Engineering T Cells to Express Tumoricidal MDA-7/IL24 Enhances Cancer Immunotherapy

Melanoma differentiation associated gene-7/IL24 (mda-7/IL24) is a unique IL10 family gene member that demonstrates profound anticancer, antiproliferative, and cytotoxic effects in a broad spectrum of histologically distinct human cancers without harming normal cells (1, 2). Other attributes of this therapeutic gene product include “bystander” antitumor activity (3), antiangiogenesis (4), immune augmentation (5), inhibition of cancer-initiating/stem cells (6), and synergy with other cancer treatment modalities, for example, radiotherapy or chemotherapy (7, 8). A phase I clinical trial in advanced cancers has established the safety and therapeutic efficacy of MDA-7/IL24, when administered intratumorally by a nonreplication-competent adenovirus (9), further supporting the potential for translation of MDA-7/IL24 into the clinic for therapy and its rational integration with other treatment regimens.

Immunotherapy based on the adoptive transfer of naturally occurring (e.g., tumor-infiltrating lymphocytes or TILs) or genetically engineered T cells (e.g., chimeric antigen receptor or CAR) can mediate cancer regression in patients with advanced disease (10, 11). While CAR T-cell therapies have shown remarkable success in the treatment of certain hematologic malignancies, such as relapsed or refractory lymphoblastic leukemia and non-Hodgkin lymphoma, their therapeutic effect in solid tumors has been limited. Accumulating clinical and preclinical evidence suggests that this partially results from heterogeneous expression of tumor antigens, immune evasion caused by immuno-editing, for example, selection of nonimmunogenic or antigen loss cancer cell variants (12, 13), a lack of sufficient specificity or numbers of TILs (14), and the impact of an immunosuppressive tumor microenvironment (TME) that limits T-cell function and persistence (15, 16). In fact, immunotherapeutic targeting of tumor-associated antigens can promote the recurrence of the antigen-negative tumor variants (17, 18). Given the well-documented highly selective tumoricidal activity of MDA-7/IL24 that is independent of antigen recognition, strategically equipping T lymphocytes with MDA7/IL24-mediated multivalent cancer-killing features may help overcome the barriers to T-cell–based immunotherapy.

In this study, we rigorously evaluated the therapeutic activity of tumor-reactive or antigen-specific T cells that were genetically engineered to carry human MDA-7/IL24 using multiple clinically relevant mouse syngenic tumor models, including a spontaneous prostate cancer in a transgenic strain of mice. We show that T lymphocytes with a high specificity of tumor recognition can serve as an excellent platform to deliver MDA-7/IL24 to the tumor sites. More importantly, MDA-7/IL24 engineering confered T cells with superior anticancer activity in eradicating established tumors and metastases, which is attributed to synergistic tumor destruction by tumor-reactive or antigen-specific T cells and therapeutic MDA-7/IL24 as well as a previously undocumented role of MDA-7/IL24 of promoting expansion and persistence of T cells in the TME.

OT-I mice, Pmel mice, B6.PL-Thy1a/CyJ (Thy1.1) mice were purchased from the Jackson Laboratory. Male C57BL/6 mice were purchased from Envigo. The Hi-Myc transgenic strain mice (19) were kindly provided by Dr. Charles Sawyers (Memorial Sloan Kettering Cancer Center, New York, NY). All experimental procedures were conducted according to protocols approved by the VCU Institutional Animal Care and Use Committee.

Mouse prostate tumor RM1 cells, mouse melanoma B16 cells, and 293T cells were purchased from ATCC. OVA or luciferase-expressing RM1 cells were generated in our laboratory (20, 21). Cells were maintained in DMEM supplemented with 10% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin (Hyclone). All cell lines were routinely tested for Mycoplasma contamination using a PCR-based Mycoplasma Detection Kit (ATCC).

OVA_257–264 (SIINFEKL) and gp100_25–33 (KVPRNQDWL) peptides were purchased from AnaSpec Inc. Recombinant human IL2, IL7, and IL15 cytokines were kindly provided by the NCI (Rockville, MD). For FACS analysis, mouse mAbs to CD4 (GK1.5), CD8a (53–6.7), CD3 (HL3), IFNγ (XMG1.2), CD25 (PC61), CD90.1 (OX-7), isotype control rat IgG2b (RTK4530), and IgG1 (RTK2071) were purchased from BioLegend; FoxP3 antibody (FJK-16s) was purchased from eBioscience. For Western blotting analysis, anti-human MDA-7/IL24 antibody was purchased from GenHunter; anti-mouse MDA-7/IL24 antibodies were purchased from R&D Systems; phospho-STAT1 (Y701), phospho-STAT3 (Y705), total STAT1, and total STAT3 antibodies were purchased from Cell Signaling Technology. Cucurbitacin I was purchased from Sigma Aldrich. BrdU Flow Kit was purchased from BD Biosciences. Polybrene was purchased from Sigma Aldrich. CytoTox 96 Non-Radioactive Cytotoxicity Assay kits were purchased from Promega. Cell Counting Kits 8 (CCK8) were purchased from Abcam.

Ex vivo expansion of tumor-sensitized T cells was performed as previously described with modifications (22, 23). Briefly, spleen and lymph nodes were collected from tumor-bearing mice and dispersed into a single cell suspension in complete RPMI1640 media at 2 × 10⁶ cells/mL. Cells were then stimulated with Bryostatin 1 (5 nmol/L) and Ionomycin (1 μmol/L) plus IL2 (80 IU/mL) for 16–18 hours followed by culture in RPMI1640 media containing IL7 and IL15 (10 ng/mL each) at 1–2×10⁶ cells/mL for 5 days. For ex vivo expansion of antigen-specific OT-I or Pmel T cells, cells were stimulated with OVA_257–264 peptide or gp100_25–33 peptide (1 μg/mL), respectively, for 16–18 hours, followed by culture in RPMI1640 media containing IL7 and IL15 (10 ng/mL each) at 1–2×10⁶ cells/mL for 5 days.

The pLenti CMV Puro DEST (w118–1) vector (Addgene plasmid # 17452) encoding human MDA-7/IL24 or the empty vector together with pMD2.G (Addgene plasmid # 12259) and psPAX2 plasmid (Addgene plasmid # 12260) were cotransfected into 293T cells using Fugene HD (Promega). Lentiviral supernatants containing control lentivirus (LV-vec) or lentivirus expressing MDA-7/IL24 (LV-mda-7) were collected 48 and 72 hours after transfection. Day 5 expanded T cells were transduced with LV-vec or LV-mda-7 (2–3 × 10⁶ cells/mL in 6-well plate at multiplicity of infection of 5 in the presence of 6 μg/mL polybrene) under centrifugation at 2,000 × g for 90 minutes at 32°C. Transduced T cells were cultured in the presence of IL7 and IL15 overnight before collection for studies.

Tumors were established by subcutaneous (s.c.) injection of 2×10⁵ tumor cells into the right dorsal flank. Mice were randomly assigned to control and treatment groups and received whole body irradiation (WBI, 4.5 Gy) followed by intravenous (i.v.) injection of tumor-sensitized or antigen-specific T cells (10⁷ in 200 μL PBS) at the indicated time points. Tumor growth was monitored in a blinded manner by measuring the perpendicular diameters of tumors. Animals were euthanized when tumors reached the size of approximately 16-mm in diameter or when mice had lost 10% of their total body weight. To establish experimental lung metastases, mice were intravenously injected with 1×10⁵ RM1-Luc cells on day 0. Mice were treated with T cells on days 2 and 8. Lung metastases were monitored using IVIS Spectrum imaging system (PerkinElmer). For analysis of leukocyte infiltration into the tumor microenvironment, tumors, or lungs were digested with collagenase D (1 mg/mL) and DNase I (20 μg/mL) to prepare single-cell suspensions as described previously (24). Tumor draining lymph nodes (DLN) and spleens were collected at indicated time points for immune analyses.

For treatment of Hi-Myc transgenic mice, 4-month-old male mice received WBI and engineered tumor-sensitized T cells (1×10⁷) that were derived from prostate cancer-bearing 6-month-old Hi-Myc mice. Treatments continued weekly for a total of 4 times. Animals were euthanized at 6 months old and prostate tissues were collected for analyses.

Bromodeoxyuridine (BrdU) incorporation into T cells was assessed by following intraperitoneal (i.p.) injection of 1 mg BrdU 12 hours before collection of lung tissues. Cell suspensions were fixed, permeabilized, and stained with FITC-conjugated anti-BrdU antibody according to the manufacturer's protocol. Frequency of BrdU⁺ cells was determined by flow cytometric analysis using a FACSCalibur flow cytometer (BD Biosciences).

Total RNA was extracted using TRIzol Reagent (Thermo Fisher Scientific). Reverse transcription and real-time reverse transcription-PCR were conducted using primers and carboxy fluorescein (FAM)-labeled probe sets (Thermo Fisher Scientific; ref. 25). Gene expression was quantified relative to the expression of β-actin. An analysis of relative gene expression data was performed using the 2^ΔΔC_t method.

Statistical analysis was performed using Sigmaplot 14.0 or GraphPad Prism 5. Data are expressed as mean ± SD values. Statistical differences between groups within experiments were determined by the two-way repeated measures ANOVA test or the Student t test. Survival mice in the experimental groups were compared using the log-rank test. Values of P < 0.05 were considered to be statistically significant.

Tumor-reactive T cells derived from RM1-OVA prostate tumor-bearing mice were expanded ex vivo using a modified protocol involving bryostatin/ionomycin/IL2 and common γ-chain cytokines (IL7/IL15; refs. 23, 26). Consistent with our previous reports, use of γ-chain cytokines resulted in a marked increase in the numbers of CD8⁺ and CD4⁺ T cells (23- and 9-fold, respectively), which preferentially exhibited a central memory phenotype (CD44⁺CD62L^high; Supplementary Fig. S1A–S1D). ELISpot analysis showed that the frequency of OVA-specific CD8⁺ T cells also increased during T-cell expansion (Supplementary Fig. S1E). To genetically engineer T cells to produce human MDA-7/IL24, we constructed a human mda-7/IL24-encoding lentivirus, that is, LV-mda-7 (Supplementary Fig. S2A and S2B). Expression and secretion of MDA-7/IL24 protein was confirmed by immunoblotting analysis of media of T-cell cultures (Fig. 1A; Supplementary Fig. S2C). The concentration of MDA-7/IL24 in the culture media 72 hours posttransduction was estimated to be 300 ng/mL. The ratio of CD8⁺ to CD4⁺ T cells and the central memory phenotype of T cells are comparable in T cells infected with LV-mda-7 or an empty virus, that is, LV-vec (Fig. 1B). Infection with LV-mda-7 did not significantly affect T-cell viability (Supplementary Fig. S2D). Transduction of T cells with either LV-vec or LV-mda-7 did not trigger the activation of T cells (Supplementary Fig. S2E).

We next examined the impact of cellular production of human MDA-7/IL24 on effector function of T cells. Cytolytic assays showed that RM1 prostate tumor or B16 melanoma-sensitized T cells, upon engineering to produce human MDA-7/IL24 (referred to T-mda-7 unless specifically indicated), were more effective than mock-transduced T-cell counterparts (T-vec) in killing corresponding cancer cell targets (Fig. 1C; Supplementary Fig. S3A). Immunoblotting anlaysis showed that MDA-7/IL24 receptors (i.e., IL20/22R) were present on RM1 and B16 tumor cells (Supplementary Fig. S3B). Recombinant human MDA-7/IL24 protein also induced cytotoxicity in RM1 and B16 tumor cells (Supplementary Fig. S3C), which is consistent with our recent reports of killing of human cancer cells by MDA-7/IL24 protein following binding to its cognate receptors. (27, 28)

To examine whether genetic engineering with mda-7/IL24 enabled antigen-specific T cells to more efficiently destroy antigen-negative tumor cells, we compared the tumoricidal activity of OVA_257–264-specific OT-I cells that had been engineered to produce MDA-7/IL24 against antigen-negative parental RM1 cells or antigen-positive RM1-OVA cells. While mock transduced OT-I cells only recognized and killed RM1-OVA cells, MDA-7/IL24-secreting OT-I cells were capable of killing both RM1-OVA and RM1 tumor cells (Fig. 1D). Similar results indicated that OT-I cells carrying MDA-7/IL24 showed increased cytotoxicity against antigen-positive targets (i.e., RM1-OVA) as compared with mock-transduced OT-I cells (Fig. 1D).

We first evaluated the therapeutic activity of MDA-7/IL24-producing OT-I cells in the treatment of mouse prostate cancer. Administration of LV-vec-modified OT-I cells to mice bearing RM1-OVA tumors resulted in a modest growth inhibition of this aggressive tumor. However, treatment with MDA-7/IL24-producing OT-I cells caused profound tumor suppression compared with mock-transduced OT-I cells, which was evident by significantly reduced tumor sizes as well as prolonged life-span (Fig. 2A and B; Supplementary Fig. S4A) and correlated with tumor-associated expression of human MDA-7/IL24 protein following transfer of MDA-7/IL24-producing OT-I cells. In addition, there was a significant elevation of ifng, tnfa, and tbet gene transcription in the tumor upon treatment with MDA-7/IL24-producing OT-I cells (Fig. 2C), suggesting a Th1-skewed tumor immune environment.

We showed that engineering T cells to produce MDA-7/IL24 enhanced their in vivo persistance or expansion after adoptive transfer, which was supported by a higher frequency of tumor-infiltrating Vα2⁺Vβ5⁺CD8⁺ OT-I cells in mice treated with T-mda-7 cells compared with those treated with T-vec cells (Fig. 2D). Treatment with T-mda-7 cells also promoted tumor infiltration by endogenous T cells and their activation based on intracellular IFNγ staining and flow cytometry analyses (Fig. 2E and F). In addition, immune activation enhanced by MDA-7/IL24 modification appeared to be systemic, because there was a higher frequency of OVA-specific T cells in the lymphoid tissues such as spleen following T-mda-7 therapy (Supplementary Fig. S4B).

We next used mouse B16 melanoma as an additional independent model to test the therapeutic effect of MDA-7/IL24-producing T cells. To evaluate the robustness of T-cell therapy, mice received tumor-sensitized T-vec or T-mda-7 cells begining at different times (day 5 or day 9) after innoculation of B16 melanoma (Fig. 3A). In the setting of low tumor burden, MDA-7/IL24-producing T cells were more efficient than mock-transduced cells in suppressing tumor growth (Fig. 3A, left). In the setting of high tumor burden, MDA-7/IL24-producing T cells continued to display therepautic activity, whereas mock-transduced T cells failed to control progression of B16 tumors (Fig. 3A, right).

Consistent with our observations in the RM1 prostate cancer model, MDA-7/IL24 modified T cells induced upregulation of tumor-suppressive immune genes (i.e., ifnγ, tnfα, inos) and downregulation of arginase 1 that is known to facilitate tumor growth (Fig. 3B; Supplementary Fig. S5A). Similarly, MDA-7/IL24 expression was associated with enhanced tumor infiltration by adoptively transferred CD90.1⁺CD8⁺ T cells with an activated phenotype (Fig. 3C and D) as well as endogenous CD8⁺ or CD4⁺ T cells (Supplementary Fig. S5B). Furtheremore, the ratio between CD4 effector T cells (CD3⁺CD4⁺CD25⁺Foxp3⁻) and regulatory T cells (Treg, CD3⁺CD4⁺CD25⁺Foxp3⁺) or the ratio between CD8 T cells (CD3⁺CD8⁺) and Treg (CD3⁺CD4⁺CD25⁺Foxp3⁺) significantly increased in the tumors after treatment with MDA-7/IL24-producing T cells (Fig. 3E). However, these changes were not observed in spleens or DLNs (Supplementary Fig. S5C).

An independent experiment was performed using T cells from Pmel-1 transgenic mice (Pmel cells) that express a TCR specific for a MHC class I–restricted T-cell epitope of the melanoma antigen gp100, i.e., gp100_25–33 that is recognized by CD8⁺ T cells (29). Administration of gp100_25–33–specific T cells secreting MDA-7/IL24 resulted in a profound tumor inhibition compared with mock-transduced T cells (Fig. 3F), which was associated with enhanced T-cell activation in the tumor sites, indicated by intracellular IFNγ staining of tumor-infiltrating CD8⁺CD90.1⁺ Pmel cells (Fig. 3G) and the frequency of IFNγ-producing T cells in the tumors (Fig. 3H).

We next studied whether MDA-7/IL24 engineering improves the efficacy of T cells for eradicating tumor metastases. Experimental lung metastases were established by intravenous injection of luciferase-expressing RM1 cells, followed by treatment with tumor-sensitized T cells that had been engineered with LV-vec or LV-mda-7. Bioluminescence imaging analysis showed that MDA-7/IL24-producing T cells were the most efficient in reducing lung metastases among all treatment groups (Fig. 4A). This superior antitumor activity of MDA-7/IL24-producing T cells was also validated by colony-forming assays using lung cell suspensions prepared from treated mice, as indicated by significantly reduced numbers of colony-forming cells (Fig. 4B). Although MDA-7/IL24 engineering did not alter the memory phenotype of T cells (Supplementary Fig. S6A), it increased the frequency and activation of adoptively transferred as well as endogenous CD8⁺ T cells (Fig. 4C and D) or CD4⁺ T cells with activated phenotypes (Supplementary Fig. S6B) in the tumor sites. This increased presence of T cells in the tumor sites correlated with elevated gene transcription of ifng and tnfa that are known to mediate Th1 immunity (Fig. 4E). In addition to the lungs with tumor lesions, MDA-7/IL24 engineering-enhanced immune activation was also evident in lymphoid tissues (Supplementary Fig. S6C and S6D).

Considering that tumor-sensitized T cells contained both CD8⁺ and CD4⁺ T cells after ex vivo expansion (Fig. 1B), we sought to define the relative contributions of CD8⁺ and CD4⁺ T cells to the enhanced antitumor activity of MDA-7/IL24-producing T cells. Mice with established RM1 tumor metastases were treated with MDA-7/IL24-engineered T cells, sorted CD8⁺ T cells, or CD4⁺ T cells, followed by colony-forming assays to assess metastatic burden in the lungs. We showed that while CD8⁺ T cells were more effective than CD4⁺ T cells in reducing colony formation, possibly due to their direct cytolytic activity, T cells containing both subsets exhibited the highest therapeutic potency (Fig. 4F).

Given the increased tumor infiltration by T cells following engineering to produce MDA-7/IL24, we tested the possibility of MDA-7/IL24 to promote expansion of adoptively transferred T cells using in vivo BrdU incorporation assays. Indeed, MDA-7/IL24-engineered CD8⁺ T cells in lung metastases showed significantly increased BrdU incorporation, indicating active proliferation when compared with mock-transduced RM1 tumor–sensitized T cells (Fig. 5A, left), which correlated with a much higher frequency of tumor-infiltrating CD90.1⁺CD8⁺ T cells (Fig. 5A, right). MDA-7/IL24 engineering also enhanced the proliferation of CD90.1⁺CD4⁺ T cells in transferred tumor-sensitized T cells or endogenous CD90.1⁻CD4⁺ T cells (Supplementary Fig. S7A). However, the MDA7/IL24-enhanced T-cell incorporation of BrdU proliferation was not evident in the lymphoid tissues (Supplementary Fig. S7B). Similar results were obtained in mice with RM1-OVA tumor metastases that were treated with OVA-specific OT-I cells producing MDA-7/IL24 (Fig. 5B).

To demonstrate the direct effect of MDA-7/IL24 production on T-cell proliferation, ex vivo expanded OT-I cells engineered with LV-vec or LV-mda-7 were stimulated with OVA_257–264 peptide, followed by ³H-thymidine incorporation assays to examine T-cell proliferation. MDA-7/IL24-producing OT-I cells proliferated more efficiently than mock-transduced counterparts upon antigen stimulation (Fig. 5C). Exogenous addition of MDA-7/IL24 protein to T-cell culture had the similar proliferation-promoting effect (Fig. 5D). Considering that MDA-7/IL24 has been reported to induce cell proliferation through the STAT1 and STAT3 pathways (30, 31), we next examined the phosphorylation of STAT1 or STAT3 in MDA-7/IL24-engineered T cells upon activation. We showed that production of MDA-7/IL24 increased the phosphorylation of STAT3, not STAT1, when compared with that in mock-transduced T cells (Fig. 5E). Furthermore, pharmacologic inhibition of the STAT3 pathway with Cucurbitacin I abrogated MDA-7/IL24-facilitated T-cell proliferation (Fig. 5F).

qRT-PCR analyses revealed that treatment with human MDA-7/IL24-producing murine T cells led to elevation of mRNA levels of human mda-7/IL24 in tumor tissues and DLNs, which was expected as a result of T-cell trafficking to the tumor sites and homing to lymphoid organs (Fig. 6A). Intriguingly, we found that in mice receiving human MDA-7/IL24-producing T cells high levels of endogenous mouse MDA-7/IL24 (FISP) were induced in the tumors (Fig. 6B), but not in the DLNs (Fig. 6B), suggesting that tumor cells could be the source of the mouse MDA-7/IL24 (FISP).

Human MDA-7/IL24 and mouse MDA-7/IL24 share 61% similarity based on sequence alignment. Previous studies support the antitumor effects of both versions of MDA-7/IL24 (2). To understand the underlying mechanism of induction of endogenous mouse MDA-7/IL24 (FISP), we cocultured RM1 tumor cells with RM1 tumor–sensitized T cells with or without human MDA-7/IL24 engineering. While the presence of unstimulated T cells only slightly upregulated mouse mda-7/IL24 (fisp) gene expression, activation of T cells with anti-CD3/CD28 antibodies induced a sharp increase of mouse mda-7/IL24 (fisp) mRNA, regardless of T cells producing human MDA-7/IL24 or not (Fig. 6C). When coculturing RM1 tumor cells with activated T cells at various ratios, a dose-dependent induction of mouse MDA-7/IL24 (FISP) was evident. However, this induction was similar when coculturing with activated MDA-7/IL24-producing T cells and mock-transduced T cells (Supplementary Fig. S8A). We further validated these results using conditioned media from T-vec or T-mda-7 cells with or without stimulation (Fig. 6D). Finally, treatment of mouse tumor cells with human MDA-7/IL24 protein or coculturing mouse tumor cells with conditioned media derived from gamma-chain cytokines expanded T cells in the presence of exogenous human MDA-7/IL24 failed to induce tumor-intrinsic mouse MDA-7/IL24 (FISP; Supplementary Fig. S8B), suggesting that activation of T cells rather than human MDA-7/IL24 contributes to elevation of tumor-associated mouse MDA-7/IL24 (FISP). A similar observation was made when coculturing B16 tumor cells with activated Gp100-specific Pmel T cells, suggesting that it was not a tumor type–specific effect (Supplementary Fig. S8C). To test the functional significance of induced mouse MDA-7/IL24 (FISP) in cancer cells, we infected mouse B16 melanoma or RM1 prostate cancer cells with lentivirus expressing mouse MDA-7/IL24 (FISP). We showed that overexpression of mouse MDA-7/IL24 (FISP) significantly reduced viability of mouse tumor cells (Fig. 6E), supporting its tumor-suppressive activity.

Finally, we evaluated the therapeutic potency of human MDA-7/IL24-producing T cells in treatment of Hi-Myc transgenic mice that spontaneously develop prostate cancer due to prostate-specific expression of c-Myc driven by two androgen responsive elements and the prostate-specific probasin promoter (19). These mice develop prostatic intraepithelial neoplasia (PIN) as early as 6 to 8 weeks of age and progress to develop invasive prostatic adenocarcinoma by 3 to 6 months of age (19, 32). Different cohorts of 4-month-old Hi-Myc mice with presumed prostate cancer were treated with tumor-sensitized T cells that had been engineered to produce human MDA-7/IL24 (Fig. 7A). Examination of the prostates after euthanizing animals at 6 months showed that T-mda-7 therapy was more effective than T-vec at inhibiting prostate cancer progression, indicated by reduced weights of prostates (Fig. 7B). Administration of MDA-7/IL24-producing T cells was associated with elevation of mRNA levels of human mda-7/IL24, tnfa, and ifng in prostate tissues, indicating trafficking of human MDA-7/IL24-producing T cells to the tumor lesions as well as their activated status (Fig. 7C). Hematoxylin and eosin staining of prostate tissue sections showed progressing adenocarcinoma in the control group, which was modestly delayed by T-vec treatment. In contrast to the pathology in these two groups, sharply reduced lesions were observed in the prostates of T-mda-7–treated mice (Fig. 7D).

It has been suggested that inadequate immune infiltration and T-cell persistence, impaired antigen processing or presentation, and immunosuppression in the TME may contribute to the limited therapeutic outcomes of adoptive cell immunotherapies, for example, CAR-T therapy, in the treatment of solid cancers. In fact, the recurrence of CD19-negative leukemia has been reported following CAR-T-cell therapy (17). In this study, we scrutinized a molecular targeted therapy based on the distinct anticancer activity of the cytokine MDA-7/IL24 delivered by adoptive immunotherapy through tumor/antigen-specific T cells. Using multiple preclinical cancer models including transgenic mice that spontaneously develop prostate cancer, we demonstrate that MDA-7/IL24-engineered T cells exhibit superior antitumor activity by counteracting multiple immune-limiting factors in the TME and targeting cancer cells beyond an antigen-specific fashion (Fig. 7E).

MDA-7/IL24 is well documented as a novel therapeutic agent that is active against a broad spectrum of cancers without harming normal cells (1, 2). In this study, we show that tumor-sensitized T cells or antigen-specific T cells engineered to carry human MDA-7/IL24 have acquired new killing features in addition to T-cell–inherent tumoricidal activity (Fig. 1). Considering that MDA-7/IL24-induced tumor toxicity does not rely entirely on T-cell receptor–mediated antigen recognition, we postulate that MDA-7/IL24-equipped T cells would have an advantage of eliminating both antigen-positive as well as antigen-negative cancer targets. Indeed, our study shows that MDA-7/IL24-producing, OVA_257–264-specific OT-I cells are capable of destroying both parental RM1 tumor cells and RM1-OVA tumor cells, whereas mock-transduced OT-I cells can only kill RM1-OVA cells (Fig. 1). Our findings support the idea that targeted delivery of MDA-7/IL24 to tumor sites by antigen-specific T cells (e.g., TILs, CAR-T) can be utilized to facilitate elimination of antigen-loss or low-expressing cancer cells within the heterogenous tumor cell populations and potentially could overcome the acquired tumor resistance to T cells. Therefore, our molecular-targeted cellular immunotherapy synergistically augments both immune and nonimmune–mediated cancer cell toxicity, which can potentially reduce tumor escape from immune-mediated attack during adoptive T-cell therapy and help achieve better tumor eradication. In support of this notion, we have demonstrated that tumor-reactive or antigen-specific T cells producing MDA-7/IL24 display significantly improved therapeutic activity in the treatment of prostate cancer and melanoma primary tumors as well as metastases (Figs. 2–4). Furthermore, treatment of the Hi-Myc transgenic mice with MDA-7/IL24-engineered T-cell therapy profoundly inhibits the progression of spontaneously developed prostate cancer compared with untreated control animals or those receiving mock-modified T cells (Fig. 7). While there is a limitation of this study relative to testing human cancer models, our findings strongly support the further scrutiny of this approach in the treatment of human cancers using MDA-7/IL24-engineered CAR/TCR-T cells.

We also demonstrated that delivery of MDA-7/IL24 by tumor-reactive or antigen-specific T cells can efficiently reprogram the TME, indicated by substantial elevation of tumor-inhibitory cytokines IFNγ and TNFα. This was associated with increased tumor infiltration by not only adoptively transferred T cells but also endogenous CD8⁺ and CD4⁺ T cells. These findings suggest that MDA-7/IL24-engineered T-cell therapy is able to skew an immunosuppressive TME toward an immunostimulatory one with a Th1-polarized phenotype. Our data are consistent with a previously proposed role of MDA-7/IL24 as a pro-Th1 cytokine (33). It is also possible that tumoricidal MDA-7/IL24 can trigger immunogenic cancer cell death, further amplifying antitumor immune responses (34). Indeed, we previously showed that MDA-7/IL24 induced cancer cell death by potentiating endoplasmic reticulum (ER) stress signals in a cancer cell–specific manner (35). Induction of an ER stress response has been identified as one common feature of immunogenic cell death (36). Nonetheless, confirming a potential role of MDA-7/IL-24 for driving immunogenic tumor cell death requires more studies.

We have made an intriguing observation that engineering T cells to produce MDA-7/IL24 results in increased levels of TILs following adoptive T-cell therapy, suggesting that MDA-7/IL24 produced by therapeutic T cells may support the survival and expansion of effector T cells after their trafficking to the tumor sites (Fig. 5). The in vivo BrdU incorporation assay reveals that those TILs derived from adoptively transferred MDA-7/IL24-producing T cells proliferate more actively than mock-transduced T cells. Surprisingly, this is not seen in those engineered T cells that have homed to the peripheral lymphoid tissues, suggesting that proliferation of TILs may require sustained engagement by antigenic signals from cancer cells or possibly tumor-associated antigen-presenting cells. Our mechanistic studies using a pharmacologic inhibitor of STAT3 reveal that MDA-7/IL24-facilitated T-cell proliferation depends on the STAT3 signaling pathway. Given the distinct features of MDA-7/IL24 in selective destruction of cancer cells and in promoting survival, proliferation, and persistence of effector T cells, future studies are warranted to investigate MDA-7/IL24-engineered TILs or CAR-T cells for potentially improved treatment of human solid cancers.

Another major finding that we have made is that treatment with MDA-7/IL24-carrying T cells induces the elevation of endogenous murine homolog of MDA-7/IL24 or FISP (Fig. 6). Although autocrine induction of MDA-7/IL24 in human cancer cells by extracellular MDA-7/IL24 protein was previously reported (37), our extensive studies using cancer-T-cell coculture, conditioned media from unstimulated or stimulated T cells, and recombinant human MDA-7/IL24 uncover that T-cell activation instead of MDA-7/IL24 from engineered T cells is responsible for inducing MDA-7/IL24 (FISP) in mouse tumor cells. Indeed, overexpression of mouse MDA-7/IL24 (FISP) causes significant inhibition of mouse cancer cells, which is supported by previous reports of mouse MDA-7/IL24 (FISP)-mediated antitumor activity (38, 39). While induction of tumor-endogenous MDA-7/IL24 (FISP) by activated T cells may also contribute to the overall antitumor response, additional studies are necessary to define its exact role in the setting of cancer immunotherapy.

In summary, we have established a novel targeted cellular immunotherapy using tumor-reactive or antigen-specific T lymphocytes engineered to carry and deliver a unique cancer cell killing cytokine, that is, MDA-7/IL24, to achieve broader cancer recognition and destruction. This T-cell–based platform for targeting MDA-7/IL24 to the tumor sites confers advantages over current adenoviral delivery of MDA-7/IL24, which can be potentially restricted by a host antiviral response (33). Because MDA-7/IL24 can initiate production of an autocrine/paracrine production of MDA-7/IL24 (37), the systemic administration of this purified cytokine protein to humans could be problematic. Genetically engineering tumor-reactive or antigen-specific T cells to carry MDA-7/IL24 enhances expansion or prolongs persistence of TILs, and enables their efficient killing of antigen-loss tumor cell variants as well as antigen-positive targets, which can help overcome tumor resistance to the current cellular immunotherapies (e.g., CAR-T) to achieve long-lasting control of cancers and metastases.

P.B. Fisher reports being co-founder and has ownership interest in InterLeukin Combinatorial Therapies, Inc. (ILCT). Virginia Commonwealth University has ownership interest in ILCT. No disclosures were reported by the other authors.

Z. Liu: Data curation, methodology, writing–original draft. C. Guo: Data curation, methodology, writing–original draft, writing–review and editing. S.K. Das: Data curation, methodology. X. Yu: Data curation, methodology. A.K. Pradhan: Data curation, methodology. X. Li: Data curation, methodology. Y. Ning: Data curation, methodology. S. Chen: Resources, data curation, methodology. W. Liu: Resources, data curation. J.J. Windle: Resources, methodology, writing–review and editing. H.D. Bear: Methodology, writing–original draft, writing–review and editing. M.H. Manjili: Formal analysis, supervision, methodology, writing–original draft, writing–review and editing. P.B. Fisher: Conceptualization, data curation, formal analysis, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing. X-Y Wang: Conceptualization, data curation, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing.

The study was supported, in part, by Department of Defense Prostate Cancer Research Awards (W81XWH-11–1-0481, W81XWH-13–1-0162), NCI grants (CA229812, CA099326, CA259599), VCU Massey Cancer Center Research Development Funds, the National Foundation for Cancer Research, and NCI Cancer Center Support Grant to VCU Massey Cancer Center P30CA16059.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.