Background. Suicide gene therapy is a promising approach for the safe exploitation of the graft-versus-leukemia effect. The insertion of Herpes Simplex Virus thymidine kinase confers an inducible suicidal phenotype upon ganciclovir (GCV) administration, thus enabling the selective elimination of T lymphocytes causing graft-versus-host disease (GvHD). Despite clinical and experimental studies substantiating the efficacy of the strategy, protocols to generate genetically modified cells (GMC) has been shown to reduce alloreactivity. The physiological CD4/CD8 ratio is inverted and GMC are enriched for “effector memory” T cells. Co-stimulation through CD28 has been shown to preserve the functional phenotype GMC. XcyteTM Dynabeads®, 4,5 μm anti-CD3 and anti-CD28 coated paramagnetic beads (bCD3/CD28) sustain T cell proliferation and can be used to obtain GMC.

Aim. To in vitro characterize human suicide GMC generated with bCD3/CD28 GMC (XcyteTM Dynabeads®, Xcyte Therapies, Inc.) and to test their ability to engraft and cause GvHD in a xenogeneic mouse model.

Results. bCD3/CD28 (bead to T cell ratio 3:1) are a potent stimulus for cell cycle entry for both CD4+ and CD8+ human T cells. This permits retroviral transduction (SFCMM#3 vector, Molmed SpA) and preservation of CD4/CD8 ratio. GMC generated with bCD3/CD28 are enriched for “central memory” T cells (CD45RA+CCR7+ 34±7%, CD28+CD27+ 67±12%, intracytoplasmic IL-2+ 14±5%, IFN-γ+ 10±3% and perforin+ 7±3%) when compared with GMC generated with anti-CD3 (CD3) alone (CD45RA+CCR7+ 17±4%, CD28+CD27+ 21±5%, intracytoplasmic IL-2+ 5±3%, IFN-γ+ 52±11% and perforin+ 22±4%). bCD3/CD28 GMC resist activation induced cell death (AxV+PI+ 12±3% vs 42±13% for CD3 GMC). When injected i.p. in NOD/SCID mice conditioned with irradiation and anti-NK depleting antibodies bCD3/CD28 GMC engraft with a faster kinetics (human chimerism at 2 weeks 14±7%) than observed for for CD3 GMC (5±2%). In this model, mice injected with unmodified human lymphocytes develop signs of xenogeneic (X-) GvHD (ruffled fur, hunched back, weight loss and finally death with massive accumulation of human T cell in lymphoid organs) by week 5. X-GvHD observed in mice injected with CD3 GMC has a significant slower course with a proportion of mice surviving week 8. X-GvHD caused by bCD3/CD28 GMC kill all the animals by week 7 (p<0,05 vs CD3 GMC). In mice with established X-GvHD caused by GMC treatment with GCV leads to a reduction in circulating GMC and modulates X-GvHD. GCV administration is not able to cure animals suffering from X-GvHD caused by unmodified T lymphocytes.

Conclusions. GMC generated with bCD3/CD28 display a “central memory” functional phenotype and are significantly more efficient than CD3 GMC in causing lethal X-GvHD. GCV administration is able to abrogate X-GvHD caused by GMC. These results validate a tool for the generation of human suicide GMC with high alloreactive potential to be utilized in clinical protocols of adoptive immunotherapy of tumors.

Author notes

Corresponding author