Early Reconstitution of T-Cell Immunity to CMV After HLA-Haploidentical Hematopoietic Stem Cell Transplantation Is a Strong Surrogate Biomarker for Lower Non-Relapse Mortality Rates

HLA-haploidentical hematopoietic stem cell transplantation (haplo-HSCT) is a readily available therapeutic option for patients suffering from high-risk hematological malignancies who lack an HLA-compatible donor or for whom there is no time to find one. To avoid GVHD, haplo-HSCT has been classically performed under conditions of strict T-cell depletion, in the form of CD34-selected (“naked”) grafts. Unfortunately, naked haplo-HSCT is followed by a profound state of immune incompetence, which associates with high non-relapse mortality (NRM) rates, mainly due to opportunistic infections. Over the last decade, different strategies have been developed in order to speed-up immune reconstitution after haplo-HSCT, including the delayed infusion of donor T cells (DLI), the selective depletion of T-cell subsets from the graft or the use of an unmanipulated graft followed by novel strategies of immune suppression.

To analyze in depth the early phenomena of immune reconstitution following haplo-HSCT in order to find early surrogate biomarkers of NRM.

From 2004 to 2010, we prospectively studied multiple parameters of T-cell immune reconstitution in 89 pts treated at our Center with haplo-HSCT. Time-points of analysis were pre-transplant, at day 30, day 90 and day 180 post-transplant. Underlying diseases included: high-risk AML 53 pts (60%), ALL 11 pts (12%), CLL 2 pts (2%), tyrosine kinase inhibitors-resistant CML 4 pts (5%), Hodgkin lymphoma 6 pts (7%), NHL 2 pts (2%), MDS with high International Prognostic Scoring System 7 pts (7%), other 4 pts (4%). Eighteen patients (20%) were given prophylactic suicide gene-modified DLI starting at day 30 post-transplant, while 64 patients (80%) received an unmanipulated graft followed by rapamycin until day 100. Overall, the incidence of grade III-IV acute GvHD was as low as 12%. Chronic extensive GvHD developed in 28% of pts. Compared with historical naked haplo-HSCT, the recovery of T-cell counts was accelerated: at day 90, median CD3⁺ cells were 378 per μL (0–2817), CD4⁺ 127 (0–804), CD8⁺ 173 (0–1922). Higher T-cell counts, however, did not clearly associate with lower NRM rates. After the initial expansion of effector memory cells (CD45RA⁻/CD62L⁻) in patients given DLI (P<0.01) and of central memory cells (CD45RA⁻/CD62⁺) in patients receiving rapamycin (P<0.05), there was a progressive normalization of the memory differentiation phenotype with late appearance of naïve cells (CD45RA⁺/CD62L⁺), which displayed a TCR spectratyping complexity score similar to pre-transplant values. Nevertheless, none of these biomarkers performed enough to be considered for surrogating lower NRM rates. In this series at high risk for CMV reactivation (CMV serostatus: H⁺/D⁺ 68%, H⁺/D^- 27%) the event was observed in 46 pts (52%) and CMV disease in 8 pts (9%), all treated according to guidelines. By using Receiver Operating Characteristics (ROC) curve analysis of CMV-specific IFN-γ ELISPOT results at day 30–90, we found that cut-off values of 1000 spots/mL allowed to discriminate with high specificity (>95%) pts that did not reactivate the virus at later time points. Strikingly, while in pts with <1000 spots/mL, the 2-yrs NRM rate was 32%, in those with >1000 spots/mL, this was 0% (P<0.05). Interestingly, when comparing NRM rates in pts that achieved or not a CD4⁺cell value of 200 per μL, the result was not significant (21% vs 30%, P=0.8).

These results clearly indicate that the early reconstitution of T-cell immunity to CMV after haplo-HSCT does not only protect from subsequent viral reactivation, but also surrogates for lower NRM rates. Moreover, they warrant the investigation of a CMV-specific IFN-γ ELISPOT cut-off value of 1000 spots/mL as a predictive biomarker in larger, multicenter series.

No relevant conflicts of interest to declare.