Adoptive Immunotherapy for CMV Infection Complicating Haploidentical Transplants: Challenges Posed by the HLA Restriction of Immuno- Dominant Donor T Cells

T-cell depleted hematopoietic stem cell transplants (HSCT) from HLA haplotypedisparate related donors are being increasingly used for the treatment of patients lacking an HLA-compatible related or unrelated donor. Acute and chronic graft versus host disease can be prevented without added prophylaxis if techniques achieving extensive T-cell depletion are employed. Improved conditioning regimens and transplants providing high doses of CD34⁺ cells have also insured consistent engraftment. However, these patients have a higher and more prolonged risk of serious viral infections than recipients of matched unmodified or T-cell depleted grafts.

Adoptive immunotherapy with donor-derived pathogen-specific T-cells can restore effective resistance in these patients. However, T-cells sensitized and propagated in vitro usually respond to only 1–2 immunodominant epitopes from any antigeneic protein presented by different donor HLA alleles. We questioned how often T-cells generated from an HLA-haplotype disparate donor would be restricted by donor-unique, rather than shared HLA alleles since such T-cells would be inactive against virus-infected cells of a haplotype disparate host. Accordingly, we tested 8 cytomegalovirus (CMV) seropositive HLA haplotype disparate parent or sibling donors, as well as 4 patients, each transplanted from one of these donors 1 or more years previously. In each case, T-cells were sensitized with autologous DCs loaded with a pool of overlapping 15-mers spanning the sequence of CMVpp65, an immunodominant protein of cytomegalovirus, for 21–28 days. Using a matrix of peptide subpools, we then mapped the epitopes eliciting IFN_γ⁺ T-cell responses and then determined their HLA-restriction by assaying their specific cytotoxicity against a panel of EBV B-cell lines loaded with the immunogenic epitope each sharing a single HLA allele with the donor. Of the 8 donors tested, 6 responded to CMV pp65 epitopes presented by HLA alleles shared with the haplotype matched recipient, while 2 donors responded exclusively to epitopes presented by donor unique HLA alleles. Recipients of transplants from the latter two donors experienced prolonged CMV viremia, one with CMV chorioretinitis, requiring months of antiviral treatment. CMV-specific T-cells isolated from each of the four transplanted patients over 1 year post transplant were specific for epitopes presented by HLA alleles shared by their donors. In the patients transplanted from donors whose CMV-pp65 specific T-cells were restricted by donor-unique HLA alleles, the epitope targeted and HLA restriction of the engrafted T-cell differed from that of the transplant donors own T-cells, suggesting that these T-cells may have developed from donor precursors differentiating in the host thymus.

Thus, in a significant proportion of cases (2/8 in this study), virus-specific T-cells generated in vitro from HLA haplotype disparate donors may be restricted by donor-unique HLA alleles and, as a consequence, inactive against virus-infected host cells. For such patients, techniques insuring selective generation of T-cells specific for epitopes presented by shared alleles should be employed. Our findings may also in part explain the prolonged susceptibility of these patients to viral infection. Since immunodominant T-cells reactive against the same epitopes are also those detected at highest frequency in the donor’s blood, they would likely be the only virus-reactive T-cells transferred in a T-cell depleted graft. As a consequence, effective CMV-specific responses might be delayed until new T cells are generated from precursors developing in the host thymus.