We studied late graft rejection in a patient who had received a kidney transplant 9-10 years earlier from his mother and who had been off all immunosuppressive drugs for 7 years at the time of graft rejection onset. The mother differed for one HLA-A (A3) and one HLA-B (B62) antigen but had only a subtype mismatch at the HLA-DR beta1 locus (donor: DR beta1*1104; recipient: DR beta1*1102). A gradual rise in serum creatinine from 1.8 to 2.0 mg/dl at year 3 prompted a biopsy, which was negative fur rejection (focal infiltrates bur no tubulitis). Ton months later the patient's creatinine had risen to > 3.4 mg/dl, and a second biopsy revealed extensive tubulitis, cellular rejection, and glomerular sclerosis. Sonicates of donor leukocytes triggered no delayed-type hypersensitivity (DTH) response above background (PBMC only) in the patient's peripheral blood leukocytes obtained prior to year 9 A gradual recovery of antidonor DTH response between year 3 and 10 closely paralleled the change from tolerant to rejection status. Antidonor antibody was also undetectable in serum prior to year 9, but a donor-re-active antibody did develop at year 10.2 shortly after the peak of DTH response. The serum level of soluble donor HLA class I B62 antigen rose > 10-fold over prerejection level at the time of the biopsy-proven rejection, suggesting a possible trigger For both the cellular and humoral immune response. Nonetheless, we found no evidence for the development of humoral or cellular immunity to maternal HLA class I. Instead, DTH analysis of memory T cells of the patient obtained after rejection showed that a single maternal HLA DR beta1*1104 allopeptide, differing by two amino acids in sequence from the peptide of the recipient (DRP1*1102), stimulated a strong memory DTH response Similarly, are found an anti-HLA class II donor-specific antibody in serum that appeared to be crossreactive with DR beta1*1104 and DR beta1*1101 but not with the recipient DR beta1*1102 antigen. The data support the idea of a profound unresponsive state at both the cellular (DTH) and humoral level toward maternal HLA class I antigens that was not reversed even during late cellular rejection, despite the release of high levels of soluble HLA class I. Furthermore, the data suggest that DTH recovery was a close correlate of the onset of rejection and this "indirect" alloresponse, like the anti-donor alloantibody response chat followed, was directed not to noninherited maternal HLA-A,B antigens but to the maternal HLA DR beta1*1104 subtype. (C) American Society for Histocompatibility and Immunogenetics, 2000. Published by Elsevier Science Inc.