Nelson J. Chao, MD, MBA
Dr. Chao indicated no relevant conflicts of interest.
Bendle GM, Linnemann C, Hooijkaas AI, et al. Lethal graftversus- host disease in mouse models of T cell receptor gene therapy. Nat Med. 2010;16:565-570.
A major component for the success of allogeneic hematopoietic cell transplantation (HCT) is the graft-versus-leukemia (GVL) effect. It is a form of adoptive immunotherapy, albeit crude, in that the donor cells are polymorphic and heterogeneous, sometimes leading to GVL but often to graft-versus-host disease (GVHD) with or without GVL. The ability to induce GVL without GVHD remains the Holy Grail in HCT. There has been great interest in specific adoptive immunotherapy. The thinking is that perhaps one could remove all the non-specific T cells that could cause GVHD and give only those that are targeting a particular antigen, be it a tumor antigen or an infectious agent such as cytomegalovirus (CMV). Tumor- (or viral-) specific T cells can be created through transfer of the genes of unique T-cell receptor (TCR) into cytotoxic T cells (CTLs). These T cells will then express high levels of the transfected TCRs, creating CTLs that will recognize the targeted antigen although it will also express its original endogenous TCR. Clinical trials that are testing these approaches are currently ongoing.
The paper from the lab of Schumacher in Amsterdam raises a significant caution to these approaches. The concern is that the exogenous TCR (composed of the α and β chains) could rearrange with the endogenous αβ chains, as each chain is expressed individually and then combined, creating a novel TCR that would be auto-reactive. Sufficient auto-reactive T cells would then lead to GVHD, especially in the setting of HCT, where there is a major component of homeostatic T-cell proliferation after lymphodepletion. The authors demonstrated this concern by transfecting αβ TCR chains directed against the model antigen ovalbumin (OVA). They then gave these cells to mice that had been lymphodepleted by irradiation followed by interleukin-2 in vivo. Within two weeks, the authors noted that the animals developed cachexia, lymphopenia, colitis, and marrow failure, similar to findings of GVHD (called TCR gene transfer induced GVHD [TI-GVHD]). This induced GVHD occurred in animals that expressed OVA in their tissues and in animals that did not. They then demonstrated that this phenomenon was directly related to the infused OVA specific T cells, which had gained specificity for unidentified host antigens through recombination with the endogenous TCR; a single α or β chain was sufficient for this new specificity. The control T cells transduced with green fluorescent protein (GFP) did not develop GVHD.
Perhaps it should not have been surprising that TI-GVHD occurred in these experiments since the introduction of a polyclonal population of T cells, even with transfected TCRs, would likely represent many TCRs that could result from cross-pairing. The use of oligoclonal or monoclonal transgenic T cells does not result in TI-GVHD. It is not clear how relevant these observations are in humans as many of these studies are just beginning. However, these data suggest that it is important to develop strategies that would prevent such crosspairing of transduced and native α and β chains. Such approaches include the addition of cystine-modified transduced TCRs that allow for a new disulfide bond, thus reducing the cross-pairing or use of unique chimeric antigen receptors (CARs) that utilize the variable regions of immunoglobulin that are not likely to recognize host minor antigens. Other approaches could be to simultaneously introduce suicide genes, so that if these transduced T cells cause TI-GVHD, they could be destroyed. These are indeed complex systems.
back to top