Preformed donor-directed anti-HLA-DP antibodies may be an impediment to successful kidney transplantation

Recent reports have suggested an association of anti-DP antibodies with allograft rejection and graft loss, in both primary kidney transplant recipients and in re-grafts [1–3]. However, in most cases, it is difficult to separate the effects of anti-HLA-DP antibodies from other anti-HLA class I and class II antibodies. Among non-conserved haplotypes, HLA-DP allele mismatches occur despite matching for HLA class I and class II loci, because of the lack of linkage disequilibrium between DP and other HLA loci [4] The patient is a 41-year-old Caucasian female with end-stage renal disease (ESRD) due to systemic lupus erythematosus. The patient had a history of one pregnancy and multiple blood transfusions. She received two prior kidney transplants from deceased donors (DD) in 1986 and 1996. The patient's panel reactive antibody (PRA) was consistently elevated (93% for class I and 77% for class II) and included donor-specific antibodies to class I and class II antigens from her first and second allografts (Appendix 1). In 2005, she received a third kidney from a DD who was matched at the allele level for A, B, Cw, DRB1, DRB3 and DQB1, but was mismatched at HLA-DPB1 (Appendix 1). The flow cytometry crossmatch using the current serum was negative with T cells, but positive with B cells. The B cell incompatibility was not due to auto-antibodies. Moreover, screening for antibodies to MHC-Class I-related Chain A (MICA) was negative. Anti-HLA-DP reactivity was still detectable and was directed against mismatched HLA-DP antigens on the donor graft; some of these antigens were repeat mismatches from the previous transplant (Table 1; Appendix 1). Collectively, these findings suggested that the reactivity in the B cell crossmatch was most likely due to anti-HLA DP antibodies. At the time of her third transplant, the patient received induction therapy with rabbit anti-thymocyte globulin (ATG) and was started on tacrolimus, mycophenolate mofetil (MMF) and prednisone. She was discharged from the hospital with a serum creatinine of 1.7 mg/dl on Day 4 after transplantation. Within 10 days after the transplant, her serum creatinine was found to be elevated at 4.0 mg/dl. Over the next 2 months, she experienced multiple acute cellular and antibody-mediated rejection episodes, with positive C4d immunohistochemistry studies and interstitial haemorrhage. She was treated with a combination of high-dose steroids, plasmapheresis, ATG, OKT3 and rituximab (375 mg/m2 × 4 doses). Her serum creatinine level was 2.8 mg/dl at 30 months after transplantation. The patient is a 50-year-old Caucasian male with ESRD due to type 1 diabetes. He also has a history of HIV infection and has received multiple blood transfusions. The patient's pre-transplant PRA was 89% for HLA class II and was weakly positive for HLA class I antigens. Antibody specificity analysis using flow specificity bead assays demonstrated the presence of antibodies against HLA-DR7, HLA-DR9 and HLA-DP antigens. The patient received a DD kidney transplant from a zero A, B, Cw, DR and DQ mismatched donor. Subsequently, high-resolution typing revealed one DRB1 allele mismatch (Appendix 1). However, anti-DR antibodies to DRB1*0101, the mismatched allele, were ruled out by Luminex SA bead assay. The flow cytometry crossmatch was negative with T cells but positive with B cells. The B cell incompatibility was not due to auto-antibodies or anti-MICA antibodies. Like the first case, these findings suggested that the reactivity in the B cell crossmatch was most likely due to anti-HLA-DP antibodies, which were demonstrated at the time of transplant (Table 1). Anti-HLA-DP antibody specificitiesa NT denotes not tested; W+ denotes weakly positive. The table shows a list of the HLA-DPB1* alleles and underneath it are amino acid residues at positions β55−57 for each listed allele. Shaded boxes are used to indicate the HLA-DPB1* alleles to which the recipient has been exposed. Amino acid residues are given in the one-letter codes. aAnti-HLA-DP antibody specificity was determined using Luminex® single antigen beads as described in the method section. The threshold of mean fluorescence intensity (MFI) for antibody reactivity was determined in our laboratory as follows: MFI < 1000 = negative; 1000–2900 = weakly reactive; >3000 = reactive. The HLA-DPA1*, DPB1* alleles that were surveyed included: DPA1*0201, DPB1*0101; DPA1*0103, DPB1*0201; DPA1*0103, DPB1* 0301; DPA1*0103, DPB1*0401; DPA1*0103, DPB1*0402; DPA1*0201, DPB1*0501; DPA1*0201, DPB1*0901; DPA1*0201, DPB1*1001; DPA1*0201, DPB1*1101; DPA1*0401, DPB1*1301; DPA1*0201, DPB1*1401; DPA1*0201, DPB1*1701 and DPA1*0201, DPB1*1901. In addition, a possible contribution of the HLA-DP α chain to anti-HLA-DP antibody specificity was ruled out based on the examination of the amino acid sequences of the variable regions of the reactive versus non-reactive HLA-DPA1* alleles (data not shown); b For interpretation of HLA-DP specificities, epitopes were assigned based on amino acid sequence variability in the first domain of the various HLA-DPB1* alleles [Table 2]. The designation includes the position of the first amino acid and the polymorphic residues of the epitope; cA weak reactivity was observed but did not reach the cutoff positive value of 1000. Anti-HLA-DP antibody specificitiesa NT denotes not tested; W+ denotes weakly positive. The table shows a list of the HLA-DPB1* alleles and underneath it are amino acid residues at positions β55−57 for each listed allele. Shaded boxes are used to indicate the HLA-DPB1* alleles to which the recipient has been exposed. Amino acid residues are given in the one-letter codes. aAnti-HLA-DP antibody specificity was determined using Luminex® single antigen beads as described in the method section. The threshold of mean fluorescence intensity (MFI) for antibody reactivity was determined in our laboratory as follows: MFI < 1000 = negative; 1000–2900 = weakly reactive; >3000 = reactive. The HLA-DPA1*, DPB1* alleles that were surveyed included: DPA1*0201, DPB1*0101; DPA1*0103, DPB1*0201; DPA1*0103, DPB1* 0301; DPA1*0103, DPB1*0401; DPA1*0103, DPB1*0402; DPA1*0201, DPB1*0501; DPA1*0201, DPB1*0901; DPA1*0201, DPB1*1001; DPA1*0201, DPB1*1101; DPA1*0401, DPB1*1301; DPA1*0201, DPB1*1401; DPA1*0201, DPB1*1701 and DPA1*0201, DPB1*1901. In addition, a possible contribution of the HLA-DP α chain to anti-HLA-DP antibody specificity was ruled out based on the examination of the amino acid sequences of the variable regions of the reactive versus non-reactive HLA-DPA1* alleles (data not shown); b For interpretation of HLA-DP specificities, epitopes were assigned based on amino acid sequence variability in the first domain of the various HLA-DPB1* alleles [Table 2]. The designation includes the position of the first amino acid and the polymorphic residues of the epitope; cA weak reactivity was observed but did not reach the cutoff positive value of 1000. The patient received induction therapy with two doses of basiliximab and was started on tacrolimus, MMF and prednisone. By 12 days after transplant, he had minimal urine output and continued to require dialysis. A kidney biopsy revealed acute cellular and antibody-mediated rejection with positive C4d staining. The patient received ATG and steroid boluses and his serum creatinine started to decrease. His serum creatinine level was 3.2 mg/dl, off dialysis, 7 months after transplantation. Detailed information regarding HLA typing and crossmatching, and anti-HLA antibody and MICA testing is provided in Appendix 2. This study was approved by the Institutional Review Board of the University of Pennsylvania. As shown in Tables 1 and 2, antibody reactivity is not restricted to individual HLA-DPB1* alleles or specificities. Instead, a shared epitope, defined by aspartic residue (D) at codon 57 of the DP beta chain, corresponds exactly to the observed antibody reactivity in case 1. Furthermore, the mismatched DPB1* alleles of the third donor (DPB1*2001), against which the first patient's serum was reactive, were also found to harbor a D at codon 57. (Table 1; Appendix 1). A similar analysis for Case 2 revealed reactivity against HLA-DP alleles that had an E residue at position β57. However, in this case, since several DPB*1 alleles with an E at codon 57 gave weak reactivity, it is possible that other amino acids of the DPB1 beta chain contribute to the antibody binding and specificity (Table 1). Multiple patterns of anti-DP antibody specificities have been previously described, with a predominance of motifs defined by amino acid residues DPβ55-57and DPβ84-87 (Table 2) [5–10]. Moreover, other amino acid mismatches between the donor and the recipient DPβ chains could have potentially contributed to T cell alloreactivity and to the cellular rejection component observed in these two cases. Amino acid sequence alignment of HLA-DPB1* allelesa Amino acid sequence alignment of variability regions within the β1 domain of HLA-DP molecule. The sequences are given in the one-letter code along with their position. aIMGT/HLA amino acid sequence alignment is based on Release 2.16.0 (January 12, 2007). Amino acid sequence alignment of HLA-DPB1* allelesa Amino acid sequence alignment of variability regions within the β1 domain of HLA-DP molecule. The sequences are given in the one-letter code along with their position. aIMGT/HLA amino acid sequence alignment is based on Release 2.16.0 (January 12, 2007). A causal role for HLA-DP could not be unequivocally demonstrated based on these two cases. However, the accelerated rejection was most likely due to a memory alloresponse to a repeat HLA-DP antigen mismatch, since the patients and the donors were matched at the allele level for HLA-A, B, Cw, DR and DQB1, and had preformed donor-specific anti-HLA-DP antibodies. Moreover, in Case 1, the mismatched HLA-DPB1* alleles of both the husband and the second donor (DPB1*0301 and DPB1*0601 respectively) carry the HLA-DPβ57 Asp (D) epitope (Appendix 1). An alloresponse to a minor histocompatibility antigen is less likely. Currently HLA-DP typing and antibody reactivity data are not used for donor allocation, making it difficult to identify and avoid organ incompatibility due to these antibodies. Our report indicates that preformed donor specific anti-HLA-DP antibodies directed towards an epitope in the beta chain can be associated with adverse transplant outcome. Further studies are needed to better define the clinical significance of presensitization to various HLA-DP epitopes in clinical transplantation. Supplementary material is available at NDT Journal online. Conflict of interest statement. The results presented in this paper have not been published previously in whole or in part, except in abstract form.