Enhancement of PSMA-Directed CAR Adoptive Immunotherapy by PD-1/PD-L1 Blockade

Abstract

Chimeric antigen receptor (CAR) T cell therapy in hematologic malignancies has shown remarkable responses, but the same level of success has not been observed in solid tumors. A new prostate cancer model (Myc-CaP:PSMA(+)) and a second-generation anti-hPSMA human CAR T cells expressing a Click Beetle Red luciferase reporter) were used to study hPSMA targeting and assess CAR T cell trafficking and persistence by bioluminescence imaging (BLI). We investigated the antitumor efficacy of human CAR T cells targeting human prostate-specific membrane antigen (hPSMA), in the presence and absence of the target antigen; first alone and then combined with a monoclonal antibody targeting the human programmed death receptor 1 (anti-hPD1 mAb). PDL-1 expression was detected in Myc-CaP murine prostate tumors growing in immune competent FVB/N and immune-deficient SCID mice. Endogenous CD3⁺ T cells were restricted from the centers of Myc-CaP tumor nodules growing in FVB/N mice. Following anti-programmed cell death protein 1 (PD-1) treatment, the restriction of CD3⁺ T cells was reversed, and a tumor-treatment response was observed. Adoptive hPSMA-CAR T cell immunotherapy was enhanced when combined with PD-1 blockade, but the treatment response was of comparatively short duration, suggesting other immune modulation mechanisms exist and restrict CAR T cell targeting, function, and persistence in hPSMA expressing Myc-CaP tumors. Interestingly, an "inverse pattern" of CAR T cell BLI intensity was observed in control and test tumors, which suggests CAR T cells undergo changes leading to a loss of signal and/or number following hPSMA-specific activation. The lower BLI signal intensity in the hPSMA test tumors (compared with controls) is due in part to a decrease in T cell mitochondrial function following T cell activation, which may limit the intensity of the ATP-dependent Luciferin-luciferase bioluminescence signal.

Keywords: BLI; CAR T cells; anti-PD1; bioluminescence imaging; human PSMA; luciferase reporters; prostate cancer.

Figure 1

Comparison of Myc-CaP Wild-Type and Myc-CaP:hPSMA(+) Tumors Growing in NOD.SCID Il2rg^−/− (NSG) Mice (A) Mice received a s.c. injection of 1.0 × 10⁶ tumor cells; tumor growth was monitored and doubling time was calculated. H&E and immunofluorescence staining with MECA 32 (blood vessels are red) and Pimonidazole (hypoxia is green); scale bar, 100 μm. (B) Human PSMA was detected with PSMA (3E6 antibody, (Dako, Agilent Technologies).

Figure 2

Pre-targeting of Anti-hPSMA CAR-Directed Human T Cells to Myc-CaP:hPSMA(+) and Myc-CaP:hPSMA(−) Tumor Cells: Winn Assay (A) BLI images of tumor burden in different experimental sets: PSMA-directed CAR T cells mixed with (i) Myc-CaP:hPSMA(+) or (ii) Myc-CaP:hPSMA(−) tumor cells (the Winn assay), and (iii) a control set (Myc-CaP:hPSMA(+) tumor cells alone, no CAR T cells). (B)BLI images of anti-hPSMA CAR T cell in different treatment sets: PSMA-directed CAR T cells mixed with (i) Myc-CaP:hPSMA(+) or (ii) Myc-CaP:hPSMA(−) tumor cells (the Winn assay) and (iii) a control set (PSMA-directed CAR T cells alone, no tumor cells). Tumor growth profiles in NOD.SCID Il2rg−/− (NSG) mice following different treatment regimens. (C) Tumor growth was monitored by BLI using a Renilla Luciferase/coelenterazine reporter in tumor cells through day 15. (D) Tumor growth was monitored by caliper measurements. PSMA-directed CAR T cells significantly delayed the growth of hPSMA(+) tumors, compared with hPSMA(−) tumors and non-treated hPSMA(+) tumors; tumor volume measurements were significantly different on days 15 and 27. (E) Time-course of the BLI signal from anti-hPSMA CAR T cells in different experimental groups; BLI signal intensity from anti-hPSMA CAR T cells was lower in the hPSMA(+) tumors compared to hPSMA(−) tumors on days 12 and 27, and significantly lower than CAR T cells injected alone (in the absence of tumor cells) on days 7 and 12 but not on day 27. The number of mice per group, n = 5.

Figure 3

Effect of T Cell Processing on Their Metabolic Properties (A and B) Changes in ECAR (A) and OCR (B) following the processing of T cells from a post harvesting, “naive” status to anti-hPSMA CAR reporter bearing T cells. To assess glycolytic function, T cells were initially incubated in medium without glucose and ECAR is assessed. The first injection is a saturating concentration of glucose (25 mM), where glucose is taken up by the cells and catabolized through the glycolytic pathway to lactate, producing ATP and protons. (A) The extrusion of protons into the surrounding medium produces a rapid increase in ECAR. This glucose-induced response is reported as the rate of glycolysis (or glycolytic flux) percent increase over basal conditions. (B and C) In the mitochondrial respiration assay, basal OCR (B) is measured under full media growing conditions, followed by sequential addition of oligomycin, FCCP (C), maximal respiration) and antimycin A/rotenone with a measurement of changes in OCR. Each data point represents mean ± SD, n = 8. Data are normalized to 350,000 cells.

Figure 4

Limited Distribution of Endogenous T Cells in PD-L1-Positive Myc-CaP Tumors (A) PD-L1 expression in Myc-CaP:hPSMA(+) tumors growing in NOD.SCID Il2rg^−/− (NSG) mice on day 27 and MyC-CaP:WT in FVB/N mice on day 20 was studied by immunofluorescence imaging (IF). (B) CD3⁺ T cells are excluded from the core of Myc-CaP:WT tumor cell nests, and are present at high density in the tumor periphery; scale bar, 200 μm for the large image and 50 μm for the insets.

Figure 5

The Effect of the PD-1/PD-L1 Axis Blockade on Myc-CaP WT Tumors (A) The effect of anti-mPD1 mAb treatment compared to non-specific IgG treatment (control) on the growth of wild-type Myc-CaP tumors in FVB/N mice; n = 5 animals per group. (B) Immunofluorescence imaging (IF) of CD3⁺ T cells in the center of an IgG-treated (control) and an anti-mPD1 mAb-treated (test) tumor; scale bar, 100 μm. (C) Profile of CD3⁺ T cells fluorescence (from tumor periphery to center) of IgG (control) and anti-mPD1 (test) treated animals. (D) Comparison of CD3⁺ T cell density (number per square millimeter tumor area), and T cell size in control and test groups.

Figure 6

Anti-hPD1 mAb Treatment Enhances Adoptive Anti-hPSMA CAR-Directed T Cell Therapy (A) Bioluminescence imaging of Myc-CaP:hPSMA(+) and Myc-CaP:hPSMA(−) tumors (transduced with the Rluc reporter) in NOD.SCID Il2rg^−/− (NSG) mice to confirm tumor location. (B) Bioluminescence imaging of anti-hPSMA CAR T cells (transduced with the tdRFP/CBLuc reporter) 10 min after i.v. infusion and at 1 and 6 days. T cells trafficked largely to the lungs, although some localization to the tumor regions was visualized. (C) Plot of the mean tumor volume, comparing three treatment groups of mice (n = 5 per group). Treatments with anti-hPD1 mAb (test) or IgG mAb (control) are indicated by the black arrows; i.v. administration of anti-hPSMA CAR T cells is indicated by the red arrow. p value across groups is given for a one-tailed t test, *p < 0.05 versus control Myc-CaP:hPSMA(−) with anti-hPD1 mAb tumors and Myc-CaP:hPSMA(+) with IgG mAb. (D) A comparison of tumor growth in individual mice after the initiation of treatment; three groups are compared: anti-hPD1 mAb treatment of hPSMA(+) tumors (test), non-specific IgG treatment of hPSMA(+) tumors (control), and anti-hPD1 mAb treatment of hPSMA(−) tumors (control, no PSMA target).

Figure 7

Distribution of Anti-hPSMA CAR T Cells at 24 hr and Analyses of Tumor Treatment Response A comparison of three treatment groups: (A) anti-hPD1 mAb treatment of Myc-CaP:hPSMA(+) tumors (test), (B) non-specific IgG treatment of Myc-CaP:hPSMA(+) tumors (control), and (C) anti-hPD1 mAb treatment of Myc-CaP:hPSMA(−) tumors (control, no PSMA target) in combination with an i.v. infusion of 10 × 10⁶ anti-hPSMA CAR T cells. Scale bars are 1 mm for the second column and 200 μm for the third and fourth columns. TUNEL(+) cells per square millimeter were 54 (A), 29 (B), and 26 (C), respectively.