HSV-TK Engineered Donor Lymphocytes Add-Backs Reduce Late Mortality and Improve Survival of High Risk Acute Leukemia after Haplo-HSCT: Results of a Phase II Multicenter Trial.

The outcome of haplo-SCT is limited by delayed immune reconstitution resulting in a high rate of late mortality and relapse. Here, we report results of a phase II multicenter trial (MM TK007) of early add-backs of donor lymphocytes genetically engineered to express the herpes simplex thymidine kinase (TK) suicide gene after haplo-SCT in inducing immune reconstitution and selective control of GvHD by ganciclovir. Twentysix advanced age pts (median age 51, 17–63) were transplanted for high risk leukemia; disease status at SCT was CR1 (8), CR2 (7), refractory (11). A median of 12.2x10⁶/kg (7.3–16.8) CD34+ selected (Clinimacs) and 1x10⁴/kg (0.8–1.4) CD3+ cells were infused after a myeloablative conditioning. 24/26 pts engrafted with a median time of 14 d (8–21) for ANC >1.0x10⁹/l and 13 d (11–24) for plt >50x10⁹/l. No immune reconstitution and no GvHD were observed in absence of TK-add-back. Sixteen pts received TK-DLI at a median dose of 107/kg with 1st infusion at d +42 and 13 pts obtained CD3+ >100/mcl at a median time of 91 d (61–127) from SCT and 24 d (14–42) from TK-DLI. Transduced cells were documented ex vivo in all pts and represented a median of 48% (10–90) of CD3+ cells. Five pts developed acute GvHD, (grade I to IV) that was always completely abrogated by ganciclovir. In patients in CR at time of SCT who were alive at d +42 and received add-backs of Tk cells, OS rate was 46% at 800 days (intention-to-treat analysis: 38% OS at 800 days post-SCT). Of significance, the cumulative incidence of TRM and relapse showed a 40% probability of mortality with a median time of death of 90 days and last event at day +166. This figure indicate that TK cell add-backs abolish late mortality after CD34+ SCT in adults. In patients in relapse at time of HSCT, a median OS of 201 days was obtained in ITT, with a significant advantage on expected survival without transplantation (60 d) and superior results as compared to haplo EBMT registry (80 d). The 2-year estimation of events of this multicenter phase II study confirm that TK-DLI is an effective tool for promoting immune reconstitution and protecting pts from late infectious mortality after haplo-SCT.

We believe that these results are due to the rapid development of a wide T cell repertoire obtained by TK cell infusions. Immunological follow-up showed Th1/Tc1 effector memory T cells, with a wide TCR repertoire in the first 3 months after SCT in all patients. High frequencies of T cells specific for CMV (median: 35 and 93 spots/10⁵ cells with CMV-infected donor and host fibroblasts) and EBV (median: 58 and 41 spots/10⁵ cells with donor and host EBV-LCL) were detected by gIFN ELISpot at time of immunereconstitution, and correlated with complete control of viral infections. Normalization of the T cell repertoire was documented by spectratype, immune-phenotype for naïve and memory T cell subsets and gIFN ELIspot 6 months after treatment.

A phase III randomized multicentric study will start in 2006.