Anti-tumor cellular immunotherapies that implement a suicide gene system can limit potential undesirable effects. In a haplo-identical bone marrow transplant clinical trial, over 90% of iCaspase-9-expressing cells were eradicated after AP1903 exposure, and signs of graft-versus-host disease disappeared. Nevertheless, low numbers of genetically modified T cells survived this treatment. We studied genetically modified cell lines (GMCL) that carried a dual iCaspase-9/ΔCD19 DNA construct (ΔCD19=truncated CD19). With AP1903 exposure, a low percentage of cells (1.47±0.67%; n=5 replications) persisted in vitro. Repeated exposures to increasing AP1903 doses generated low (GMCLLR) and high AP1903-responders (GMCLHR), which expressed different levels of surface ΔCD19 and intracellular iCaspase-9. Compared with GMCLHR, GMCLLR exhibited higher methylation of 5'-long-terminal repeat (LTR) promoters, both in the number of sequences with at least one methylated CpG (16 vs 51.5%, respectively) and in the number of CpG islands (1.2 vs 8.9%, respectively). Four days of 5-azacytidine exposure reduced methylation and increased ΔCD19 and iCaspase-9 expression. Interestingly, LTR demethylation restored GMCLLR sensitivity to AP1903 by 24.3-fold (1.8 vs 43.8%) without affecting GMCLHR. We showed that 5'-LTR-methylation inhibited transgene expression and caused AP1903 hypo-responsiveness. Treating with a hypomethylating agent restored AP1903 sensitivity. This approach can be applied in further clinical trials to improve iCaspase-9 response if low response is detected.