Impaired T-cell immune reconstitution is a major complication after allogeneic bone marrow transplantation (BMT) and is particularly exacerbated in the setting of graft-versus-host disease (GVHD). Conventional approaches to reduce GVHD, such as T-cell depletion or pharmacologic immunosuppression, typically fail to enhance T-cell immunity and often further exacerbate this problem. An alternative strategy to mitigate GVHD severity is the selective elimination of graft-versus-host-reactive donor T cells by using an incorporated thymidine kinase suicide gene. This approach has been shown to effectively reduce GVHD, although the effect of this strategy on T-cell reconstitution is unresolved. We addressed this question in a murine BMT model (C57BL/6 [H-2(b)] --> AKR/J [H-2(k)]) in which donor and recipient differ at major and minor histocompatibility antigens. Lethally irradiated AKR recipients transplanted with T cell-depleted bone marrow plus thymidine kinase-positive T cells followed by post-BMT ganciclovir (GCV) administration had more prompt and complete normalization of the T-cell repertoire than phosphate-buffered saline-treated GVHD control animals. By 60 days after transplantation, mice administered GCV had T-cell repertoires that were virtually indistinguishable from those of mice that underwent transplantation with T cell-depleted bone marrow alone (no GVHD controls) when assayed by T-cell receptor (TCR) spectratyping. In contrast, phosphate-buffered saline-treated animals had persistent skewing in most Vbeta families. T cells obtained from GCV-treated mice also had significantly higher in vitro proliferative responses after posttransplantation inoculation with ovalbumin than GVHD animals, indicating that CD4(+) T-cell responses against a nominal antigen were better preserved in these chimeras. Finally, GCV-treated mice had augmented immune reconstitution in response to exogenous interleukin-7 administration, as evidenced by increased overall spleen cellularity and absolute numbers of T and B cells. This was in contrast to GVHD control animals, which had a blunted response to interleukin-7 administration. These data indicate that GVHD severity can be significantly reduced by selective elimination of alloreactive donor T cells without compromise of T-cell immunity. Moreover, in light of previous studies demonstrating that this strategy can reduce GVHD without loss of alloengraftment and antileukemia reactivity, further examination of this approach in humans seems warranted.