No predictive value of cytotoxic or helper T-cell precursor frequencies for outcome when analyzed from the graft after stem cell transplantation.

The predictive value of limiting dilution analyses (LDA) measuring cytotoxic and helper T-lymphocyte precursor (CTLp and HTLp) frequencies for outcome after stem cell transplantation (SCT) is still a matter of debate. One reason may be that CTLp and HTLp frequencies are determined in peripheral blood mononuclear cells (PBMC) and this responder cell population does not reflect the cell type composition of the graft. We assessed whether CTLp and HTLp LDA can predict complications after human leukocyte antigen (HLA)-identical SCT when CTLp and HTLp frequencies are analyzed in PBMC of the respective stem cell graft [bone marrow (BMMC) or granulocyte colony-stimulating factor (G-CSF)-mobilized PBMC] and compared to PBMC of PB. Host-specific CTLp frequencies measured in 25 patients and HTLp frequencies analyzed in 6 patients were low in all responder cell sources. CTLp and HTLp frequencies seen against HLA-mismatched unrelated third-party cells were high, but third-party-specific CTLp and HTLp frequencies were lower in G-CSF-PBMC than in PBMC ( p=0.047 for CTLp frequencies). Host-specific CTLp frequencies analyzed in all responder cell sources did not predict acute or chronic graft-versus-host disease (GVHD). Lower CTLp frequencies were detected in all responder cell sources from patients who relapsed after SCT than in patients without relapse, but the differences between both groups were statistically significant only in PBMC. In conclusion, a significant correlation was detected only between relapse and CTLp frequencies measured in PBMC. The lower frequency of third-party-specific cells in G-CSF-PBMC indicates that the mobilization procedure with G-CSF itself may influence results.