mHAs play a critical role in GVHD and GVL immune responses after allogeneic HSCT. Several mHAs have previously been characterized in humans including four autosomal and eight H-Y T cell epitopes (5
). With only two exceptions (14
), these antigens have been found to be HLA class I–restricted peptides recognized by CD8+
CTLs. Increased numbers of CD8+
CTLs specific for some of these mHAs have been found in patients with active acute GVHD (22
), but the role of mHA-specific CD4+
T cells in the pathogenesis of GVHD has received less attention. To date, only one HLA class II–restricted human mHA has been identified in T cells from a patient with acute GVHD (14
) and another HLA class II–restricted mHA has been identified as a target of hematopoietic graft rejection (17
). Here we describe a new HLA class II–restricted mHA presented by HLA-DRB1*1501 and recognized by CD4+
helper T cells. This epitope is the first T cell mHA associated with chronic GVHD. Like one of the previously identified HLA class II mHAs, this antigen is encoded by the male-specific gene DBY. Interestingly, another class II–restricted antigen encoded by DBY has also been identified in a murine model of GVHD after female to male transplantation (23
). These observations in different species as well as the association with both acute and chronic GVHD suggest that DBY antigens are highly immunogenic to CD4+
female T cells that have not previously been exposed to these antigens during thymic maturation.
By using unstimulated PBMCs in our ELISPOT assay, we unequivocally demonstrated a relatively high level of expansion of DBY-specific T cells in vivo after transplantation. Phenotypic analysis of patient PBMCs 21 mo after HSCT showed that ~25% of PBMCs were CD3+
T cells (not depicted). Because the peptide ELISPOT assay detected ~75 spot-forming cells/106
PBMCs, the frequency of CD4+
T cells specific for a single DBY peptide can be calculated to be ~0.03%. This relatively high frequency is similar to that observed for EBV-specific CD4+
T helper cells (24
). This high level of expansion of DBY-specific CD4+
T cells persisted for more than 1 yr and coincided with persistence of active chronic GVHD that required continuous immune-suppressive therapy. The persistence of this vigorous T cell response despite immune-suppressive therapy also suggests a high level of ongoing antigenic stimulation in vivo and an important role for DBY-specific T cells in the pathogenesis of chronic GVHD in this individual.
Among the 93 peptides representing the complete amino acid sequence of DBY, only 1 peptide was recognized by unstimulated patient cells. Although we cannot exclude the possibility that other DBY epitopes were also targeted, the DBY30–48
epitope appeared to be highly immunodominant in the immune response to DBY. Vogt et al. (14
) have previously identified a different antigenic DBY peptide, DBY176–187
, presented by HLA-DQ5. In their study, the corresponding DBX homologue was not recognized by the DBY-specific T cell clone, and X-Y–disparate residues included in the epitope were shown to be critical for its recognition by donor T cells. DBY30–48
also contains two amino acids that distinguish this peptide from its DBX homologue, but unlike the epitope presented by HLA-DQ5, both DBY30–48
were presented by HLA-DRB1*1501 and recognized by the CD4+
T cell clone. In dose-response experiments, the T cell clone responded equally well to autologous EBV B cells pulsed with either peptide. Similarly, patient PBMCs obtained at different times after transplant also responded similarly to both peptides. Class II–restricted epitopes are usually composed of a 9-mer core peptide flanked by stretches of amino acids on each side (25
). The core peptide interacts directly with the HLA groove and the TCR, whereas flanking regions, sticking out of the HLA groove provide additional stability to the complex. In DBY30–48
, the putative 9-mer core is 100% identical between DBX and DBY homologues. The only disparate residues are located in the flanking regions and these amino acids do not appear to affect the affinity for HLA class II molecules (DRB1*1501) or recognition by the TCR. In our experiments, relatively high concentrations of antigenic peptides were required to trigger response of the 2F9 clone, suggesting that these cells express a low affinity TCR. Remarkably, T cells specific for the previously characterized DQ5-restricted DBY peptide were also dependent on similarly high concentrations of peptide antigens in functional assays (14
). It is currently unknown whether a general trend can be established from these two individual cases.
Polymorphic residues need not be included in the actual T cell epitope to influence the immunogenicity of an antigen. For example, recognition of HA-8 mHA results from the differential processing of an allelic gene product due to polymorphic residues located outside the actual epitope (6
). We considered this possibility for DBY30–48
but we did not find any evidence of differential processing of the DBX and DBY homologous peptides. Importantly, both male and female DCs expressing HLA-DRB1*1501 were recognized at the same level after maturation. In addition, male and female DCs were equally effective in cold target inhibition assays, confirming the endogenous expression and recognition of both the DBY30–48
peptides by mature DCs. In these experiments, the processing of DBY30–48
followed the endogenous pathway of presentation similar to what has previously been described for DBY176–187
). However, the presentation of DBY30–48
required maturation of the DCs with either poly-IC or LPS. Previous studies have demonstrated that antigen processing in DCs is modified during maturation, resulting in the presentation of a new range of cryptic peptide antigens at the cell surface (26
). Our findings suggest a similar method of antigen presentation for both DBY30–48
In addition to DBY-reactive T cells, the patient also developed an antibody response to the same antigen. Using synthetic DBY peptides we mapped the antibody response to two distinct B cell epitopes. In previous experiments we found that antibody responses specific for DBY mHA developed in ~50% of male patients who received stem cell transplants from female donors (19
). A map of the common DBY B cell epitopes was also established, demonstrating that antibody responses were primarily directed against regions of disparity between DBY and DBX. The two epitopes identified in this study are among these frequently targeted epitopes. It is interesting to note, however, that both DBY T cell epitopes characterized thus far (DBY30–48
) were not found to be common B cell targets, presumably because the corresponding portions of the DBY protein are not readily accessible to the B cell receptor. These results further illustrate the complementary function of the two arms of the adaptive immune system to cover a wide range of epitopes within a given antigen.
Antibodies to a single DBY epitope first developed between 6 and 12 mo after HSCT, the date of the first reactive sample. Subsequently, the titer of this antibody response gradually increased until 16 mo after transplant. The acquisition of a new B cell antigen between 16 and 21 mo provided further evidence that B cell immunity to DBY was still evolving at that time. In comparison, the CD4+ helper T cell response, which was first detected 8 mo after transplant, continued to be directed against a single epitope. Also in contrast to the T cell response, the B cell response was specific for DBY and did not cross-react with recombinant DBX in Western blot assays. Antibody specificity for DBY was confirmed at the peptide level using an ELISA assay. Thus, although the patient developed both T and B cell immunity to DBY after HSCT, these results show that it was only the B cell response that distinguished between DBY and DBX.
After allogeneic HSCT, the patient developed acute GVHD, which evolved into chronic GVHD 3–4 mo after transplant. Therefore, our experiments were performed primarily with samples collected in the context of chronic GVHD. In this respect, the dual recognition of DBX and DBY by patient T cells is remarkable as it suggests the development of T cell immunity to an autoantigen. Similarities between chronic GVHD and autoimmune diseases have been well documented. In particular, common skin lesions in chronic GVHD are often similar to those seen in scleroderma or systemic lupus erythematosus. A striking hallmark of autoimmune diseases is the development of antibodies as well as T cells reactive with a variety of autoantigens (27
). Based on our findings, it is attractive to speculate that coordinated B and T cell responses similarly contribute to the pathogenesis of chronic GVHD. Donor specificity for recipient mHA is likely to be required for initiation of acute GVHD, but our studies suggest that the subsequent development of chronic GVHD might be associated with the recognition of T cell epitopes that do not discriminate between recipient and donor cells. Nevertheless, maintenance of chronic GVHD might be dependent on continuing specific recognition of recipient antigens by donor B cells. Although DBY is not expressed on the cell surface, this protein is widely expressed in many tissues and anti-DBY antibodies may contribute to the formation of immune complexes, which can enhance the uptake and presentation of the antigen from dying cells. As noted in our studies, maturation of DCs by nonspecific mechanisms can also lead to the presentation of DBY epitopes to donor T cells in vivo. Taken together, these mechanisms may well be contributing to the persistence of chronic GVHD in the patient we studied. Further studies in other individuals will be needed to confirm whether similarly coordinated B and T cell immune responses are commonly found in patients with chronic GVHD. In this regard, it should also be noted that DBY is only one of several Y-encoded proteins with closely related X homologues that are widely expressed in many tissues. Similar responses to other H-Y antigens may also play a role in the pathogenesis of chronic GVHD.
Previous mHA have been identified using T cells established in vitro and selected for their ability to differentially recognize recipient cells from donor cells in functional assays. Although such approaches have resulted in the characterization of several target antigens of acute GVHD, this method also biases the type of target epitopes that can be identified. In our series of experiments, unselected patient T cells were screened for reactivity against a large panel of potential target peptides. We subsequently isolated the reactive cells in vitro and further analyzed the specificity of the immune response. The dual recognition of DBX and DBY peptides was unexpected and illustrates the validity of our screening strategy for the identification of relevant antigens in GVHD. Although T cells specific for this new DBY epitope persisted at high levels for more than 1 yr, these cells would not have been detected if only clones reactive with recipient cells and not donor cells had been selected to characterize the GVHD response.
In summary, this study represents the first demonstration of a coordinated T cell–B cell response to mHA in the clinical context of chronic GVHD. These studies reveal a complex immune response that includes elements of autoimmunity as well as alloimmunity. The alloimmune response to DBY appears to be mediated primarily by donor B cells. In contrast, the autoimmune response to both DBY and DBX is mediated primarily by donor CD4+ T cells. These observations further emphasize the intricate balance that develops during reconstitution of donor immunity after allogeneic stem cell transplantation when immunologic distinctions between autoantigens and persisting alloantigens become uncertain.