Mixed T cell receptor dimers harbor potentially harmful neoreactivity

Abstract

Adoptive transfer of T cell receptor (TCR)-transduced T cells may be an attractive strategy to target both hematological malignancies and solid tumors. By introducing a TCR, large numbers of T cells with defined antigen (Ag) specificity can be obtained. However, by introduction of a TCR, mixed TCR dimers can be formed. Besides the decrease in TCR expression of the introduced and endogenous TCR, these mixed TCR dimers could harbor potentially harmful specificities. In this study, we demonstrate that introduction of TCRs resulted in formation of neoreactive mixed TCR dimers, composed of the introduced TCR chains pairing with either the endogenous TCR alpha or beta chain. Neoreactivities observed were HLA class I or class II restricted. Most neoreactive mixed TCR dimers were allo-HLA reactive; however, neoreactive mixed TCR dimers with autoreactive activity were also observed. We demonstrate that inclusion of an extra disulfide bond between the constant domains of the introduced TCR markedly reduced neoreactivity, whereas enhanced effectiveness of the introduced TCR was observed. In conclusion, TCR transfer results in the formation of neoreactive mixed TCR dimers with the potential to generate off-target effects, underlining the importance of searching for techniques to facilitate preferential pairing.

Fig. 1.

TCR td virus-specific T cells demonstrate neoreactivity. T cells were tested against a broad LCL panel. IFN-γ production of three of the five different virus-specific T cell populations transduced with three different TCRs is depicted, namely of (A) pp50-specific T cells isolated from healthy individual CVO, (B) pp50-specific T cells of healthy individual UKL, and (C) IE-1–specific T cells isolated from healthy individual MBX. As a control for the reactivity of the endogenous and introduced TCR, LCLs with the restricting HLA molecules were pulsed with the relevant peptides (endo-TCR and intro-TCR, respectively). IFN-γ production depicted is representative of three separate experiments performed in duplicate.

Fig. 2.

Neoreactivities caused by mixed TCR dimers. pp50 T cells from CVO were transduced with (A) HA1.M7-TCR α or β chains or with (B) CMV-TCR α or β chains, and IE-1-T cells from MBX were transduced with (C) HA2.6-TCR α or β chains and tested against the LCL panel for neoreactivity. As a control for the reactivity of the endogenous TCR, LCLs with the restricting HLA molecules of the endogenous virus-specific TCR were pulsed with the viral peptides (A–C; endo-TCR). IFN-γ production depicted is representative of two separate experiments.

Fig. 3.

Neoreactivity of mixed TCR dimers is both HLA class I and II restricted. HLA restriction of neoreactive mixed TCR dimers of (A) BV1⁺ HA1.M7-TCRβ td CVO pp50 T cells, (B) BV1⁺ CMV-TCRα td CVO pp50 T cells, or (C) BV1⁺ HA2.6-TCRβ td MBX IE-1 T cells was tested in IFN-γ production assay using blocking antibodies. Relevant HLA typing of the LCLs is indicated, with the HLA typing of the blocked LCL shown in bold. IFN-γ production depicted is representative of three separate experiments performed in triplicate.

Fig. 4.

Neoreactivity also directed against normal human cell subsets. (A) CMV-TCRα td and non–td CVO pp50 T cells were tested against HLA-B58⁺ LCL ZIL and HLA-B58⁻ LCL from CVO and CD14⁺, CD19⁺, and CD4⁺ MACS-isolated cell subsets derived from HLA-B58⁺ healthy individual IGN. HA-2.6-TCRβ td and non–td MBX IE-1 T cells were tested against LCLs and CD14⁺, CD19⁺ and CD4⁺ MACS-isolated cell subsets derived from (B) HLA-DR17⁺ MBX or (C) HLA-DR17⁺ NGI. HLA-DR17⁻ LCL EBM was included in the experiment as a control. T cells were tested against resting cell subsets (ex vivo) or activated cell subsets. CD14⁺ cells were either activated into immature DCs (iDC) or mature DCs (mDCs) using activating cytokines. CD19⁺ cells were activated using activating cytokines and CD40L (B act). CD4⁺ cells were activated using PHA (T act). IFN-γ production depicted is representative of two separate experiments.

Fig. 5.

Neoreactivities are markedly diminished using cysteine modified TCRs. (A) BZLF-1 T cells from healthy individual UKL were transduced with two separate retroviral vectors encoding either the unmodified HA1.M7-TCR α and β chains (HA1.M7-TCR WT; black bars) or cysteine-modified HA1.M7-TCR α and β chains (HA1.M7-TCR SS; gray bars), sorted on the bases of high eGFP and ΔNGF-R expression and tested against several LCLs for neoreactivity in duplicate. (B) CVO pp50 T cells were transduced with retroviral vectors containing T2A-linked unmodified HA1.M7-TCR α and β chain (HA1.M7-TCR 2A WT; black bars) or T2A-linked cysteine-modified HA1.M7-TCR α and β chain (HA1.M7-TCR 2A SS; gray bars), sorted on high ΔNGF-R expression and tested in duplicate against several LCLs for neoreactivity. As a control, non–td UKL BZLF-1 T cells and CVO pp50 T cells (non-td; white bars) were tested against the same LCL panel. IFN-γ production depicted is representative of three separate experiments.