Protective Capacity of Virus-Specific T Cell Receptor-Transduced CD8 T Cells In Vivo

Abstract

The transfer of T cell receptor (TCR) genes by viral vectors represents a promising technique to generate antigen-specific T cells for adoptive immunotherapy. TCR-transduced T cells specific for infectious pathogens have been described, but their protective function in vivo has not yet been examined. Here, we demonstrate that CD8 T cells transduced with the P14 TCR specific for the gp33 epitope of lymphocytic choriomeningitis virus exhibit protective activities in both viral and bacterial infection models in mice.

TEXT

The transfer of T cell receptor (TCR) genes into primary T cells by viral vectors represents a powerful method to generate antigen-specific T cells for adoptive immunotherapy. This approach is widely used in immunotherapy against tumors (reviewed in references 11, 13, and 20). T cells specific for viruses, bacteria, and parasites have also been generated by TCR gene transfer (2, 6, 8, 9, 12, 14, 17, 18, 21, 23), but we are not aware of any report analyzing the protective activity of TCR-transduced T cells in infection models in vivo. To address this issue, we used the well-characterized major histocompatibility complex (MHC) class I-restricted P14 TCR (15) specific for the gp33 epitope of lymphocytic choriomeningitis virus (LCMV) to assess the protective activity of TCR-transduced CD8 T cells in infection models in mice.

Splenocytes from B6.PL-Thy1a/CyJ (B6.Thy1.1) mice were transduced with P14 TCR α and β genes as described previously (10). Anti-CD3/CD28-activated splenocytes were transduced on days 1 and 2, followed by in vitro culture in interleukin-15 (IL-15) (50 U/ml). On day 6, cells were used for adoptive transfer, and at this time point, 51% ± 6% of CD8 T cells coexpressed the transduced P14 TCR Vα2 and Vβ8 chains (Fig. 1A). The proliferative potential of P14 TCR-transduced CD8 T cells (P14-td) was tested in adoptive transfer experiments. A total of 10⁵ P14-td cells (Thy1.1⁺) were transferred intravenously (i.v.) into (B6.Thy1.1 × B6.Thy1.2) F1 mice that had been infected i.v. with 200 PFU LCMV (strain WE) 1 day before. A strong expansion of the P14-td cells was observed, peaking 1 week postinfection (p.i.) followed by decline (Fig. 1B, filled circles). Expansion of P14-td cells was antigen specific since these cells did not expand after infection of the recipient mice with the LCMV variant 8.7 (16), which carries a point mutation in the gp33 epitope (Fig. 1B, open circles). P14-td cells persisted (>6 weeks p.i.) in LCMV-immune recipient mice and coexpressed the transduced P14 TCR Vα2 and Vβ8 chains at increased percentages compared to the initially transferred T cells (Fig. 1B, right, versus Fig. 1A). Adoptive transfer of TCR-transduced T cells into sublethally irradiated hosts has recently been shown to cause lethal graft-versus-host disease (GVHD) due to formation of self-reactive TCRs (4). The LCMV-infected recipients of P14 TCR-td cells described here were healthy, and overt pathologies were not observed. This is most likely due to the fact that the recipient mice in our setting were not irradiated. In addition, we used a P2A-containing retroviral vector which reduces mispairing of endogenous and introduced TCR chains, and we did not apply high-dose IL-2 administration (4).

Fig 1.

Generation and characteristics of P14 TCR-transduced cells. (A) Flow cytometric analysis of spleen cells 5 days after retroviral transduction with P14 TCR genes. A representative dot plot (left) gated on CD8 T cells showing expression of the transduced P14 TCR chains and a statistical graph (right) revealing transduction efficiencies are depicted. Dots in the statistical graph represent values from 8 independent experiments. The horizontal dashed line shows the mean percentages of Vα2⁺ Vβ8⁺ cells in the absence of transduction. (B) Expansion of 10⁵ adoptively transferred P14-td T cells (CD8⁺ Vα2⁺ Vβ8⁺) in the recipient mice infected with wild-type LCMV (filled circles) or LCMV variant 8.7 (open circles). Percentages of P14-td (Thy1.1⁺1.2⁻) cells of peripheral blood lymphocytes (PBL) of individual mice that were bled at the indicated time points p.i. are shown. Data are derived from 2 independent experiments with 4 or 5 mice per group. A representative dot plot displaying TCR Vα2 and Vβ8 expression gated on donor Thy1.1⁺1.2⁻ cells from the spleen 6 weeks p.i. is shown on the right. (C) Cell surface phenotype and IFN-γ production determined by intracellular cytokine staining of P14-td cells isolated from noninfected (open circles) and LCMV-infected (closed circles) hosts. P14-td cells from the spleens of noninfected and infected recipients were analyzed 2 and 6 weeks, respectively, after cell transfer. Iono, ionomycin. Data are derived from 3 mice per group. Statistical significance was evaluated with Student's t tests using InStat3 software (GraphPad). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

To determine whether the transferred P14-td cells establish memory characteristics prior to LCMV challenge, P14-td cells in recipient mice with and without LCMV infection were compared (Fig. 1C). Under both conditions, most of the P14-td cells expressed CD44 and CD127 at high levels. The effector cell marker KLRG1 (22) was only detected on a subset of P14-td cells from LCMV-infected mice. In contrast to P14-td cells from LCMV-immune mice, P14-td cells in the absence of in vivo priming failed to produce significant amounts of gamma interferon (IFN-γ) after stimulation with the cognate gp33 peptide antigen. Even after stimulation with phorbol myristate acetate (PMA)-ionomycin, IFN-γ production by P14-td cells from uninfected mice was considerably lower than that of P14-td cells from LCMV-infected mice. Thus, P14-td cells in the absence of in vivo priming showed an antigen-experienced CD44^high phenotype but weak effector activity.

The protective activity of P14-td cells was first examined in the LCMV infection model. B6 recipient mice were infected with 200 PFU LCMV-WE (i.v.), and 1 day later, graded numbers of P14-td cells (CD8⁺ Vα2⁺ Vβ8⁺) were transferred (i.v.). On day 4, expansion of P14-td cells in the spleen was assessed by flow cytometry and LCMV titers were determined (3). For comparison, CD8 T cells from Thy1.1⁺ P14 TCR-transgenic (P14-tg) mice were used side by side in these experiments. Similarly to P14-td cells, P14-tg cells were also activated by anti-CD3/CD28 stimulation followed by in vitro culture in the presence of IL-15 before transfer. The data shown in Fig. 2A demonstrate that the adoptively transferred P14-td cells expanded in the recipient mice almost as efficiently as P14-tg cells after LCMV infection. Most importantly, P14-td cells were able to decrease LCMV titers in the spleen of recipient mice with a cell-per-cell efficacy comparable to that of P14-tg cells (Fig. 2B).

Fig 2.

Clonal expansion and anti-LCMV activity of transduced versus transgenic P14 T cells. B6 mice were first infected with LCMV (200 PFU i.v.), and 1 day later, the indicated numbers of P14 TCR-transduced (P14-td) or P14 TCR-transgenic (P14-tg) cells were adoptively transferred (i.v.). Mice were analyzed 4 days after infection. (A) Percentages of P14-td (left) and P14-tg (right) cells within CD8 T cells in the spleen. (B) LCMV titers in the spleen determined by standard virus plaque assay. Dots in the graphs represent values from individual mice with at least 4 mice per group. Data are derived from 2 or 3 independent experiments. (C) CD8 T cells (10⁶) transduced with OT-1 or P14 TCRs were transferred into B6 mice followed by infection with 200 PFU of wild-type LCMV or of LCMV variant 8.7. LCMV titers were determined on day 4 p.i. Data are from 3 mice per group. Horizontal dashed lines mark the detection limits. Statistical significance was evaluated with the Mann-Whitney U test using InStat3 software (GraphPad). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

To control the antigen specificity, splenocytes were also transduced with the OT-1 TCR specific for ovalbumin using the same retroviral vector backbone (10). The retroviral transduction efficacies of OT-1 and P14 TCRs, determined by specific TCR staining, were comparable (data not shown). In contrast to P14-td cells, OT-1-td cells failed to lower LCMV titers (Fig. 2C, left side). As an additional specificity control, recipient mice were also infected with LCMV variant 8.7. In contrast to wild-type LCMV, P14-td cells failed to decrease LCMV 8.7 titers. Thus, the observed antiviral activity of P14 TCR-td cells was antigen specific.

Next, we examined whether P14-td cells were capable of inhibiting replication of vaccinia virus (VV) in ovaries by intraperitoneally (i.p.) infecting B6 mice with 2 × 10⁶ PFU of recombinant VV expressing LCMV-GP (Vacc_LCMV-GP) (7). It is known from previous work that large numbers of effector-like T cells have to be present before viral challenge to observe protective activity in this infection model (1). Similar to P14-tg cells, P14-td cells given 2 days before infection decreased VV titers in ovaries by 1 to 2 logs in this “demanding” infection model (Fig. 3A). Finally, we tested whether P14-td cells could confer protection against infection with recombinant Listeria monocytogenes expressing the gp33 epitope (rLm_gp33) (19). P14-td cells were adoptively transferred into B6 recipients, and 1 day later, mice were infected with 2 × 10⁴ CFU rLm_gp33. As shown in Fig. 3B, bacterial loads in spleen and liver were reduced by 2 to 3 logs in recipients of P14-td cells on days 2 and 3 p.i., indicating that the TCR-transduced cells mediated protective activity in the L. monocytogenes infection model.

Fig 3.

Anti-vaccinia virus and anti-L. monocytogenes activity of P14 TCR-transduced cells. (A) The indicated numbers of P14 TCR-transduced (P14-td) or P14 TCR-transgenic (P14-tg) cells were adoptively transferred into B6 mice. On day 2, mice were infected with 2 × 10⁶ PFU rVV_LCMV-GP and VV titers in ovaries were determined on day 5 p.i. The symbols in the graphs represent values from individual mice with at least 4 mice per group. Data are derived from 5 (P14-td) and 2 (P14-tg) independent experiments. (B) A total of 5 × 10⁶ P14-td cells were adoptively transferred into B6 mice. On day 1, mice were infected (i.v.) with 2 × 10⁴ CFU recombinant Listeria monocytogenes cells expressing the gp33 epitope (rLm_gp33). Bacterial loads in spleen and liver were determined on days 2 and 3 after infection. Horizontal dashed lines mark the detection limits. Symbols in the graphs represent values from individual mice with at least 7 mice per group. Data are derived from 5 independent experiments. Statistical significance was evaluated with the Mann-Whitney U test using InStat3 software (GraphPad). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Taken together, our data demonstrate that CD8 T cells transduced with the P14 TCR showed protective activity in three different infection models. The protective activities of P14-td cells were more pronounced in the LCMV model compared to the L. monocytogenes or VV model. Our data further revealed that P14-td cells isolated from noninfected recipient mice failed to produce significant levels of IFN-γ without further in vivo priming. It is well known that IFN-γ is crucial for control of L. monocytogenes (5), and thus, P14-td cells that are able to produce this cytokine more efficiently might be more potent in this infection model. Such cells may be generated by adding polarizing cytokines such as IL-12 to the T cell cultures. In the LCMV protection assay, P14-td cells without functional memory characteristics showed potent activity. This result fits well with our previous observation that even naïve P14 T cells from TCR-tg mice can considerably lower LCMV titers in the spleen after adoptive transfer (24). Since T cells efficiently home to the spleen, the main site of LCMV replication, virus clearance by adoptively transferred T cells is facilitated in the LCMV model. This is distinct from the VV infection model, in which viral titers were assessed in the ovaries, which harbor limited numbers of T cells in the absence of an infection. In the VV model, effector memory TCR-td cells with increased homing capacity to peripheral organs are likely to be more efficient.