Antigen presented by tumors in vivo determines the nature of CD8+ T-cell cytotoxicity

Abstract

The biological relevance of the perforin and Fas ligand (FasL) cytolytic pathways of CD8(+) T lymphocytes (CTL) for cancer immunotherapy is controversial. We investigated the importance of these pathways in a murine renal cell carcinoma expressing influenza viral hemagglutinin as a defined surrogate antigen (Renca-HA). Following Renca-HA injection, all FasL-dysfunctional FasL(gld/gld) mice (n = 54) died from Renca-HA tumors by day 62. By contrast, perforin(-/-) (51%; n = 45) and Fas(lpr/lpr) (55%; n = 51) mice remained tumor-free at day 360. Blocking FasL in vivo inhibited tumor rejection in these mice. Moreover, established Renca-HA tumors were cleared more efficiently by adoptively transferred HA(518-526)-specific T-cell receptor-transgenic CTL using FasL rather than perforin. Strikingly, a range of mouse tumor cells presenting low concentrations of immunogenic peptide were all preferentially lysed by the FasL but not the Pfp-mediated effector pathway of CTL, whereas at higher peptide concentrations, the preference in effector pathway usage by CTL was lost. Interestingly, a number of human renal cancer lines were also susceptible to FasL-mediated cytotoxicity. Therefore, the FasL cytolytic pathway may be particularly important for eradicating Fas-sensitive tumors presenting low levels of MHC class I-associated antigens following adoptive T-cell therapy.

Figure 1. Antitumor role of Fas ligand in vivo

Tumor cells were injected in BALB/c WT, SCID, Pfp^−/−, FasL^gld/gld and Fas^lpr/lpr mice. (A) Survival of Renca-HA (1 × 10⁴ subcutaneous) tumor-bearing SCID (n = 43), WT (n = 55), Pfp^−/− (n = 45), FasL^gld/gld (n = 54) and Fas^lpr/lpr (n = 51) mice. *P <0.0001 (WT compared to SCID; or Pfp^−/− or Fas^lpr/lpr compared to FasL^gld/gld) (two-sided log-rank test). (B) Lung metastases counts on day 14 following tumor (0.5 × 10⁶ intravenous) injection. In some groups, mice were administered intraperitoneally with mAb MFL4 against Fas ligand or its isotype IgG (0.5 mg each) on day 0, 3, 7 and 10. *P < 0.0001 (ANOVA, two-sided) for WT (n = 23), Fas^lpr/lpr (n = 18) and Pfp^−/− (n = 23) compared to SCID (n = 13) or FasL^gld/gld (n = 18), or for MFL4-treated WT (n = 13), Pfp^−/− (n = 13) or Fas^lpr/lpr (n = 8) compared to equal numbers of control isotype-treated mice. Data from 7 (A), and 6 (B) independent experiments.

Figure 2. Activation of CTL in Renca-HA-bearing mice

(A) Lung metastases counts on day 14 following Renca-HA (3 × 10⁶ intravenous) in BALB/c WT, Tg, Tg^Pfp−/− and Tg^gld/gld mice. *P < 0.0001 (ANOVA, two-sided) for Tg (n = 8) and Tg^Pfp−/− (n = 9) compared to WT (n = 8) and Tg^gld/gld (n = 9) mice (pool of 2 independent experiments). **P < 0.0001 for Tg^gld/gld compared to WT. (B) Total number of cells ± SD in tumor-draining (axillary and inguinal) and contralateral LN harvested on day 7 post Renca-HA (5 × 10⁶ subcutaneous) injection in WT, Tg, Tg^Pfp−/− and Tg^gld/gld mice (n = 4). *P < 0.0001 (ANOVA, two-sided) for the draining LN compared to the contralateral LN. (C) Dot plots of K^d:HA-tetramer versus CD25, CD44 and intracellular IFN-γ on gated CD8⁺ cells in the tumor-draining LN. Numbers in the quadrants represent % positive cells. Plots representative of 2 independent experiments.

Figure 3. Fas ligand effector pathway is more efficient than Pfp in vivo

Lung metastases counts on day 14 following Renca-HA (2 × 10⁶ intravenous) injection in RAG2^−/− mice (A) that were transferred with the indicated numbers of the activated Tg^Pfp−/− and Tg^gld/gld LN CTL intravenously on day 7; *P = 0.004, **P < 0.0001 (ANOVA, two-sided) for Pfp^−/− (n = 10) compared to the corresponding gld (n = 10) group; (B) that were administered with FasL-neutralizing MFL4 or its isotype IgG on day 0, 2 and 5 post transfer of the activated Tg and Tg^Pfp−/− CTL (0.1 × 10⁶). *P < 0.0001 (ANOVA, two-sided) for MFL4-treated (n = 9) compared to the corresponding control Ab-treated (n = 9) group. (C) The activated Tg^Pfp−/− and Tg^gld/gld LN CTL (5 × 10⁶ intravenous) were transferred intravenously in RAG2^−/− mice bearing 10-day-established Renca-HA tumor metastases. Representative sections of lungs and WT mouse control thymus harvested a day after CTL transfer and stained with CD8α antibody. Bar = 25 µm, all images. Data representative of 2 independent experiments. (D) Bar graphs depict IFN-γ secretion in the day 4 supernatants from the HA-stimulated Tg, Tg^Pfp−/− and Tg^gld/gld LN CTL cultures. Values are means ± SD of triplicate samples in a representative of two independent experiments. Histograms represent the % expression of granzyme B, perforin and FasL on gated CD8⁺ TCR Vβ8.1⁺ cells in the HA-stimulated Tg LN CTL as compared to naïve Tg LN T cells, with numbers depicting the mean fluorescence intensities for naïve (red) and activated (green) Tg T cells.

Figure 4. Renca-HA tumor cells express Fas in vivo

(A) Fas (M-20)-stained sections (20× magnification) of Renca-HA solid tumors from Fas^lpr/lpr and FasL^gld/gld mice, Fas^lpr/lpr thymus (negative control) and WT mice (positive control). Bar = 25 µm, all images. (B) Dot plots of Fas (Jo2) versus MHC class-I K^d (right panel) on gated HA⁺ cells (left panel) in the tumor cells isolated from day-12-established Renca-HA tumors in RAG2^−/− mice. Numbers represent % positive cells. (C) Dot plots of Fas versus MHC class-I K^d (right panel) on Renca-HA or Renca tumor cells treated overnight with medium or with 500 U/ml of TNF-α and IFN-γ. Numbers represent % positive cells. Data representative of 2 independent experiments.

Figure 5. FasL and Pfp effector pathways play non-redundant roles

(A) Percent specific lysis of Renca-HA or Renca cells, pre-treated with TNF-α and IFN-γ (500 U/ml each, overnight), pulsed with various concentrations of HA peptide and then incubated for 18 h with activated Tg, Tg^Pfp−/− and Tg^gld/gld CTL at effector:target ratio 25:1. *P < 0.001 (two-sided unpaired t test) with respect to other CTL. (B) Percent specific lysis of Renca-HA cells pulsed with 10⁻⁶ M HA peptide by Tg^gld/gld CTL pretreated 2 h with medium or with concanamycin A (upper panel). Percent specific lysis of Renca-HA cells pre-treated with medium or with TNF-α and IFN-γ (500 U/ml each, overnight) by Tg^Pfp−/− CTL in the presence of FasL-neutralizing Ab or its isotype control (lower panel). *P < 0.001 (two-sided unpaired t test) with respect to medium or isotype control. (C) Percent specific lysis of 4T1, C-26 and A-20 tumor cells pulsed with 10⁻¹⁰ M (lo) or 10⁻⁶ M (hi) HA peptide at effector:target ratio 25:1. *P < 0.001 (two-sided unpaired t test) with respect to the respective lysis values by Tg or Tg^Pfp−/− CTL. (D) Percent decrease in cell number of medium− or human-TNF-α− and IFN-γ (100 U/ml, overnight)−pretreated human renal cancer lines ACHN, TK-10 and UO-31 incubated with pan-caspase inhibitor zVAD-FMK or control analog zFA-FMK (20 µM) one hour before the treatment with FasL (100 ng/ml). *P < 0.001 (two-sided unpaired t test) with respect to the control. Values are means ± SD (n = 3) from a representative of 6 (A), 2 (B, and C) and 3 (D) independent experiments.

Figure 6. Antigen density on tumor targets influences CTL response and effector cell degranulation

(A) Gating schematic for sorting of HA^negative, HA^low, HA^intermediate and HA^high cells in Renca-HA cells following staining with HA-specific H28-E23 Ab. (B and C) Percent specific lysis of sorted or unsorted Renca-HA cells by activated Tg CTL in the presence of medium, concanamycin A, FasL-neutralizing Ab, MFL4, or a combination of concanamycin A and MFL4. *P < 0.001 (two-sided unpaired t test) with respect to respective control groups. (D) Dot plots of TCR Vβ8.1 versus CD107a on Tg CTL following incubation with sorted Renca-HA tumor cells for 6 h in the presence of monensin. Numbers represent % positive cells. Plots representative of 3 independent experiments.