A GVHD kill switch helps immune reconstitution

Jeffrey J. Molldrem and Sijie Lu

In this issue of Blood, Zhou et al report long-term follow-up and detailed analysis of immune reconstitution associated with a different suicide gene strategy to abrogate graft-versus-host disease (GVHD).1

The use of alternative stem cell sources for allogeneic stem cell transplantation (SCT) such as HLA haplotype-identical donors is increasingly more common, although haploidentical donor SCT is associated with an increased incidence and severity of GVHD compared with SCT with HLA-matched donors. However, delayed and incomplete immune recovery beyond the first year after haploidentical SCT results in more-frequent opportunistic infections,2 which can account for up to 40% of nonrelapse mortality. Thus, therapeutic strategies to reduce GVHD and accelerate immune reconstitution are needed to improve clinical outcome after haploidentical SCT and to extend this potentially curative procedure to more patients suffering from hematological malignancies. A novel strategy that has been tested involves gene modification of donor lymphocytes with a “suicide gene” such as HSV-TK, which can be activated by administration of ganciclovir, selectively eliminating the transduced T cells.2

Although this strategy effectively abrogates GVHD,3,4 the HSV-TK transgene is immunogenic, thus limiting long-term persistence of the transduced T cells, and ganciclovir is required to control viral infections after SCT.

In this study, 10 pediatric patients with acute lymphoblastic leukemia, acute myelogenous leukemia, biphenotypic leukemia, and myelodysplastic syndrome received haploidentical SCT that was CD34-selected to achieve a 4- to 5-log reduction in T cells, and patients received no other GVHD prophylaxis. After initial engraftment, patients received escalating doses of donor-derived T cells that were first depleted of alloreactivity by eliminating CD25+ cells activated during a mixed lymphocyte reaction with recipient-derived Epstein-Barr virus-transformed lymphocytes ex vivo. The allodepleted T cells were then gene modified with a retrovirus encoding an inducible safety switch, caspase 9. Inducible human caspase 9 (iC9) is a hybrid protein consisting of a human FK506-binding protein (FKBP12) linked to a modified human caspase 9 lacking the caspase recruitment domain (CARD). On binding of a synthetic chemical dimerizer such as AP1903 to FKBP12, caspase 9 is activated and leads to apoptosis of the cell.5

Because the report is a pilot study involving a small number of patients, we must be cautious about drawing broad conclusions. Nevertheless, it is an important study that extends the early report by these investigators with iC9-expressing T cells (iC9-T cells) to control GVHD.6 Zhou et al now report the long-term outcome of 10 patients who received infusions of iC9-T cells. They found that iC9-T cells can persist long-term (median, 3.5 years). In addition, circulating iC9-T cells were predominantly CD8+, but they were clonally diverse. Most important, GVHD that occurred after infusion of the iC9-T cells was permanently abrogated after a single dose of the inducer molecule AP1903 in 5 of the patients. Interestingly, although AP1903 induced more than a 90% reduction of circulating iC9-T cells within 2 hours, a small fraction of iC9-T cells eventually reemerged in the peripheral blood, but without recurrent GVHD.

Interestingly, the study also suggests a beneficial effect of iC9-T cells beyond control of clinical GVHD. For example, infusion of iC9-T cells was associated with faster immune recovery of a clonally diverse population of T cells compared with historical results of haploidentical SCT. Importantly, the authors found that iC9-T cell infusion was associated with protective T cell immunity against cytomegalovirus, adenovirus, Epstein-Barr virus, and Aspergillus fumigatus, which are significant causes of mortality and morbidity after SCT.7

Moreover, although the iC9-T cells were primarily central and effector memory T cells, the iC9-negative T cells that expanded in association with iC9-T cells contained a larger number of phenotypically naive cells. Because the study involved pediatric patients, it is possible this rapid recovery of naive T cells would be expected to be more modest in older patients because of reduced thymic function.

It is also intriguing that the authors observed effective antiviral T cell responses that were preserved even after some patients received AP1903. In contrast, antileukemia immunity might not have persisted because leukemia relapsed in 3 of 4 patients treated with AP1903. This raises a question of why antiviral immunity persisted but graft-versus-leukemia (GVL) did not. Potential clues to this question might derive from the observation that although 3 of 4 AP1903-treated patients with acute lymphoblastic leukemia relapsed, a fourth patient with acute myelogenous leukemia has not relapsed after nearly 4 years of follow-up after SCT. Because acute lymphoblastic leukemia is less susceptible to GVL effects compared with acute myelogenous leukemia, it remains possible that GVL immunity was not affected by AP1903 treatment, as the authors also speculated. Future studies will need to address this issue.8

The results of this study are encouraging, and they support continued investigation of iC9-T cells for the treatment of GVHD. Future trials should address whether T cell allodepletion is indeed necessary before infusion of iC9-T cells because the iC9-T cells were so effectively eliminated by AP1903, and excluding this step would enable more widespread application of iC9 as a kill switch for treating GVHD. In addition, this study raises interesting other questions about the mechanism that results in more rapid immune reconstitution associated with iC9-T cells and about how the antiviral immunity is maintained even after the induction of apoptosis of the iC9-T cells.


  • Conflict-of-interest disclosure: The authors declare no competing financial interests.