Research ArticleImmunologyOncology Open Access | 10.1172/JCI179014

Zhenyu Dai,1 Zheng Zhu,1 Zhiyao Li,1 Lei Tian,2,3 Kun-Yu Teng,1 Hanyu Chen,1 Li-Shu Wang,1 Jianying Zhang,4 Laleh Melstrom,5 Michael A. Caligiuri,1,6,7 and Jianhua Yu2,3

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Dai, Z. in: JCI | PubMed | Google Scholar

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Zhu, Z. in: JCI | PubMed | Google Scholar |

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Li, Z. in: JCI | PubMed | Google Scholar

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Tian, L. in: JCI | PubMed | Google Scholar

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Teng, K. in: JCI | PubMed | Google Scholar

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Chen, H. in: JCI | PubMed | Google Scholar |

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Wang, L. in: JCI | PubMed | Google Scholar

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Zhang, J. in: JCI | PubMed | Google Scholar

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Melstrom, L. in: JCI | PubMed | Google Scholar

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Caligiuri, M. in: JCI | PubMed | Google Scholar |

1Department of Hematology and Hematopoietic Cell Transplantation, City of Hope National Medical Center, Los Angeles, California, USA.

2Division of Hematology/Oncology, Department of Medicine, School of Medicine, and

3The Clemons Family Center for Transformative Cancer Research, Chao Family Comprehensive Cancer Center, University of California, Irvine, California, USA.

4Department of Computational and Quantitative Medicine,

5Division of Surgical Oncology, Department of Surgery, and

6Hematologic Malignancies Research Institute, City of Hope National Medical Center, Los Angeles, California, USA.

7City of Hope Comprehensive Cancer Center, Los Angeles, California, USA.

Address correspondence to: Jianhua Yu, 839 Medical Sciences Ct., Sprague Hall 212, Irvine, California, 92697, USA. Phone: 949.824.3926; Email: jianhuay@uci.edu. Or to: Michael A Caligiuri, 1500 East Duarte Rd., Duarte, California, 91010, USA. Phone: 626.218.6041; Email: mcaligiuri@coh.org.

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Find articles by Yu, J. in: JCI | PubMed | Google Scholar

Authorship note: ZD, ZZ, ZL, and LT contributed equally to this work.

Published April 15, 2025 - More info

Published in Volume 135, Issue 8 on April 15, 2025
J Clin Invest. 2025;135(8):e179014. https://doi.org/10.1172/JCI179014.
© 2025 Dai et al. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Published April 15, 2025 - Version history
Received: January 4, 2024; Accepted: February 17, 2025 View PDF Related article:

Abstract

Prostate stem cell antigen (PSCA) is highly and preferentially expressed on the surface of pancreatic ductal adenocarcinoma (PDAC) cells, raising the promise of tumor-selective cell-based immunotherapies. In this issue of the JCI, Dai et al. harness PSCA for the development of an off-the-shelf chimeric antigen receptor (CAR) invariant natural killer T (iNKT) cell–based treatment for PDAC. Through in vitro experiments and in vivo models, the authors demonstrate selectivity and therapeutic efficacy of PSCA CAR_sIL15 iNKT cells against both gemcitabine-sensitive and -resistant PDAC cells with comparable antitumor activity for freshly produced and frozen off-the-shelf PSCA CAR_sIL15 iNKT cells. This development opens another potential therapeutic option for pancreatic cancer.

Authors

Rachel Elizabeth Ann Fincham, Joe Poh Sheng Yeong, Hemant Mahendrakumar Kocher

× Abstract

Pancreatic ductal adenocarcinoma cancer (PDAC) continues to pose a significant health burden, with a 5-year survival rate of only 10%. Prostate stem cell antigen (PSCA) is highly expressed on the surface of tumor cells of most PDAC patients, with minimum expression in most normal tissues. Here, we generated cryopreserved, off-the-shelf, allogeneic PSCA chimeric antigen receptor (CAR) invariant NKT (iNKT) cells using human peripheral blood mononuclear cells as a cell source. In multiple in vitro and in vivo PDAC models, freshly manufactured PSCA CAR_sIL-15 iNKT cells and frozen-thawed, off-the-shelf PSCA CAR_sIL-15 iNKT cells demonstrate comparable efficacies, and both show remarkable suppression of PSCA-positive and gemcitabine-resistant PDAC. Importantly, off-the-shelf cryopreserved PSCA CAR_sIL-15 iNKT cells show equivalent efficacy when compared with PSCA CAR T cells using the same PSCA CAR and in the same PDAC model; however, PSCA CAR_sIL-15 iNKT cells do not appear to induce systemic toxicity or graft-versus-host disease, thus allowing for multiple infusions to control recurrent disease. Collectively, our study suggests that PSCA CAR_sIL-15 iNKT cells merit clinical investigation for PDAC patients exhibiting positive PSCA expression. The therapy could be given as a single agent or in combination with established therapeutic modalities for PDAC.

Introduction

Pancreatic ductal adenocarcinoma cancer (PDAC) accounts for more than 90% of pancreatic cancer cases and is one of the most aggressive solid tumors, characterized by a high rate of morbidity and mortality. It accounts for 7% of all cancer-related deaths, and the general 5-year survival rate for PDAC patients is just 10% (1). These alarming statistics raise concerns, as PDAC is predicted to become the second leading cause of cancer-related deaths by 2030 (2). The poor prognosis of PDAC can be attributed to several factors, including the absence of specific symptoms, leading to diagnosis at advanced stages with local and/or distant metastases. Additionally, PDAC cells exhibit a high resistance to standard chemotherapy, and the tumor microenvironment (TME) is characterized by highly immunosuppressive and metabolic challenges. To date, salvage chemotherapy regimens remain the primary option for treating advanced PDAC. Gemcitabine (2’,2’-difluorodeoxycytidine), a nucleoside analog, is the first-line intervention to treat advanced PDAC. However, despite its use, overall survival rates remain unsatisfactory, leaving few alternatives for patients who have failed gemcitabine-based therapy (3, 4). Therefore, it is imperative to prioritize the development of innovative and effective therapies to fight PDAC.

In recent years, substantial progress has been made with chimeric antigen receptor T (CAR-T) cell therapies, primarily for the treatment of lymphoid malignancies. These therapies have shown promise by inducing remissions and improving long-term relapse-free survival in B cell leukemia, lymphoma, and multiple myeloma. Unfortunately, the results of clinical trials indicate that CAR-T cell therapy has had limited success in treating solid tumors, including PDAC (5, 6). Several barriers must be overcome, such as the challenge of limited CAR-T cell infiltration at tumor sites and the immunosuppressive effects of the TME. These effects lead to impaired CAR-T cell proliferation, onset of exhaustion, and thus reduced efficacy. In addition, inserting CAR genes into polyclonal activated T lymphocytes results in cell products with high functional heterogeneity, which may compromise their antitumor potential and increase the risk of toxicity, including cytokine storm (7). Furthermore, the current application of CAR-T cells is primarily autologous, aimed at avoiding graft-versus-host disease (GvHD). This limits recurrent administration and widespread distribution while incurring high costs. The development of allogeneic, off-the-shelf CAR-T cell therapies is still in progress. Studies are ongoing to explore the potential of specific lymphocyte subsets, such as NK cells, γ δT, IL-9–secreting T cells, or NKT cells, which have been reported to be superior in terms of cell-mediated cytotoxicity, tumor infiltration, or desired cytokine production. These investigations represent a promising avenue for CAR-based immunotherapy (7–9).

Type I NKT cells (or invariant NKT [iNKT] cells) are an evolutionarily conserved sublineage of T cells that express the invariant TCR-α chain (Vα24-Jα18) (10). They possess unique characteristics that are intermediate between NK and T cells and exhibit an ability to recognize self- and microbial-derived glycolipids presented by the monomorphic human leukocyte antigen (HLA) class I–like molecule CD1d (10). Unlike HLA molecules, which exhibit genetic polymorphism and ubiquitous expression, CD1d gene expression is monomorphic and limited to specific cell types. This unique characteristic minimizes the potential for autologous or allogeneic iNKT cells to cause toxicity, regardless of HLA allele expression (11). iNKT cells offer distinct mechanistic advantages over bulk T cell populations when applied to CAR-based immunotherapy. They exhibit the capacity to traffic to solid tumors in response to chemokines produced by tumor cells, stromal cells, and tumor-associated macrophages (TAM) (12). This migration of iNKT cells into primary tumors correlates with better outcomes in various types of tumors (12, 13).

Several studies have demonstrated that donor-derived iNKT cells can effectively inhibit GvHD while preserving their antitumor activity (14, 15). In pediatric leukemia patients who received haploidentical transplants, the reconstitution of iNKT cells in peripheral blood has been associated with long-term remission (16). Furthermore, during the preparation of our manuscript, it was reported that autologous CAR-NKT cells have superior antitumor activity compared with CAR-T cells (17). Given these promising features, here we explore the potential of using iNKT cells for CAR modification in the treatment of PDAC.

Targeting the right antigen is critical to ensure the safety and effectiveness of CAR-based therapy. Prostate stem cell antigen (PSCA) is a glycosylphosphatidylinositol-linked cell-surface antigen that plays a critical role in promoting cell-cycle progression and boosting tumor cell proliferation. Moreover, the presence of metastasis and advanced clinical stages in prostate cancers is directly linked to the levels of PSCA protein (18, 19). Also, the overexpression of PSCA in PDAC begins during the early stages of malignant transformation and can be detected in 60%–80% of patients diagnosed with PDAC. In contrast, PDAC expression in normal pancreatic tissue is very low, providing a strong rationale for PSCA-targeting immunotherapy (20, 21). Previously, we reported that PSCA CAR NK cells expressing soluble IL-15 (sIL-15) showed significant and specific tumor-suppressive effects on PSCA+ PDAC, both in vitro and in vivo (22). These results underscore the promising potential of targeting PSCA to treat PDAC.

The objective of this study was to develop off-the-shelf human PSCA CAR iNKT cells expressing sIL-15 to enhance the antitumor functions of iNKT cells without inducing toxicity for sustained control of PDAC tumors. In addition, Liu et al. demonstrated that IL-15–expressing iNKT cells are not subjected to TAM inhibition or hypoxia, thereby significantly increasing their antimetastatic activity (23). Therefore, sIL-15 was also incorporated into our CAR construct (PSCA CAR_sIL-15). Our results demonstrate that PSCA CAR_sIL-15 iNKT cells exhibit enduring antitumor efficacy in vitro and in vivo without causing notable toxicity in multiple models, including the orthotopic PDAC model and the metastatic PDAC model. Moreover, we observed that PSCA CAR_sIL-15 iNKT cells efficiently kill PDAC cells that had developed resistance to first-line standard chemotherapy with gemcitabine in vitro and in vivo. Off-the-shelf PSCA CAR_sIL-15 iNKT cells were further validated to have comparable antitumor capabilities even after undergoing a freeze-thaw cycle without risk of GvHD when compared with PSCA CAR_sIL-15 T cells generated from the same donor and to fresh PSCA CAR_sIL-15 iNKT cells. These preclinical evaluations provide a robust foundation for exploring the clinical applications of allogeneic off-the-shelf PSCA CAR_sIL-15 iNKT cells.

Results

PSCA CAR_sIL-15 iNKT cells demonstrate excellent in vitro expansion and have low expression of exhaustion markers. Human primary iNKT cells isolated and expanded from human PBMCs were engineered to express soluble IL-15 alone (sIL-15 iNKT) or both PSCA CAR and sIL-15 (PSCA CAR_sIL-15 iNKT; Figure 1A). Our constructs also included a truncated EGFR (tEGFR) as a marker to detect successful transduction and as a safety switch, the latter of which allows for in vivo depletion of PSCA CAR_sIL-15 iNKT cells by administering a clinical-grade anti-EGFR antibody, cetuximab (22). The sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells showed high (97%) iNKT purity, identified by the marker of TCR Vα24-Jα18. Furthermore, the overall transduction efficiencies, around 42%, were similar for both sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells, as assessed by EGFR expression 3 days after the cells underwent transduction. We also detected the basal apoptosis levels of sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells, showing that both populations exhibited very low basal levels (~5%) of apoptosis (Figure 1, B–D). These data demonstrate our newly generated PSCA CAR_sIL-15 iNKT cells show high purity, have a high rate of CAR transduction, and are generally healthy several days after transduction. After in vitro culture with α-galactosylceramide (α-GalCer), both sIL-15 and PSCA CAR_sIL-15 iNKT cells can be expanded more than 5,000-fold (Figure 1E) and the expanded iNKT cells display low surface density expression of exhaustion markers LAG-3, PD-1, and TIM-3 (Figure 1F). These data demonstrate the successful engineering, manufacturing, and expansion of PSCA CAR_sIL-15 iNKT cells.

Figure 1

iNKT cells expressing the PSCA CAR_sIL-15 construct demonstrate excellent in vitro expansion and show low surface expression of exhaustion markers. (A) Schematic diagrams of the clinical grade vectors. tEGFR was included as both a detection marker and a safety switch, allowing for in vivo iNKT cell depletion by administering an anti-EGFR antibody. (B) Representative flow cytometric analysis of PSCA CAR_sIL-15 iNKT cells and sIL-15 iNKT cells shows the proportion of CD3 and iNKT (TCR Vα24-Jα18) expression 2 days after transduction. sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells show high NKT purity of approximately 97%. The experiment was conducted with 3 donors with similar results. (C) The transduction ratio of PSCA CAR_sIL-15/sIL-15 iNKT cells was detected by measuring tEGFR expression 2 days after transduction and analyzed by flow cytometry. The transduction efficiencies, approximately 42%, were similar in both sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells. The experiment was conducted with 3 donors with similar results. SSC, side scatter. (D) The level of apoptosis of sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells was measured by the coexpression of annexin V and 7-AAD 2 days after transduction by flow cytometry. Both sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells exhibited very low levels of apoptosis. (E) Quantification of PSCA CAR_sIL-15 and sIL-15 iNKT cell fold expansion following 12 days of secondary expansion (mean ± SD, n = 3). Not significant (Student’s t test). Both sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells can be expanded more than 5,000-fold. (F) Surface expression of exhaustion markers LAG-3, PD-1, and TIM-3 on sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells, measured by flow cytometry. The results are displayed as mean ± SD (n = 3). Not significant (2-way ANOVA).

PSCA CAR_sIL-15 iNKT cells exhibit potent antitumor activity against human PDAC cell lines in vitro. Previously, we reported that PSCA is highly expressed in primary PDAC tumor samples, and this expression was correlated with a poor prognosis for patients (22). In our subsequent functional validation to assess the antitumor activity of PSCA CAR_sIL-15 iNKT cells, we first measured the expression levels of PSCA in 5 different human PDAC cell lines using flow cytometry. We found that Capan-1, MIA PaCa-2, and Aspc-1 cells highly expressed PSCA, while Panc-1 and BxPC-3 cells had low PSCA expression (Figure 2A). To evaluate the anti-PDAC capability of PSCA CAR_sIL-15 iNKT cells in vitro, PSCA CAR_sIL-15 iNKT cells or sIL-15 iNKT cells were cocultured with different PDAC cell lines at an effector: target (E:T) ratio of 1:1 for 6 hours. Compared with PSCA CAR_sIL-15 iNKT cells alone, PSCA CAR_sIL-15 iNKT cells expressed higher levels of the iNKT cell-activation markers CD69 and CD25 after coculturing with PSCA+ Capan-1 cells and PSCA+ MIA PaCa-2 cells, but not when cocultured with PSCA– BxPC-3 cells (Figure 2, B and C). In contrast, sIL-15 iNKT cells did not exhibit this phenotype (Figure 2, B and C).

Figure 2

PSCA CAR_sIL-15 iNKT cells demonstrate PSCA+ PDAC cell-specific activation. (A) Surface density expression of PSCA on human PDAC cell lines was measured by mean fluorescent intensity (MFI) using flow cytometry. Capan-1, MIA PaCa-2, and Aspc-1 cells highly expressed PSCA, while Panc-1 and BxPC-3 cells had low PSCA expression. (B) Summary of percentages of sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells positive for CD69 and CD25 following a 24-hour coincubation with Capan-1, MIA PaCa-2, or BxPC-3 (gated on iNKT cells). Data are presented as mean ± SD (n = 3). (C) Representative flow cytometric analysis shows the expression of CD69 and CD25 on sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells after coincubation with target cells. (D) Representative flow cytometric analysis (left) and summary graph (right) show CD107a expression on sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells after coincubation with target cells. (E) Representative flow cytometric analysis (left) and summary graph (right) show TNF-α expression in sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells after coincubation with target cells. (F) Representative flow cytometric analysis (left) and summary graph (right) show IFN-γ expression of sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells after coincubation with target cells. All experiments were repeated using ≥ 3 donors and presented as mean ± SD (B, D, E, and F). Statistical analyses were performed using 1-way ANOVA, with P values corrected for multiple comparisons using the Holm-Šídák method.*P < 0.05; **P < 0.01; ***P < 0.001.

Degranulation is a prerequisite for immune cell perforin-mediated killing. PSCA CAR_sIL-15 iNKT cells upregulated expression of CD107a (a surrogate marker for degranulation) when cocultured with PSCA+ PDAC cells but not with PSCA– PDAC cells. In contrast, PSCA+ PDAC cells and PSCA– PDAC cells did not activate sIL-15 iNKT cells (Figure 2D). Moreover, PSCA CAR_sIL-15 iNKT cells produced more proinflammatory cytokines, TNF-α and IFN-γ, in response to PSCA+ cells, compared with PSCA– PDAC cells and with sIL-15 iNKT cells (Figure 2, E and F, respectively).

Next, the cytolytic function of PSCA CAR_sIL-15 iNKT cells was assessed using real-time cell analysis (RTCA). PSCA CAR_sIL-15 iNKT cells demonstrated robust killing activity against PSCA+ tumor cell lines, including Capan-1 cells, MIA PaCa-2 cells, and Aspc-1 cells, in contrast with sIL-15 iNKT cells. However, neither sIL-15 iNKT cells nor PSCA CAR_sIL-15 iNKT cells exhibited cytotoxicity against the PSCA– cell line Panc-1 and both showed relatively modest but equivalent killing against the PSCA– cell line BxPC-3 (Figure 3, A and B, and Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/JCI179014DS1). Collectively, our in vitro data indicate that the difference in activation, degranulation, cytokine secretion, and cytolysis between PSCA CAR_sIL-15 iNKT cells and sIL-15 iNKT cells is specific for the expression of the PSCA CAR on iNKT cells as well as for PSCA expression on the tumor cell lines.

Figure 3

PSCA CAR_sIL-15 iNKT cells demonstrate potent and specific cytotoxicity against human PSCA+ PDAC cell lines in vitro. (A) RTCA results measuring cytotoxicity of sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells against PSCA+ Capan-1, PSCA+ MIA Paca-2, and PSCA+ Aspc-1 or PSCA– Panc-1 and PSCA– BxPC-3 tumor cells at an E:T ratio of 1:1. Experiments were repeated with 3 donors. (B) Representative microscopic images show the killing as noted in A after 90 hours of coincubation. Experiments were repeated with 3 donors. (C) Freshly isolated human primary NK cells were cultured in the presence of supernatants from nontransduced iNKT (NT supernatant), sIL-15 iNKT, PSCA CAR iNKT, or PSCA CAR_sIL-15 cells for 2 days. Capan-1 cells were labeled with 51Cr and served as target cells. The labeled target cells were added to the cultured NK cells in the presence of respective supernatants for an additional 12 hours. The cytotoxicity levels were measured by 51Cr release assay. n = 4 donors. NT versus PSCA, P = 0.2218; NT versus sIL-15, P < 0.0001; PSCA versus PSCA CAR sIL-15, P < 0.0001; PSCA versus sIL-15, P < 0.0001; sIL-15 versus PSCA s15, P = 0.1157. Statistical analyses were performed by 1-way ANOVA with P values corrected for multiple comparisons by Bonferroni’s method.

To explore the function of sIL-15 within the construct, we measured the bystander effect of IL-15, secreted by engineered iNKT cells, on nontransduced, freshly isolated NK cells and T cells. For this purpose, the supernatants of nontransduced iNKT cells (NT), sIL-15 iNKT cells, PSCA CAR iNKT cells, and PSCA CAR sIL-15 iNKT cells were collected. Freshly isolated human NK cells or T cells were cultured in these 4 different supernatants for 2 days and the levels of NK cell and T cell cytotoxicity were measured by 51Cr release assay using Capan-1 cells as target cells. Compared with supernatants from the NT group, the supernatants of PSCA CAR iNKT did not activate T cells and NK cells, while the supernatants of sIL-15 iNKT cells and PSCA CAR_sIL-15 iNKT cells resulted in significantly higher levels of NK and T cell activation when compared with supernatants from the NT or PSCA (Figure 3C and Supplemental Figure 2). These data suggest that the IL-15 from these supernatants can activate NK cells and T cells via a bystander effect. Thus, IL-15 produced by our engineered NKT cells can be released into the extracellular milieu with the potential for a local effect on cell activation.

PSCA CAR_sIL-15 iNKT cells show superior therapeutic activity in 2 in vivo PDAC metastasis models without notable toxicity. To validate the in vitro antitumor effectiveness of PSCA CAR_sIL-15 iNKT cells noted above, we established 2 PDAC metastatic orthotopic models for in vivo study. Previously, we demonstrated that a combination of i.p. and i.v. injections of PSCA CAR NK cells killed PSCA+ Capan-1 cells in the pancreas and those that metastasized to the liver and lung (22). Therefore, we combined i.p. and i.v. injections of PSCA CAR_sIL-15 iNKT cells to treat PDAC tumor-bearing mice. The procedure for establishing the Capan-1 cell–based metastatic PDAC model and the treatment are depicted in Figure 4A. Briefly, 2 × 105 PSCA+ Capan-1 cells expressing the firefly luciferase (FFL) gene were i.p. injected into NOD-scid IL2Rgammanull (NSG) mice on day 0. Three days after tumor implantation, mice were treated with a single dose of 3 × 106 PSCA CAR_sIL-15 iNKT cells by i.p. injection combined with 1.5 × 106 PSCA CAR_sIL-15 iNKT cells by i.v. injection. Saline and sIL-15 iNKT cells were administered through the same routes as control. Progression of tumors was monitored by bioluminescence (BLI) until week 9, and survival data were recorded. Compared with the 2 control groups, PSCA CAR_sIL-15 iNKT cells significantly inhibited the progression of metastatic PDAC and significantly prolonged the survival of the tumor-bearing mice, reaching 100% survival by day 80, while allowing for maintenance of body weight compared with control mice in vivo (Figure 4, B–D, and Supplemental Figure 3). We also compared the antitumor effects in vivo of i.p. plus i.v., i.p. alone, and i.v. alone. For this purpose, mice were injected with 5 × 105 Capan-1-luc cells on day 1. On day 7, mice were randomly divided into 4 groups: group 1, PBS; group 2, i.p. injection of 4 × 106 PSCA CAR_sIL-15 iNKT cells per mouse; group 3, i.v. injection of 4 × 106 PSCA CAR_sIL-15 iNKT cells per mouse; group 4, i.p. injection of 2 × 106 plus i.v. injection of 2 × 106 of PSCA CAR_sIL-15 iNKT cells per mouse. Tumor burden and survival were monitored as above. Group 2 (i.p. injection) showed significantly greater survival compared with group 3 (i.v. injection). Furthermore, there was no difference in survival between group 2 and group 4 (the median survival day of group 2 was 71 days versus the median survival of 77.5 days for group 4) (Supplemental Figure 4).

Figure 4

In vivo assessment of PSCA CAR_sIL-15 iNKT cells. (A) Treatment schema for i.p. plus i.v. injection of PSCA CAR_sIL-15 iNKT cells in a human metastatic PDAC model established by i.p. injection of PSCA+ Capan-1_luc cells into NSG mice. The image was created in BioRender. (B) Tumor growth, directly correlated with color intensity, was monitored by BLI until week 9. (C) Graphical depiction of BLI from B up to day 32. The results are displayed as mean ± SD (n = 4). *P < 0.05; **P < 0.01 (2-way ANOVA). (D) Overall Kaplan–Meier survival curve. **P < 0.01 (log-rank test, n = 5). Compared with the 2 control groups, PSCA CAR_sIL-15 iNKT cells significantly inhibited the progression of metastatic PDAC and prolonged the survival of the tumor-bearing mice. (E) Schematic diagram as in A but with MIA PaCa-2_luc PDAC tumor cells. (F) Representative images of the pancreas and liver from each treatment group at the endpoint of the in vivo experiment. Red arrows mark metastatic tumors in the liver. PSCA CAR_sIL-15 iNKT cells demonstrated strong therapeutic effects, as evidenced by their ability to kill MIA PaCa-2 cells in the pancreas and the liver. (G) Summary of relative fold change in BLI over 15 days as shown in H. The results are displayed as mean ± SD (n = 5). **P < 0.01; ***P < 0.001 (2-way ANOVA). (H) The growth of the tumor was monitored by BLI imaging until week 6. (I) Overall Kaplan-Meier survival curve. ***P < 0.001 (log-rank test, n = 5). PSCA CAR_sIL-15 iNKT cells completely eradicated PDAC in vivo. sIL-15 iNKT cells were inferior to PSCA CAR_sIL-15 iNKT cells but also exhibited some degree of efficacy in delaying tumor progression in mice bearing MIA PaCa-2 cells. (J) Assessment of blood cells and HGB on day 15 after PDAC cell transplantation (12 days after treatment with PBS, sIL-15 iNKT cells, or PSCA CAR_sIL-15 iNKT cells). Values represent mean ± SD (n = 5). Not significant (2-way ANOVA).

We constructed another metastatic PDAC model using the PSCA+ cell line MIA PaCa-2 cells to confirm the therapeutic capability of PSCA CAR_sIL-15 iNKT cells depicted in Figure 4E. In this model, PSCA CAR_sIL-15 iNKT cells also demonstrated a strong therapeutic effect, as evidenced by their ability to kill MIA PaCa-2 cells in the pancreas and the liver (Figure 4F), completely eradicate PDAC in vivo, maintain remission, and significantly extend survival compared with the untreated and sIL-15 iNKT–treated groups (Figure 4, G–I). Furthermore, a hematological analysis of blood samples from treated mice revealed that PSCA CAR_sIL-15 iNKT cell treatment had no significant adverse effect on blood cell counts and hemoglobin (HGB) levels when compared with both the untreated group and the sIL-15 iNKT cell group (Figure 4J).

PSCA CAR_sIL-15 iNKT cells exert superior therapeutic activity in an orthotopic PDAC model without notable toxicity. We then conducted an extensive evaluation of PSCA CAR_sIL-15 iNKT cells in an orthotopic PDAC model (Figure 5A). This model showed locoregional cancer cell spread and liver metastasis, thus mimicking the condition of PDAC patients. To establish the model, 2 × 105 FFL-expressing MIA PaCa-2 cells were injected intrapancreatically on day 0. On day 3, mice received a single dose of 3 × 106 PSCA CAR_sIL-15 iNKT cells by i.p. injection and 1.5 × 106 PSCA CAR_sIL-15 iNKT cells by i.v. injection. In this orthotopic PDAC model, PSCA CAR_sIL-15 iNKT cells efficiently eliminated macroscopic evidence of MIA PaCa-2 cells in situ within the pancreas and decreased metastatic lesion formation in the liver (Figure 5B). Treatment with PSCA CAR_sIL-15 iNKT cells resulted in complete macroscopic clearance of orthotopic tumors and reached 100% survival at 80 days (Figure 5, C–E) without affecting peripheral blood counts and HGB compared with the untreated group and sIL-15 iNKT cell treatment group (Figure 5F).

Figure 5

PSCA CAR_sIL-15 iNKT cells eliminate human PDAC cells in an orthotopic tumor model and maintain long-term tumor-free survival. (A) Schematic diagram of treatment with PSCA CAR_sIL-15 iNKT cells in a human orthotopic PDAC model established by i.p. injection of MIA PaCa-2_luc cells into NSG mice. The image was created in BioRender. (B) Representative images of the pancreas and the liver of each group in the MIA PaCa-2-transplanting PDAC mouse model at the endpoint of the in vivo experiments. Red arrows mark metastatic tumors in the liver. In this orthotopic PDAC model, PSCA CAR_sIL-15 iNKT cells efficiently eliminated carcinoma in situ within the pancreas and decreased metastatic lesion formation in the liver. (C) Summary statistical data of mouse tumor burden changes of each treatment group. The results are displayed as mean ± SD. **P < 0.01; ****P < 0.0001 (2-way ANOVA). n = 7 for the untreated and PSCA CAR_sIL-15 groups. n = 6 for the sIL-15 group. (D) The growth of the tumor was monitored by BLI imaging until week 8. (E) Overall Kaplan-Meier survival curve. ***P < 0.001 (log-rank test). n = 7 for the untreated and PSCA CAR_sIL-15 groups. n = 6 for the sIL-15 group. Treatment with PSCA CAR_sIL-15 iNKT cells resulted in complete clearance of orthotopic tumors and reached 100% survival. (F) Assessment of blood cell populations on day 15 after PDAC cell transplantation (12 days after iNKT cell treatment). Peripheral blood counts and HGB in the PSCA CAR_sIL-15 iNKT group were not changed compared with the untreated group and sIL-15 iNKT cell treatment group. Values represent mean ± SD (n = 7 for the untreated and PSCA CAR_sIL-15 groups. n = 4 for the sIL-15 group.). 2-way ANOVA.

Gemcitabine-resistant PDAC can be overcome by PSCA CAR_sIL-15 iNKT cells. Gemcitabine-based therapy is a standard first-line therapy for patients with advanced PDAC (24). However, invariable tumor recurrence after gemcitabine results in relapse and progression, allowing for reduced patient survival (25). Therefore, we next determined whether PSCA CAR_sIL-15 iNKT cells can kill gemcitabine-resistant (GR) PDAC in vitro and in vivo. In this scenario, we generated 2 GR cell lines (Capan-1 GR and MIA Paca-2 GR) by exposing parental Capan-1 and MIA PaCa-2 cells to escalating concentrations of gemcitabine for 9 months, as previously reported (26). The GR cell lines, Capan-1 GR and MIA PaCa-2 GR, showed a modest increase in the expression of PSCA compared w