median transfusion exposure at the time of autoantibody formation was 24 RBC units (range, 3-341 units). The autoantibody reacted as a panagglutinin in 11 cases but had anti-e specificity in 3 patients. RBC-bound complement also was detected in 5 patients. Clinically significant hemolysis was documented in 4 patients, each of whom had both RBC-bound IgG and C3. The development of RBC autoantibodies was associated with the presence of alloantibodies. The authors concluded that formation of warm autoantibodies in association with transfusions is not rare in pediatric patients with SCD. Aygun and associates21 reviewed the medical records for pediatric and adult SCD patients who had received transfusions over a 10-year period. They found that 8 and 9 percent of pediatric and adult patients, respectively, developed RBC autoantibodies. They noted that partial or extended RBC matching had not been utilized. Singer and coworkers22 evaluated 64 patients who received transfusion with thalassemia and found that 14 (22%) developed RBC alloantibodies, in part owing to RBC phenotypic differences between donors (mostly white) and patients (75% Asian). Sixteen patients (25%) developed a positive DAT and had autoantibodies in their sera. Eleven autoantibodies were IgG panagglutinins, and 5 were IgM antibodies of which 2 had anti-i specificity. Of the 16 autoantibodies, 13 were transient and did not cause significant hemolysis. Three patients developed severe AIHA and required prolonged or recurrent treatment with steroids. Two patients required splenectomy. Autoimmunization was associated with alloimmunization and with the absence of the spleen (44 and 56%, respectively). One of the most interesting findings relating to autoantibody formation following transfusion is contained in reports by Salama and Mueller-Eckhardt 23 and Ness and coworkers.24 These investigators found that in Germany23 and the US,24 it was common for positive DATs associated with delayed hemolytic and serologic transfusion reactions (DHTR and DSTR, respectively) to remain positive for up to 300 days. The antibody eluted from the DAT+ RBCs had a similar specificity as the original alloantibody causing the DHTR and/or DSTR. Sometimes panagglutinins, suggesting autoantibodies, were also demonstrable. These are extraordinary findings because transfused RBCs cannot be present at 300 days. One must assume that the original alloantibody developed autoantibody characteristics, sometimes obvious (i.e., panagglutinin also present), sometimes not. Thus, it seems that transfusion-induced alloantibody formation is commonly associated with autoantibody formation, although AIHA is rare. It has been previously suggested that the same phenomenon may occur in posttransfusion purpura (PTP) where an HPA-A1-negative patient makes an anti-HPA-A1 that eventually reacts with the HPA-A1-negative platelets (PLTs), including the patient's own.23,25 The much shorter t1/2 of PLTs compared to RBCs probably accounts for the difference in clinical severity between the PLT and RBC syndrome. The hypothesis is supported by a PTP marmoset model.26,27 When one species of marmoset was immunized with PLTs from another species of marmoset, thrombocytopenia developed. This was associated with PLT antibody that was detectable in the recipient's serum. This antibody reacted with donor PLTs but did not react, or reacted very weakly, with the recipient's PLTs. Nevertheless, a consistent finding was PLT-associated IgG on the recipient's PLTs. In those animals that recovered from the thrombocytopenia, IgG-coated PLTs were shown to be circulating for 30 to more than 100 days. IgG antibody eluted from these PLTs reacted equally well with PLTs from other marmosets of the same species as the recipient or donor. These findings were in contrast to the finding that the serum antibody would react strongly with PLTs from marmosets of the same species as the donor, but reacted only moderately with PLTs of 75 percent, and weakly with PLTs of 25 percent, of marmosets of the same type as the recipient. The weaker reactions, with 25 percent of the same species, were similar in strength to these observed when autologous PLTs taken from marmosets in remission were tested. This suggested two populations of antibody, one with stronger cross-reactive properties, that were adsorbed by the recipient's PLTs. There are some suggestions in the literature, and in my own experience, that transfusion-induced autoantibodies tend to be transient. It is not clear whether the associated AIHA is more transient than that usually associated with idiopathic AIHA. In an insightful review on autoimmunity in 1965, Dameshek28 mentioned seeing transfusion-induced autoantibodies and suggested that it is actually an “isoimmune reaction which results in a temporary state of ‘pseudo’-autoimmunity. The stimulation of exogenous transfusion is required to set off the reaction: it is not self-perpetuating.” The more modern concept of epitope spreading,29 a process whereby epitopes distinct from and cross-reactive with an inducing epitope become major targets of an ongoing immune response, may apply here. There are immunologic data that are of interest in the association of autoimmunization after transfusion. Kaminski and coworkers30 showed that T lymphocytes from multiply transfused patients develop cytotoxicity against autologous cells; this was not necessarily associated with tissue damage. Paglieroni and associates31 showed that, after transfusions, IgM levels and the absolute number of B cells that coexpress CD5 (i.e., a B-cell subpopulation thought to produce autoantibodies) rose to twice the levels before transfusion. Existing autoantibodies (e.g., rheumatoid factor and antinuclear antibody) increased in four of five patients who received transfusions. Stahl et al.32 showed that, following transfusion, there are alterations in self-reactive antibody repertoires of plasma IgM and IgG that are independent of the presence of a specific immune response to RBC antigens. Phage display technology and human MoAb characteristics are yielding interesting data pertinent to the discussions above. Marks and coworkers33 and Ouwehand and coworkers34 isolated several blood-group-specific antibodies (anti-B, -HI, -D, -E, and -Kpb) from an RNA-derived “naive” phage-displayed V-gene repertoire. RNA was obtained from the mononuclear (MNCs) of two nonimmunized D+ individuals, one group O and one group A. Three of the antibodies (anti-D, -Kpb, and -HI) were made against antigens for which the donors were positive. For example, the RNA-derived B-lymphocyte repertoire from D+ donors contained anti-D reactivity. The serologic reactivity of the anti-D resembled that of a typical “cold” antibody. The affinity for the D antigen was low (kD 5.2 × 107/M), as was expected from a VH gene repertoire derived from µ-positive B lymphocytes. The anti-E (derived from an E– donor) cross-reacted with E– RBCs at 4°C. Ouwehand and colleagues34 believed that these data were in concordance with observations of others that the B-cell repertoire contains a considerable level of self-reactivity. They also believed that their data provided a molecular basis for the serologic cross-reactivity (“mimicking”) characteristics commonly described in AIHA.35 Thompson and colleagues36 studied a panel of 72 human MoAbs (anti-A, -D, -C, -c, -E, -e, -G, -Jka, and -Jkb). Although these antibodies were highly specific for their respective blood group antigens, they were found to react with unrelated foreign antigens and autoantigens (DNA, thyroglobulin, hemoglobin, rabbit and rat tissues, rheumatoid factor). IgM MoAbs showed cross-reactivity more often than did IgG antibodies. Only 5 of 34 IgM antibodies (1 of 13 anti-D, 2 of 7 anti-E, 1 of 13 anti-c, and 1 of 2 anti-A) failed to demonstrate any cross-reactivity. The results clearly demonstrated that a considerable proportion of antibodies, produced in response to exogenous antigen, can show multispecific properties identical to those of natural autoantibodies. Thompson and colleagues36 suggested that multispecificity is a feature of antibodies produced in response to exogenous antigens. Their conclusions were that autoantibodies associated with autoimmune disease may be one of three classes: 1) antibodies that apparently are monospecific for the autoantigen; 2) antibodies that show multispecific properties but whose primary antigen is the autoantigen; and 3) autoantibodies that show multispecific properties and whose autoreactivity is unrelated to the primary antigen (e.g., blood group antibodies). To further complicate the etiology of autoantibodies after transfusion, some old data suggest that transfused RBCs may not be necessary. Mohn and associates37 performed a unique set of experiments that led to the production of autoantibodies in humans but did not involve transfusion of donor RBCs. These investigators transfused volunteers with potent incompatible Rh (anti-CD) antibodies. Approximately 240 mL of antibody-containing plasmas were transfused into three recipients of cDE/cDE, CDe/cDe, and cde/cde phenotypes. Acute hemolytic anemia ensued in the D+ recipients. DATs on the RBCs were positive immediately, remaining so for 168 and 202 days, respectively, for the cDE/cDE and CDe/CDe recipients. Eluates from the RBCs of the cDE/cDE recipient revealed activity for his pretransfusion RBCs and cDe/cde RBCs to the end of the study. A surprising finding was the presence of a specific anti-E in the RBC eluate. This antibody appeared on the 70th day after transfusion and persisted at a high level through the 229-day observation period. The authors concluded that the anti-E was a new autoantibody made by the recipient himself. They hypothesized that the autoantibodies were produced in response to alterations brought about at the Rh sites on the RBC membrane by the antigen-antibody reactions between the D antigen receptors on the recipient's RBCs and the transfused anti-CD alloantibodies, involving the E antigen receptors in the autoimmunization process by reason of their proximity. It has been suggested that perhaps AIHA sometimes occurs because lymphocytes, from previous transfusions, survive and create a “graft vs. host” scenario where the donor lymphocytes produce alloantibody that sensitizes the host's RBCs, thus mimicking an autoantibody.2,3,25 I was stimulated to consider this when finding out that approximately 10 percent of blood donors have been transfused and thus might have “memory lymphocytes” that would recognize foreign RBC antigens (i.e., the recipient). If one adds on the number of female donors who have been exposed to foreign RBC antigens through pregnancy, this could mean that at least one-third of donors may be able to transmit such “memory lymphocytes.” The survival of foreign WBCs, even in immunologically normal people, is now established and has been termed microchimerism.38,39 Strong associations of microchimerism, from pregnancy and from autoimmune disease, have been described.38,39 A publication in 1993 lends support to the suggestion that microchimerism may sometimes be a cause of AIHA. Ishikura and coworkers40 described a patient who received a liver transplant at age 11 months. He received multiple transfusions. On Day 198 the DAT became positive. An autoantibody was eluted from the patient's RBCs. Severe AIHA developed. Allotyping of the patient and donor's IgG were identical, but the RBC eluates showed multiple heavy-chain (Gm) allotypes, indicating that the autoantibody was not produced by the liver donor or the recipient, but probably by transfused lymphocytes. In conclusion, the article by Young et al.1 draws attention to a phenomenon that probably occurs commonly but only manifests obvious clinical ill effects on rare occasions. The phenomenon deserves much more attention because it involves basic immunologic mechanisms that have broad implications.
No takes yet. Share an insight, caveat, or question.
George Garratty (2004) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: