T cells bearing a chimeric antigen receptor against prostate-specific membrane antigen mediate vascular disruption and result in tumor regression

Abstract

Aberrant blood vessels enable tumor growth, provide a barrier to immune infiltration, and serve as a source of protumorigenic signals. Targeting tumor blood vessels for destruction, or tumor vascular disruption therapy, can therefore provide significant therapeutic benefit. Here, we describe the ability of chimeric antigen receptor (CAR)-bearing T cells to recognize human prostate-specific membrane antigen (hPSMA) on endothelial targets in vitro as well as in vivo. CAR T cells were generated using the anti-PSMA scFv, J591, and the intracellular signaling domains: CD3ζ, CD28, and/or CD137/4-1BB. We found that all anti-hPSMA CAR T cells recognized and eliminated PSMA(+) endothelial targets in vitro, regardless of the signaling domain. T cells bearing the third-generation anti-hPSMA CAR, P28BBζ, were able to recognize and kill primary human endothelial cells isolated from gynecologic cancers. In addition, the P28BBζ CAR T cells mediated regression of hPSMA-expressing vascular neoplasms in mice. Finally, in murine models of ovarian cancers populated by murine vessels expressing hPSMA, the P28BBζ CAR T cells were able to ablate PSMA(+) vessels, cause secondary depletion of tumor cells, and reduce tumor burden. Taken together, these results provide a strong rationale for the use of CAR T cells as agents of tumor vascular disruption, specifically those targeting PSMA. Cancer Immunol Res; 3(1); 68-84. ©2014 AACR.

Conflict of interest statement

Conflict of Interest: None

©2014 American Association for Cancer Research.

Figures

Figure 1. Design and characterization of an anti-vascular CAR

(A) Lentiviral cassette design for the J591-based CAR constructs: Pζ, P28ζ, PBBζ, P28BBζ and the FR28BBζ specificity control. (B) Schematic representation of the P28BBζ CAR. (C) Representative transduction efficiencies, as measured by eGFP reporter, for the CAR constructs in primary human T cells. Closed histograms depict untransduced cells; open histograms depict transduced T cells. (D) Human PSMA expression on the HMEC-1 and the engineered HMEC-1PSMA endothelial cell lines. PSMA was detected by humanized J591 Ab (open histogram) and compared with human IgG control (closed histogram). (E) Overlay of histograms comparing the proliferative capacity of the T-cell groups 5 days after co-culture with HMEC-1PSMA. (F) Quantitative comparison of proliferation between CAR constructs harboring different signaling domains. The geometric mean fluorescence intensity of Cellvue® Claret staining was measured for the CAR-positive T cells after 5 days in co-culture with either HMEC-1 or HMEC-1PSMA (E:T = 2:1). To normalize across multiple donors, the MFI of CellVue® staining on CAR+ T cells after co-culture with HMEC-1PSMA was divided by the MFI of the CAR+ T cells after co-culture with the HMEC-1. *P < 0.05; data are means ±SEM from three independent donors. (G) Cytolytic activity of the various T cells after 18 h co-culture with HMEC-1 or HMEC-1PSMA. Lysis was measured via luciferase assay (specific lysis = 100 – (luciferase signal treated/luciferase signal untreated × 100)). Data are means ±SD from triplicate cultures.

Figure 2. PSMA is expressed on the vasculature of primary and metastatic cancer.

br>(A) Representative staining for PSMA (top row), as well as PSMA and the endothelial marker, CD34 (bottom row). Sequential sections were cut and stained for either PSMA alone, or PSMA with CD34. Images in each column depict the same region within the same core. Original magnification, 200×; scale bar, 50 μm. (B) A heat map showing the presence or absence of PSMA on the vasculature of tumors taken from subjects with resected ovarian cancer (n = 15). (C) Representative tumor core staining for both PSMA (red) and CD34 (brown). Black arrows indicate CD34+ vessels that are negative for PSMA expression; red arrows indicate dual-positive (CD34+PSMA+) vessels. (D) Level of PSMA expression on tumor endothelial cells (CD45−CD31+) from three subjects with gynecologic cancer. Tumor endothelial cells were collected by CD45 depletion followed by CD31 enrichment. PSMA was detected by humanized J591 Ab (open histogram) and compared with human IgG control (closed histogram). (E) The percentage of PSMA-positive CD45−CD31+ endothelial cells remaining after overnight co-culture with CAR T cells (E:T = 1:1). Percentage = treated/untreated × 100. (F) IFNγ production by P28BBζ T cells after co-culture with CD31 enriched or depleted targets (E:T = 1:1). Culture supernatants were collected at 18 h and IFNγ was measured by ELISA.

Figure 3. P28BBζ CAR T cells target human tumor vasculature expressing PSMA.

br>(A) Antigen-specific destruction of HMEC-1PSMA microvessels after 48 h co-culture with P28BBζ T cells. HMEC-1 or HMEC-1PSMA endothelial cells (RFP, red) were seeded atop a Matrigel basement membrane and allowed to form microvessels for 8 h prior to co-culture with CAR-bearing T cells (eGFP, green) (E:T = 3:1). Images were taken at 0, 24, and 48 h. Original magnification, 40×; scale bar, 1 mm. (B) Immunofluorescence images of the HMEC-1PSMA endothelial cells at initiation (0 h) and termination (48 h) of co-culture with the P28BBζ T cells. Original magnification, 100×; scale bar, 500 μm. (C) Experimental design for the HMEC-1 pilot study. Mice were injected s.c. on each flank with 1.0×106 HMEC-1 (left flank) and 1.0×106 HMEC-1PSMA (right flank) endothelial cells. T cells were administered concurrently via i.v. injection. Treated mice received 5.0×106 CAR-positive T cells. (D) Representative dot plots from treated or control mice showing human CD45 and CAR (eGFP) expression. Matrigel plugs were retrieved upon sacrifice (11 d) and were digested to yield a single-cell suspension. Cells were subsequently stained for human CD45 and analyzed for eGFP expression using a flow cytometer. (E) CAR T cells as a percentage of the cells isolated from the Matrigel plugs. *P = 0.001, as determined by two-tailed Student’s t test. n = 2 mice in the FR28BBζ and P28BBζ; n = 1 mouse in the untreated (PBS) group. Data are means ±SEM. (F) CAR T cells as a percentage of total splenocytes. Spleens were isolated from mice upon sacrifice, mechanically disrupted, and analyzed for the presence of CAR T cells (eGFP). *P = 0.009, as determined by two-tailed Student’s t test. n = 2 mice in the FR28BBζ and P28BBζ; n = 1 mouse in the untreated (PBS) group. Data are means ±SEM.

Figure 4. P28BBζ CAR T cells eliminate PSMA+ vascular neoplasms.<

br>(A) Human PSMA expression on native (top row) and engineered (bottom row) murine endothelial cell lines. PSMA was detected by humanized J591 Ab (open histogram) and compared with human IgG control (closed histogram). (B) IFNγ production by P28BBζ T cells after co-culture with endothelial targets (E:T = 3:1). Culture supernatants were collected at 18 h and IFNγ was measured by ELISA. Representative donor shown; data are means ±SD of triplicate cultures. (C) Cytolytic activity of P28BBζ T cells after 18 h co-culture with endothelial cell targets. Cell lysis measured by chromium release. Representative donor shown; data are means ±SD of triplicate cultures. (D) Tumor injection schematic and experimental design for the MS1 hemangioma study. Mice were injected s.c. on each flank with 1.0×107 MS1 (left flank) and 1.0×107 MS1PSMA (right flank) endothelial cells. Hemangiomas were allowed to develop for 24 days prior to i.v. injection of T cells. (E) Tumor progression, as measured by luciferase luminescence, in mice receiving PBS, FR28BBζ, or P28BBζ T cells. Mice were given a single injection of 5.0×106 CAR-positive T cells (arrows) 24 days after inoculation with tumor. n = 5 mice per group; data are means ±SEM. *P < 0.05, as determined by two-tailed Student’s t test, for the P28BBζ treated group when compared to the FR28BBζ control group at the indicated time points. (F) Representative hemangiomas at the time of sacrifice (day 53). (G) Experimental design for the H5V hemangiosarcoma tumor study. Mice were injected i.v. with 5.0×105 H5VPSMA endothelial cells. Lung tumors were allowed to engraft for 3 weeks prior to administration of T cells. (H) Tumor progression, as measured by luciferase luminescence, in mice receiving PBS, FR28BBζ, or P28BBζ T cells. Mice were given three injections of 5.0×106 CAR-positive T cells beginning 21 days after tumor inoculation (arrows). n = 5 mice per group; data are means ±SEM. *P < 0.05, as determined by two-tailed Student’s t test, for the P28BBζ treated group when compared to the FR28BBζ control group at the indicated time points. (I) Survival curve for mice with H5VPSMA tumors. P28BBζ treated mice were sacrificed on day 108. P = 0.002, as determined by log-ranked (Mantel-Cox) test. n = 5 mice per group. (J) Tumor luminescence in representative mice before (day 20) and after T cell treatment (day 45).

Figure 5. The MS1/ID8VEGF tumor model closely mimics normal tumor physiology.<

br>(A) Tumor injection schematic and experimental design for the MS1/ID8VEGF characterization study. Mice were given a single s.c. injection of 1.0×106 ID8VEGF tumor cells either alone or with 1.5×107 MS1 or 1.5×107 MS1PSMA endothelial cells and monitored for 41 days. (B) Representative staining for CD31 (green) and PSMA (blue) from an MS1PSMA/ID8VEGF tumor (d 41). Original magnification, 200×; scale bar, 100 μm. (C) The percentage of MS1PSMA-derived endothelial cells (CD31+) in ID8 tumors at time of sacrifice. PSMA was detected by humanized J591 Ab. *P < 0.0001, as determined by two-tailed Student’s t test. n = 4 mice per group; data are means ±SEM. (D) The portion of CD31+ endothelial cells present in tumor digests reported as a percentage of the total number of live cells collected. n = 4 mice per group; data are means ±SEM. (E) Tumor progression in mice harboring ID8VEGF, MS1/ID8VEGF, or MS1PSMA/ID8VEGF tumors. Tumor volumes were calculated via caliper measurement. n = 4 mice per group; data are means ±SEM. (F) Representative images taken from the same mouse on day 17 and 41 showing eGFP radiance and firefly luminescence. (G,H) Correlation of endothelial cell signal (firefly luciferase, luminescence) with tumor signal (eGFP, radiant efficiency). Linear regression analysis was performed and the slope of best fit is shown (solid lines) for individual mice with either MS1/ID8VEGF (f) or MS1PSMA/ID8VEGF (g) tumors. (I,J) Statistical analysis describing the correlation of luminescence and radiant efficiency in mice with MS1/ID8VEGF (h) and MS1PSMA/ID8VEGF (i) tumors. Measurements were made 17–41 days after tumor inoculation. Luciferase and radiant efficiency measurements were plotted against one another (n = 7) and Pearson correlation coefficients were calculated for each mouse (r). Statistical significance was evaluated by two-tailed test (P).

Figure 6. CAR T cells ablate PSMA+ vasculature in solid tumors<

br>(A) Experimental design for the MS1/ID8VEGF tumor study. 6.7×105 ID8VEGF tumor cells were mixed with either 1.0×107 MS1 (left flank) or 1.0×107 MS1PSMA (right flank) endothelial cells and injected s.c. on the opposite flanks of each mouse. Tumors were allowed to engraft for 25 days before treatment with CAR T cells. (B) The persistence of the MS1 and MS1PSMA endothelial cells after treatment with PBS, FR28BBζ, or P28BBζ T cells. Mice were given three injections of 5.0×106 CAR-positive T cells beginning 25 days after tumor inoculation (arrows). The presence (or absence) of the MS1 and MS1PSMA endothelial cells was monitored via luciferase luminescence. n = 5 mice per group; data are means ±SEM. *P < 0.05, as determined by two-tailed Student’s t test, for the P28BBζ treated group when compared to the FR28BBζ control group at the indicated time points. (C) The percentage of PSMA positive tumor endothelial cells (CD45−CD31+) remaining after T cell administration to mice harboring MS1/ID8VEGF and MS1PSMA/ID8VEGF tumors. PSMA was detected by humanized J591 Ab. *P < 0.0001, as determined by two-tailed Student’s t test. n = 5 mice in the PBS and FR28BBζ groups; n = 2 mice in P28BBζ group. Data are means ±SEM. (D) The total percentage of tumor endothelial cells (CD45−CD31+) remaining in the MS1PSMA/ID8VEGF tumors after T cell administration. n = 5 mice in the PBS and FR28BBζ groups; n = 2 mice in P28BBζ group. Data are means ±SEM.

Figure 7. CAR T cells induce secondary loss of tumor cells and regression of solid tumors.

br>(A) Experimental design for the MS1/ID8VEGF tumor study. 6.7×105 ID8VEGF tumor cells were mixed with either 1.0×107 MS1 (left flank) or 1.0×107 MS1PSMA (right flank) endothelial cells and injected s.c. on opposite flanks of the same mice. Tumors were allowed to engraft for 22 days before T cell administration. (B) Tumor progression in mice receiving PBS, FR28BBζ, or P28BBζ T cells. Mice were given three injections of 5.0×106 CAR-positive T cells beginning 22 days after tumor inoculation (arrows). The impact of T cell administration on the ID8VEGF tumor cells was measured via luciferase luminescence (left column) and tumor volumes were calculated using caliper measurements (right column). n = 5 mice per group; data are means ±SEM. *P < 0.05, as determined by two-tailed Student’s t test, for the P28BBζ treated group when compared to the FR28BBζ control group at the indicated time points. (C) The fold-change in ID8VEGF tumor cell luminescence from the MS1/ID8VEGF and MS1PSMA/ID8VEGF tumors in mice receiving P28BBζ T cells. Luminescence measurements were normalized to those made prior to treatment (day 21). P values were determined by two-tailed Student’s t test and reflect the statistical significance of changes between day 24 and day 28. n = 5. (D) P28BBζ T cells do not elicit bystander killing of ID8VEGF when cultured in the presence of MS1PSMA. MS1PSMA cells were titrated into wells with ID8VEGF tumor cells and T cells at the described ratios and were cultured for 18 h. Lysis was measured via luciferase assay (specific lysis = 100 – (luciferase signal treated/luciferase signal untreated × 100)). Data are means ±SD from triplicate cultures.