Proportions of T-Cell CD8+ Subsets Lacking CD28 Expression at Day 30 after Myeloablative Allogeneic Stem Cell Transplantation Are Predictive for Relapse and Chronic GVHD.

The curative potential of allo-SCT for malignancies derives from the progressive reconstitution of the immune system and the development of effective anti-tumor immunity, but GVHD and disease relapse remain considerable obstacles to improvement in overall outcomes. Because in recipients target antigens are persisting, donor-derived T-cell responses may be expected to lead to the accumulation of a sizable proportion of differentiated T-cells, as happens following infection with persisting pathogens. A few cross-sectional studies have pointed to the preponderance of certain memory T-cell subsets associated with chronic GVHD (cGVHD), but the subset identified differed between studies. Inasmuch as qualitative T-cell recovery takes months to years to complete and there is substantial variability in time to development of GVHD or relapse, serial analysis might be more suitable to unveil early changes in T-cell subset composition attributable to transplantation-related events. From October 2003 on, 55 pts who underwent an allo-SCT after myeloablative conditioning were monitored prospectively in terms of clinical post-graft complications, including graft rejection, infections, GVHD and relapse. Blood samples were obtained on days 30±2, 60±3, 90±5, 180±10 and 365±15 post-transplant. Naive (CD45RA+CCR7+), central memory (TCM, CD45RAnegCCR7+), effector memory (TEM, CD45RAnegCCR7neg), and terminally differentiated effector (TTD, CD45RA+CCR7neg) were enumerated within the CD4+ and CD8+ pools, and the percentage of cells coexpressing CD28 was calculated within each eight subsets. The degree of donor-derived T-cell chimerism was assessed by real time PCR (sensitivity ≤ 1%). Median follow-up was 733 d (404–1251). Dynamics of CD4+ and CD8+ naive, TCM, TEM, and TTD were similar between the pts who developed cGVHD (n=15) and those who did not and between pts who relapsed and those who did not. However, costaining to detect CD28 demonstrated contrasting differences between cGVHD and relapse. At day 30, pts who subsequently relapsed (n=17) had elevated percentages of cells keeping CD28 expression within CD8+ T-cell subsets (TCM, p=.001; TCM, p=.021; and TTD, p=.007). Conversely, pts who subsequently developed cGVHD (n=15; only one relapsed) had diminished percentages of CD28+ cells within the two CD8+CCR7+ subsets at day 30 (p=.002 and p=.034, respectively). Loss of CD28 expression is known to be a hallmark of CMV infection but multivariate analysis ruled out, however, a confounding effect of CMV. Adjusted hazard ratios were 0.10 (95% CI, 0.01-0.76; p=.026) and 5.56 (95% CI, 1.16-25.00; p=.032) with CD28neg cells 16.7% of all CD8+ TCM at day 30 for relapse and cGVHD, respectively. Furthermore, pts with relapse had more often mixed chimerism at day 30 while those with cGVHD had more often full-donor chimerism (p=.042 and p=.023, respectively).

CONCLUSION: This prospective study is the first to associate an early contrasting change in CD8+CD28neg T-cells with the risk of relapse and cGVHD after a myeloablative conditioning. Determination at day 30 of the proportions of CD8+ T-cell subsets expressing CD28 and of the level of T-cell chimerism could assist in predicting risk of relapse and cGVHD and help build an algorithm for the management of immunosuppressive treatment.