RNAi-mediated TCR knockdown prevents autoimmunity in mice caused by mixed TCR dimers following TCR gene transfer

Abstract

Genetically modified T cells that express a transduced T cell receptor (TCR) α/β heterodimer in addition to their endogenous TCR are used in clinical studies to treat cancer. These cells express two TCR-α and two TCR-β chains that do not only compete for CD3 proteins but also form potentially self-reactive mixed TCR dimers, composed of endogenous and transferred chains. To overcome these deficits, we developed an RNAi-TCR replacement vector that simultaneously silences the endogenous TCR and expresses an RNAi-resistant TCR. Transduction of the virus-specific P14 TCR without RNAi resulted in unequal P14 TCR-α and -β chain surface levels, indicating heterodimerization with endogenous TCR chains. Such unequal expression was also observed following TCR gene optimization. Equal surface levels of the introduced TCR chains were however achieved by silencing the endogenous TCR. Importantly, all mice that received cells transduced with the native or optimized P14 TCR developed lethal TCR gene transfer-induced graft-versus-host-disease (TI-GVHD) due to formation of mixed TCR dimers. In contrast, TI-GVHD was almost completely prevented when using the RNAi-TCR replacement vector. Our data demonstrate that RNAi-assisted TCR replacement reduces the formation of mixed TCR dimers, and thereby significantly reduces the risk of TI-GVHD in TCR gene therapy.

Figure 1

Intronic miRNA results in superior transgene expression compared with exonic miRNA. (a) A miRNA targeting the T cell receptor (TCR)-β chain (miRβ) was introduced into GFP-encoding MP71 vectors at following positions: 5′ intronic (5′ int.), 5′ exonic (5′ ex.) and 3′ exonic (3′ ex.). LTR, long terminal repeat; PRE, post-transcriptional regulatory element; SA, splice acceptor; SD, splice donor; Ψ, packaging signal. (b) TCR surface levels and GFP expression of transduced polyclonal splenocytes were determined by flow cytometry. The percentage of gated cells is indicated. (c,d) Median fluorescence intensity (MFIs) were compared to the parental control vector (contr.). Transduction (Td) efficiencies: contr. GFP (83%, 78%), 5′ int. miRβ-GFP (66%, 68%), 5′ ex. miRβ-GFP (52%, 45%), 3′ ex. miRβ-GFP (62%, 63%). (e) GFP-encoding MP71 vectors expressing either each miRNA for the silencing of the TCR α (miRα) and β chain (miRβ) separately or both miRNAs together (miR). (f,g) TCR/CD3 complex surface levels and GFP expression were determined and compared to the parental vector (contr.). Td efficiencies: contr. (68%, 62%); miR-GFP (44%, 37%); miRα-GFP (51%, 50%); miRβ-GFP (47%, 42%). Plots show the mean of two independent experiments ± SD (n = 2).

Figure 2

Silencing of the endogenous T cell receptor (TCR) supports the expression of an RNAi-resistant second TCR. (a) The miRNA vectors were used for the expression of an RNAi-resistant P14 TCR cassette. (b) The P14 TCR-α chain (TCR Vα2) and the P14 TCR-β chain (TCR Vβ8) surface levels of transduced polyclonal T cells were analyzed by flow cytometry. CD8 T cells are shown. Numbers indicate the percentage of cells in each the quadrant gate. Nontransduced T cells served as control (contr.). MFIs of the transduced cells were calculated using the gates R1 (miR-P14), R2 (miRα-P14), and R3 (miRβ-P14). (c) MFIs of cells expressing only one miRNA were compared to cells expressing both miRNAs. Transduction efficiencies: miR-P14 (62%, 62%); miRα-P14 (73%, 65%); miRβ-P14 (72%, 70%). Plot shows the mean of two independent experiments ± SD (n = 2).

Figure 3

RNAi-assisted T cell receptor (TCR) replacement results in equal surface levels of the transferred TCR chains. (a) Polyclonal T cells were transduced with four P14 TCR vectors differing from each other by the presence of the intronic miR cassette (miR) and the usage of the optimized P14 TCR cassette (opt). Cells were stained either for CD8 and for both P14 TCR chains (TCR Vα2, TCR Vβ8) or for CD8 and using a P14 TCR-specific MHC multimer (Db-GP33). CD8 T cells are shown and the percentages of the gated cells are indicated. Nontransduced T cells served as control (contr.). (b) Median fluorescence intensity (MFIs) of transduced CD8 T cells were compared to nontransduced CD8/Vα2Vβ8 T cells (end. TCR). (c) Plot shows the proportion of cells expressing both P14 TCR chains that bind the Db-GP33 multimer. (d) Plot shows the fold change of the Db-GP33 multimer MFI relative to the P14 vector-transduced sample. Horizontal bar (group mean, indicated on top). Vertical bar (SD). Combined data of six independent experiments are shown in b, c, and d (n = 6). Transduction efficiencies were comparable (Supplementary Figure S3b).

Figure 4

miR-T cell receptor (TCR) gene-modified T cells show antitumor reactivity and increased antigen-specific proliferation. (a) Transduced polyclonal T cells were analyzed either for the expression of both P14 TCR chains (P14 TCRαβ) or binding of the P14 TCR-specific MHC multimer (Db-GP33) at the day of adoptive T cell transfer (ATT). 2 × 10⁶ congenic CD8/P14 TCRαβ T cells (B6.SJL) were transferred into recipients (C57BL/6) that received 3 days before 1 × 10⁶ GP33-expressing tumor cells (B16-GP33) intravenously. (b) Plot shows the number of macroscopically visible experimental (exp.) lung metastasis 21 days after ATT. *The mice of the control group without ATT were sacrificed after 14 days. (c,d) Transferred T cells in peripheral blood were analyzed at day 7 after ATT. Results were confirmed by a second measurement at day 14 (Supplementary Figure S5). Symbols represent individual mice. Horizontal bar (group mean, indicated on top). Vertical bar (SD). Group sizes: P14 (n = 6), miR-P14 (n = 7), P14opt (n = 6), miR-P14opt (n = 7), control (n = 5).

Figure 5

RNAi-assisted T cell receptor (TCR) protein replacement severely reduces the incidence of TI-GVHD caused by mixed TCR dimers. (a,b) TI-GVHD is induced by P14 vector-transduced (Td) T cells but not by P14 TCR-transgenic (Tg) T cells. Cell populations containing 1 × 10⁶ P14 TCR-expressing CD8 cells were transferred into irradiated mice. Ten days after adoptive T cell transfer (ATT), the mice were treated for 3 days with IL-2 twice a day and monitored for autoimmune symptoms. The control group was treated identical but received nontransduced T cells. A Kaplan–Meier plot of disease-free survival and the change of body weight at day 14 after ATT are shown in a and b, respectively. Horizontal bar (group mean, indicated on top). Vertical bar (SD). Group sizes, Td efficiencies: control (n = 7), P14 Td (n = 8, 20%), P14 Tg (n = 5). (c,d) TI-GVHD is reduced in mice receiving P14 TCR-transduced T cells with silenced endogenous TCR. Results of two independent experiments are shown in c (Kaplan–Meier plot) and d (change of body weight at day 12 after ATT). Two mice of the P14 group had to be sacrificed because of TI-GVHD at day 11 and were therefore not included in d. Horizontal bar (group mean, indicated on top). Vertical bar (SD). Group sizes, Td efficiencies: control (n = 15), P14 (n = 16, 55%, 56%), miR-P14 (n = 16, 35%, 62%), P14opt (n = 15, 44%, 21%), miR-P14opt (n = 16, 65%, 57%).