Abstract 203

INTRODUCTION:

Occurrence of a robust Graft versus Leukemia (GvL) effect is at the basis of the curative efficacy of Hematopoietic Stem Cell Transplantation (HSCT). Although leukemia-specific antigens have been characterized, most of the antileukemic potential of transplantation resides in alloreactivity towards patient-specific antigens, such as minor and major histocompatibility antigens. This is particularly relevant in the context of transplantation from related haploidentical and matched unrelated donors (MUD), in which mismatched HLA molecules are potent targets for alloreactive donor T cells. Still, upon in vivo selective pressure by donor T cells, acute myeloid leukemia can undergo genomic rearrangements which result in loss of the patient-specific HLA haplotype, a mechanism which our group recently demonstrated to be frequently responsible for leukemia relapse after haploidentical HSCT (Vago et al. NEJM, 2009).

METHODS:

103 patients who underwent a partially HLA-matched transplantation for Acute Myeloid Leukemia (AML), MyeloDysplastic Syndrome (MDS), or Chronic Myelomonocytic Leukemia (CMML) at the San Raffaele Hospital from 2002 to present, were included in our analysis. For 67 patients, 26 of whom transplanted in complete remission, the stem cell donor was related haploidentical. For the remaining 35 patients, 23 of whom transplanted in complete remission, the donor was unrelated and mismatched for an average of 2/12 HLA alleles. All patients received donor T cells as part of the transplantation protocol. Post-transplantation follow-up comprised monthly bone marrow examination, with Short Tandem Repeat (STR) chimerism analysis and HLA typing performed in parallel on the marrow aspirate samples. The same analyses were performed also for an additional patient, referred to our attention for relapse of AML after two MUD transplantations performed in another center, both from HLA-C and DPB1-mismatched donors, and several donor lymphocyte add-backs. In cases of relapse with suspected loss of the mismatched HLA alleles, STR chimerism and HLA typing were performed also on purified leukemic blasts.

RESULTS:

Disease relapse occurred in 28/67 and 8/35 patients after haploidentical or MUD transplantation performed at our center, respectively. After haploidentical transplantation, 10/28 relapses (35.7%) were due to mutated leukemic blasts which had lost the patient-specific HLA haplotype. Median time to relapse in these patients was 281 days after HSCT (range 67–390), and all patients had received a high dose of donor T cells (median 289×106 CD3+ cells/kg, range 90–583). Upon detection of the mutated leukemic blasts, five of these ten patients were enrolled to receive a subsequent transplantation from a different stem cell donor, mismatched for the remaining HLA haplotype. None of the 8 relapses after MUD transplantations performed at our center displayed loss of the mismatched HLA. However, in the additional patient referred to our attention after two transplantations from partially HLA-mismatched unrelated donors, we documented loss of the HLA haplotype carrying mismatched alleles in the leukemic blasts responsible for leukemia relapse.

CONCLUSIONS:

With respect to our original series of five relapses with leukemic loss of the mismatched HLA haplotype, we describe here five additional cases based on the same mechanism, further consolidating the utmost clinical relevance of this escape mechanism from the GvL effect mediated by haploidentical donor T cells. Prompt detection of the mutant leukemic blasts allowed us to avoid infusion into the patients of donor T cell add-backs, predictably inefficacious in controlling these relapses, and to quickly enroll these patients into a salvage transplant from a different donor, mismatched for the remaining HLA haplotype. Moreover, we demonstrate that even in the case of fewer HLA mismatches, such as in MUD transplants, the genomic rearrangement we described can lead to clinical relapse. Given the constant increase in the number of transplants performed from partially HLA-mismatched donors worldwide, we recommend that post-transplantation HLA typing of bone marrow samples should be routinely performed at disease relapse to detect mutant leukemic blasts, and to guide therapeutic strategies targeted against them.

Disclosures:

Bonini:MolMed S.p.A.: Consultancy.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution