Tumor-Targeted Human CD4⁺ CD25^hi Regulatory T Cells Effectively Inhibit T Cell –mediated Rejection of CD19⁺ Tumors in An in Vivo xenotransplant Tumor Model

The genetic targeting of human T cells to selected tumor antigens offers a novel means to investigate human immunobiology and treat cancer. T cells may be genetically modified to target specific antigens through the introduction of genes encoding chimeric antigen receptors (CARs). We have previously demonstrated that human T cells targeted in this manner to the CD19 antigen, expressed on normal B cells as well as most B cell tumors, eradicate systemic human CD19⁺ B cell malignancies in SCID-Beige mice. However, in the clinical setting, the anti-tumor efficacy of these T cells may be impaired by endogenous suppressive elements of the host immune system, including CD4⁺ CD25^hi Foxp3⁺ regulatory T cells (Tregs). Significantly, Tregs are often increased in the blood and infiltrate the tumor of cancer patients which has been correlated with poor patient outcome and ineffective anti-tumor immunity. In order to study the in vivo impact of Tregs on adoptive therapy with CD19 targeted effector T cells, we developed a murine model wherein human Tregs, similarly targeted to the tumor, are infused prior to adoptive transfer of targeted cytotoxic T cells. To do so, we initially isolated natural Tregs from healthy donor peripheral blood mononuclear cells. Isolated Tregs were subsequently modified to express CARs through retroviral gene transfer. Subsequently, CAR⁺ Tregs were rapidly expanded either by activation on NIH-3T3 fibroblasts modified to express CD19 and the CD80 costimulatory ligand (3T3(CD19/CD80)), or non-specifically using CD3/CD28 antibodycoated magnetic beads. Expanded CAR⁺ Tregs exhibited potent suppressive function in vitro inhibiting both effector T cell proliferation as well as cytotoxicity. In vivo, CAR⁺ Tregs specifically traffic to established tumor in SCID-Beige mice. Significantly, injection of CD19-targeted Tregs into SCID-Beige mice bearing established human CD19⁺ tumors at 24 hours prior to infusion with CD19-targeted effector T cells, completely abrogated effector T cell function even at Treg:Teff ratios as low as 1:8. We further found that full suppression was dependant both on Treg localization to the tumor site as well as in vivo activation through the CAR. Finally, we show that a pre-conditioning regimen with low-dose cyclophosphamide, which failed to eradicate tumor, was able to reverse the CAR⁺ Treg mediated inhibition and restore the anti-tumor activity by the targeted effector T cells. In conclusion, we have developed a robust model ideally suited to the study of in vivo Treg-Teff interactions. Furthermore, the data generated from this model to date have significant implications with respect to the application of adoptive T cell therapies in the clinical setting. Namely, the presence of endogenous Tregs at the site of tumor is likely to significantly compromise the anti-tumor activity of adoptively transferred tumor targeted T cells. This inhibition may be reversed by preconditioning regimens designed to eradicate endogenous Tregs. The findings presented here should be considered in the design of future clinical trials utilizing T cell-based adoptive therapies of cancer.