BACKGROUND: Long-term T-cell survival is pivotal for the development of effective therapeutic approaches against pathogens and cancer, since the success of immunotherapy requires the generation of a robust, safe but also durable immune response. Even if it is established that memory cells can survive and persist for years, little is known about the requirements for their long-term persistence. Suicide gene therapy after T-cell depleted haploidentical hematopoietic stem cell transplantation (haplo-HSCT) provides a unique model to study memory T cells. In this setting, patients receive the post-transplant infusion of donor-derived gene-modified memory T lymphocytes retrovirally transduced to express the Herpes Simples Virus Thymidine Kinase (TK) suicide gene and the DLNGFR selection marker. The presence of a safety switch allows the infusion into patients of a broad T-cell repertoire in the absence of immune suppression, while the surface marker enables unambiguous detection and close monitoring of gene-modified cells circulating in treated patients. In the present work we characterize the immunological profile of a cohort of long-term survivors after suicide gene therapy and we studied the fate of persisting TK cells to shed light on memory T cell dynamics in vivo and to unravel the requirements for long-term persistence directly in humans.RESULTS: We studied 10 adult patients who underwent haplo-HSCT and infusion of suicide-gene modified donor T cells (median dose: 1.9x107 cells/kg, range:1-39.5x106) for high-risk hematologic malignancies between 1995 and 2012. Three out of 10 patients (33%) experienced GvHD early after HSCT; in all cases, ganciclovir (GCV) administration proved effective in abrogating the adverse reaction. At a median follow-up of 7 years (range 2-14), all patients were in complete remission and free of GvHD, and displayed a complete and broad donor-derived immune system characterized by physiological counts of NK cells, B lymphocytes, γδ T cells and naïve and memory CD4+ or CD8+ T cells.TK cells were detected in all patients, at low levels (median=4cells/uL), even in patients treated with GCV. Ex vivo selection of pure TK-cells confirmed the presence of functional transduced cells, thus directly demonstrating the ability of memory T cells to persist for years. Importantly, GCV sensitivity was preserved in long-term persisting TK cells, independently from their differentiation phenotype. Longitudinal follow up revealed that TK cells circulated in patients at stable levels and displayed a conserved phenotype comprising effector memory (TEM), central memory (TCM) and stem memory (TSCM) T cells. The low level of Ki-67 positivity suggested the maintenance of a pool of gene-modified memory cells through homeostatic proliferation. Polyclonality was demonstrated by sequencing among TK cells of thousands of diverse TCRs with a broad usage of V and J alpha and beta genes.The number of TK cells persisting at the longest follow-up did not correlate with the amount of infused cells, but instead with the peak of TK cells measured within the first months after infusion, suggesting that antigen recognition is dominant in driving in vivo expansion and persistence of memory T cells. Accordingly, we documented the persistence of CMV and Flu-specific TK cells only after post-transplant CMV reactivation or after Flu infection. Characterization of TK cell products infused to patients showed that the amount of infused TSCM cells positively correlates with early expansion and long-term persistence of gene-marked cells. By combining sorting of memory T-cell subsets with sequencing of integration sites, TCRα and TCRβ clonal markers, we longitudinally traced T-cell clones from infused products to late follow-up time-points. We showed that although T cells retrieved long-term are enriched in clones originally shared in different memory T-cell subsets, dominant long-term clonotypes preferentially originate from infused TSCM clones, suggesting that TSCM might play a privileged role in the generation of a long-lasting immunological memory.CONCLUSION: In a completely restored immune system, suicide gene-modified donor T cells persist for up to 14 years in treated patients. Long-term persistence of memory T cells is determined by antigen exposure, and by the original phenotype of infused cells, according to a hierarchical model in which TSCM are superior to TCM and TEM/EFF. DisclosuresLambiase:MolMed S.p.A: Employment. Traversari:MolMed S.p.A: Employment. Bordignon:MolMed S.p.A: Membership on an entity's Board of Directors or advisory committees. Bonini:MolMed S.p.A: Consultancy.