Recombination of Endogenous TCR Chains with Retrovirally Introduced TCR Chains Can Result in Mixed T Cell Receptor Dimers Harbouring Harmful Alloreactivity

T cell receptor transfer to engineer tumor specific T cells is being explored as a strategy for adoptive immunotherapy. By retroviral introduction of T cell receptors (TCRs), large numbers of T cells with defined antigen specificity can be obtained. The in vivo efficacy of adoptively transferred TCR engineered T cells has been demonstrated in mouse studies and recently the first clinical trial with TCR engineered T cells was performed in melanoma patients. However, a potential drawback of TCR gene transfer is the formation of mixed TCR dimers. Chains of the introduced TCR can pair with the endogenous TCR chains, resulting in unknown specificities, and potentially in harmful reactivity against patient HLA molecules. We investigated whether TCR gene transfer leads to the generation of new detrimental reactivities by creating T cells that expressed mixed TCR dimers. To be able to discriminate between the antigen specificity of the mixed TCR dimers and the introduced as well as the endogenous TCR, we transduced mono-specific T cells with seven different antigen specific TCRs. As mono-specific T cells we used CMV-pp50 specific HLA-A1 restricted T cells. The transduced T cells were analyzed for newly acquired specificities against a large HLA-typed EBV-LCL panel covering almost all HLA class I and II molecules. We transduced several polyclonal virus specific T cell populations with the seven different antigen specific TCRs, and showed that in all T cell populations at least one of the seven TCR-transduced populations acquired new alloreactivities. Furthermore, by randomly combining TCR alpha and beta chains derived from different T cell clones we created 60 mixed TCR dimers of which 17 acquired alloreactivity. These results indicate that recombination of the introduced TCR chains with the endogenous TCR chains frequently gives rise to new allospecificities. To ascertain that the newly acquired alloreactivities were exerted by mixed TCR dimers, we introduced only TCR alpha or beta chains into CMV-pp50 specific monoclonal T cells, and demonstrated for example, that the introduction of a CMV pp65 specific TCR alpha chain led to a newly acquired reactivity that was HLA B58 restricted. The introduction of only the beta chain of a minor histocompatibility antigen (mHag) HA-1 specific TCR led to a newly acquired HLA B52 specific reactivity. Furthermore, we analyzed whether mixed TCR dimers consisting of conserved TCRs with the same specificity could acquire new harmful reactivity. We recombined mHag HA-2 specific TCR alpha and beta chains from 4 different T cell clones. Of the 12 mixed TCR dimers, a combination of the mHag HA-2 specific TCR alpha chain derived from the HA2.6 T cell clone with the mHag HA-2 specific beta chain of clone HA2.19 resulted in alloreactivity that was HLA DQ3 restricted. These results indicate that each recombination of TCR chains after TCR gene transfer can potentially result in a harmful new reactivity. In conclusion, mixed TCR dimers due to pairing of endogenous TCR chains with introduced TCR chains acquire potentially dangerous reactivities, both class I and class II restricted. To limit the chance of generating self- or alloreactive T cells, TCRs may be constructed allowing selective pairing of the TCR alpha chain with the corresponding TCR beta chain. Alternatively, we propose to use virus specific T cells as host cells for TCR gene transfer. Since they consist of a restricted TCR repertoire, the number of different chimeric TCRs formed will be limited. By introducing into these T cells as controls only the alpha or beta chain of the TCR of interest, the reactivity of these T cells and harmful reactivities of the mixed TCR dimers can be tested against different patient derived cell types.