Assessing donor chimerism level among CD3 T, CD4 T, CD8 T, and NK cells predicts subsequent graft rejection, GVHD, and relapse after allogeneic HCT with nonmyeloablative conditioning

We previously showed that low levels of day-14 CD3 T and NK (CD56) cells donor chimerism predicted graft rejection, whereas high levels of day-28 CD3 T-cell donor chimerism predicted acute graft-versus-host disease (GVHD) after HCT with nonmyeloablative conditioning. Here we investigate whether assessing chimerism levels among CD4 T cells and CD8 T cells, and also the absolute number of lymphocyte subsets of donor and host origins, would lend greater precision to our initial observations. We analyzed data from 157 patients receiving HCT after conditioning with 2 Gy TBI +/− fludarabine as treatment for AML (n= 22), ALL (n= 4), CML (n= 13), CLL (n= 19), MDS (n= 26), MM (n= 24), NHL (n= 30), HD (n= 14), RCC (n= 4), and WASP deficiency (n= 1). Postgrafting immunosuppression included MMF and CSP. A total of 97 patients received grafts from HLA-identical siblings, and 60 patients received grafts from HLA-matched unrelated donors. Lymphocyte subsets were isolated from peripheral blood by flow cytometry on days 14, 28, and 42. The proportion of cells of donor origin (chimerism levels) were assessed by VNTR-PCR and quantified by phosphor imaging. Eighteen patients (11%) had graft rejection. Day-14 donor chimerism levels< 50% among CD3 T (P =.0007), CD4 T (P =.03), and NK cells (P =.003) but not CD8 T cells predicted graft rejection. High absolute numbers of CD3 T (P =.002) and NK cells (P= .002) of host origin on day 14 were each associated with increased risks of graft rejection when treated as continuous linear variables. Grades 2, 3, and 4 acute GVHD were seen in 40%, 9%, and 5% of patients, respectively. High donor chimerism levels on day 14 among CD3 T (P= .02), CD4 T (P =.03), and CD8 T cells (P =.02) but not NK cells were each associated with increased risks of grades 2–4 acute GVHD. High absolute numbers of CD4 T (P =.04) and CD8 T cells (P =.04) of donor origin on days 14–42 were each associated with increased risks of grade 2–4 acute GVHD when treated as continuous linear variables, whereas high donor CD3 T (P= .002), CD8 T (P= .006), and NK cell (P= .002) chimerism levels from days 14–42 were associated with decreased risks of relapse. No statistically significant correlations between absolute numbers of donor cells and risks of relapse were found. These data suggest that assessing CD3, CD4, CD8, and NK cell donor chimerism levels and determining absolute numbers of CD3 and NK cells of host and donor origins are useful for predicting HCT outcomes after nonmyeloablative conditioning.