Of Mice, Not Men: No Evidence for Graft-versus-Host Disease in Humans Receiving T-Cell Receptor–Transduced Autologous T Cells

Abstract

A recent article by Bendle et al.1 reported a high incidence of lethal graft-versus-host disease (GVHD) in mice receiving a lymphodepleting regimen followed by syngeneic cells transduced with genes encoding T-cell receptors (TCRs). The GVHD was manifested as cachexia, anemia, loss of hematopoietic reconstitution, pancreatitis, colitis, and death. A recent in vitro study by van Loenen et al. suggested that introduction of new TCRs into human lymphocytes could lead to the generation of mixed-TCR dimers with alloreactivity, although no in vivo data were presented.2 We have treated 106 patients in the Surgery Branch of the National Cancer Institute with autologous T cells transduced with seven different gammaretroviral vectors encoding antitumor TCRs (Table 1; refs. 3 and 4 and unpublished observations) and have seen no evidence of GVHD in any patient. Each of these patients received a lymphodepleting regimen of 60 mg/kg of cyclophosphamide for 2 days followed by 5 days of fludarabine at 25 mg/m2. The TCRs were of human origin in 77 patients and of mouse origin in 29 patients. In addition, we treated 6 patients who received the same cyclophosphamide-fludarabine regimen plus 600 cGy whole-body irradiation and TCR-transduced cells, and none of these patients showed any evidence of GVHD. An additional 44 patients were treated with these TCR-transduced cells without the lymphodepleting regimen, and none developed GVHD. The transduced TCRs in all of these studies manifested potent activity based on in vitro assays and the ability of transferred cells to mediate tumor regression in some patients. Table 1 Patients receiving autologous T-cell receptor–transduced peripheral blood lymphocytes or autologous tumor-infiltrating lymphocytes after the same lymphodepleting regimen We compared the clinical course of patients who received TCR-transduced cells with 115 patients treated by us who received the adoptive transfer of autologous nontransduced tumor-infiltrating lymphocytes following the same nonmyeloablative chemotherapy regimen.5,6 As shown in Table 2, there were no meaningful differences in clinical parameters between the two groups, including the number of days to recovery of absolute neutrophil count to 500 cells/mm3, absolute lymphocyte count to 1,000 cells/mm3, or platelets to 30,000 cells/mm3. In addition, there was no difference from start of treatment to discharge from the hospital, in weight loss or maximum liver function tests such as bilirubin. No patient in either group developed pancreatitis or prolonged colitis. There were no treatment-related deaths in any of the 106 patients who received TCR-transduced peripheral blood lymphocytes (PBLs). Table 2 Analysis of 103 patients who received T-cell receptor–transduced peripheral blood lymphocytes compared with 115 patients who received tumor-infiltrating lymphocytes It thus appears that in contrast to the report of mouse models by Bendle et al., humans receiving autologous T cells transduced with human or mouse TCRs did not develop evidence of GVHD. How then do we reconcile these divergent data? Bendle et al. claimed that a reduced level of GVHD occurs with transfer of transduced monoclonal compared with polyclonal populations; however, all our patients received polyclonal populations of transduced PBLs. To obtain the large number of cells needed to treat patients, human cells were often cultured for approximately 1 month as compared with only a few days of culture of mouse cells. However, 25 of the 106 patients received anti-TCR transduced cells cultured for only 4–8 days. In the human trials interleukin-2 administration was begun on the day of cell infusion as compared with 10 days after cell infusion in the mouse. Bendle et al. also reported a lower level of GVHD using an internal ribosomal entry site sequence rather than a 2A ribosomal skip sequence to link the TCR α- and β-chains, although we saw no GVHD using either construct nor the use of a PGK internal promoter (Table 1). Our clinical trials in patients with metastatic cancer refractory to standard treatments have resulted in objective cancer regressions using autologous cells transduced with the anti-MART-1:27-35, the anti-gp100:154-162 (refs. 2 and 3), as well as the anti-NY-ESO-1 and the anti-CEA TCRs (unpublished observations). Although a great deal has been learned from the use of mouse models, the article by Bendle et al. and the analyses included in our response emphasize a potential problem of using individual mouse models to predict toxicities in the human. Had Bendle and colleagues' paper been published before the conduct of human trials utilizing TCR-transduced cells, the ability to obtain regulatory approval for these human trials would have been severely limited. This study emphasizes the need to evaluate toxicities of gene therapy in humans utilizing a careful dose-escalation approach and raises cautions about undue reliance on models in inbred mice that attempt to replicate the human clinical situation.