Mesothelin-Specific Cell-Based Vaccine Generates Antigen-Specific

ORIGINAL ARTICLE Mesothelin-speciﬁc cell-based vaccine generates antigen- speciﬁc immunity and potent antitumor effects by combining with IL-12 immunomodulator

M-C Chang1,2, Y-L Chen3, Y-C Chiang1, T-C Chen1, Y-C Tang1, C-A Chen1, W-Z Sun2 and W-F Cheng1,4,5

Ovarian cancer is a gynecologic malignancy with a high mortality rate. In the present study, we developed a novel cell-based vaccine, Meso-VAX, to generate mesothelin antigen-specific immune responses and immunotherapy against ovarian cancer. Mesothelin, a secreted protein anchored at the cell membrane, has recently been identified as a potential new tumor antigen for ovarian cancer. In this study, mice vaccinated with Meso-VAX and adeno-associated virus (AAV)-IL-12 exhibited dramatic increases in the number of mesothelin-specific CD4+ helper and CD8+ cytotoxic T-cell precursors, higher titers of anti-mesothelin Abs and in vitro tumor killing activity, and all of these mice were tumor-free after 60 days of tumor challenge. In addition, a significant reduction in peritoneal tumors and longer survival were noted in the mice vaccinated with Meso-VAX combined with AAV-IL-12. CD4+ helper and CD8+ cytotoxic T lymphocytes were essential for the antitumor effect generated by Meso-VAX combined with AAV-IL-12. The post-vaccination sera of the mice vaccinated with Meso-VAX and AAV-IL-12 also showed mesothelin-specific complement-dependent cell-mediated cytotoxicity. Our results suggest that a Meso-VAX cell-based vaccine combined with AAV- IL-12 can generate antigen-specific immunological responses and antitumor effects on ovarian cancer.

Gene Therapy (2016) 23, 38–49; doi:10.1038/gt.2015.85

INTRODUCTION histocompatibility complex) class I and class II-restricted 14 Ovarian cancer has one of the highest mortality rates among epitopes. The main disadvantage of autologous and allogeneic gynecologic malignancies.1 Around 75% of patients were diag- whole-cell tumor vaccines is their limited ability to stimulate an nosed at an advanced stage in the 1980s,2,3 and the overall immune response. However, the best method to enhance host survival (OS) rate was only 19–30% from 1955 to 1993.2,3 The immunity to generate effective antitumor responses with whole- standard treatment is surgical tumor debulking, followed by cell tumor vaccines remains unresolved. platinum-based chemotherapy.4,5 Although platinum-based Mesothelin is a secreted protein anchored on the cell 15 chemotherapy results in a response rate of around 80% in all membrane by a glycosylphosphatidylinositol linkage. Mesothe- stages of ovarian cancer,6 50–70% of patients experience lin is absent or present at low levels in normal ovarian tissues, but 16,17 recurrence and ultimately succumb to the malignancy.7 The highly-expressed in ovarian cancer cells, and Hassan et al. 15 development of novel drugs and new therapeutic strategies for suggested that it can be highly immunogenic. In a previous ovarian cancer are therefore urgently needed. study, mesothelin was found to inhibit paclitaxel-induced When developing an effective therapy for ovarian cancer, a apoptosis through a PI3K (phosphoinositide 3-kinase)/Akt-depen- strategy is needed to attack cancer cells only without destroying dent pathway, suggesting that mesothelin can reduce the normal cells and to avoid the development of drug resistance. chemotherapeutic sensitivity of ovarian cancer.17 Mesothelin has Immunotherapy meets these criteria, and it could be used as an also been reported to be a potential tumor marker for the adjuvant to surgery or in combination with chemotherapy or progression of gastric15,16,18,19 and ovarian cancers,17 and it has other biologic therapy such as chemo-immunotherapy or also been identified in triple-negative breast cancer patients.20 bio-chemo-immunotherapy. Thus, mesothelin represents a potential target antigen for the The identification of tumor-associated antigens (TAAs) and host development of immunotherapy against different types of cancer immunity stimulation are essential steps in the development of a such as malignant mesothelioma, gastric, ovarian and breast cancer vaccine.8,9 Various formulations have been investigated as cancers. sources of TAAs for vaccinations, including defined peptides or The tumor cell-based vaccine WF-0 was originally derived from TAAs derived from the entire tumor proteome,10 and whole-cell C57BL/6 mouse peritoneal cells transduced with human papilloma tumor vaccines have been investigated for decades.11–13 virus type 16 E6 and E7 genes to generate immortalized clones. Allogeneic or autologous tumor cells are processed (lysates or The immortalized peritoneal cells were further transfected with irradiated cells) to optimize the release of their antigens by the activated c-Ha-ras gene and then further injected into athymic providing a full complement of TAAs, including MHC (major nude mice, which subsequently developed peritoneal-based

1Department of Obstetrics and Gynecology, Medicine College of Medicine, National Taiwan University, Taipei, Taiwan; 2Department of Anesthesiology, Medicine College of Medicine, National Taiwan University, Taipei, Taiwan; 3Department of Obstetrics and Gynecology, Cathay General Hospital, Taipei, Taiwan; 4Graduate Institute of Clinical Medicine, Medicine College of Medicine, National Taiwan University, Taipei, Taiwan and 5Graduate Institute of Oncology, Medicine College of Medicine, National Taiwan University, Taipei, Taiwan. Correspondence: Dr W-F Cheng, Department of Obstetrics and Gynecology, National Taiwan University Hospital, 7, Chung-Shan South Road, Taipei 100, Taiwan. E-mail: [email protected] Received 15 November 2014; revised 12 June 2015; accepted 23 June 2015; accepted article preview online 11 August 2015; advance online publication, 17 September 2015 Antigen-specific cell-based immunotherapy M-C Chang et al 39 tumors and ascites. The isolated tumor cells from athymic mice bearing mice. The post-vaccination sera of the mice vaccinated were called WF-0, and a previous study confirmed that WF-0 cells with Meso-VAX and AAV-IL-12 also showed enhanced mesothelin- could not generate tumors in immunocompetent mice.21 specific complement-dependent cytotoxicity. Furthermore, there Moreover, due to a lack of endogenous mesothelin expression, were significant preventive and therapeutic effects to inhibit other genes can be introduced into WF-0 cells to generate tumor growth and prolong survival in the mice vaccinated with possible ovarian cancer antigen-specific immune responses and Meso-VAX and AAV-IL-12. These results indicate that the antigen- possible antitumor effects. Thus, WF-0 cells could be an ideal specific cell-based vaccine, Meso-VAX, can enhance the TAA, tumor cell-based vaccine to investigate potential ovarian cancer mesothelin, and specific T cell-mediated immunity and antitumor immunotherapy strategies. effects when combined with the IL-12 cytokine. Tumor-associated A mesothelin-expressing ascitogenic malignant tumor model cell-based vaccines combined with AAV-IL-12 may be able to (WF-3) demonstrating morphological features of intra-peritoneal break through immune tolerance, and represent an innovative tumorigenesis has been developed.21 In this model, the WF-3 approach in the development of cancer immunotherapy for tumor cells express high levels of mesothelin, commonly observed ovarian cancer. in intra-peritoneal tumors such as ovarian cancer and malignant mesothelioma. Thus, the WF-3 tumor model would serve as an excellent pre-clinical tumor model to study clinically relevant RESULTS issues such as the screening and evaluation of the effectiveness of Ectopic-expressing human mesothelin in murine ovarian cancer therapeutic interventions by monitoring the level of mesothelin. cells To date, whole-cell (including autogenic or allogenic) vaccines RT-PCR and flow-cytometry analysis showed that the Meso-VAX have been designed to induce specific antitumor immune tumor cells expressed higher levels of human mesothelin RNA responses.22,23 However, the application of cell-based vaccines is (Figure 1a) and protein (Figure 1b) compared with the paternal limited because only a small number of TAAs has been identified. WF-0 and WF-0/mock cells. This indicated that WF-0 cells Another problem is that tumor cells may escape from the immune transfected with human mesothelin gene could express human response due to immunological editing. Therefore, mesothelin as mesothelin. a potential TAA for ovarian cancer can be expected to generate immune tolerance in the host. Vaccination with Meso-VAX and AAV-IL-12 enhanced the In this study, we introduced the human mesothelin gene into mesothelin-specific cell-mediated and humoral immunity WF-3 tumor cells to generate the xenogenic antigen-specific, The representative figures of flow-cytometry analysis of cell-based vaccine Meso-VAX. The combination of Meso-VAX mesothelin-specific interferon (IFN)-γ-secreting CD4+ helper and adeno-associated virus (AAV)-IL-12 was further tested to T-cell precursors in the various groups (Figure 2a) showed investigate whether this strategy could break immune tolerance that the mice vaccinated with Meso-VAX and AAV-IL-12 to generate antitumor effects and enhance both mesothelin- (0.051 ± 0.0014%) had a significantly higher percentage of specific cellular and humoral immunity. Vaccination with mesothelin-specific IFN-γ-secreting CD4+ T-cell precursors Meso-VAX and AAV-IL-12 enhanced the number of mesothelin- compared with the other groups (PBS: 0.003 ± 0.0007%; specific CD4+ helper and CD8+ T cytotoxic cell precursors and WF-0: 0.003 ± 0.0008%; WF-0/mock: 0.005 ± 0.0012%; Meso-VAX: generated higher titers of anti-mesothelin Abs in WF-3 tumor 0.003 ± 0.0005%; AAV-IL-12: 0.009 ± 0.0007%; WF-0 with AAV-IL-12:

Figure 1. Characterization of human mesothelin expression in Meso-VAX cells. (a) RNA expression levels of mesothelin in various WF-0 transfectants by RT-PCR. Note: The mesothelin expression was signiﬁcantly increased in the Meso-VAX cells compared with the parental WF-0 and WF-0/mock cells. (b) Protein expression of mesothelin in various WF-0 transfectants using ﬂow-cytometry analysis. Note: The Meso-VAX tumor cells expressed human mesothelin compared with the paternal WF-0 and WF-0/mock tumor cells. Data from all of the experiments were from one representation of at least two experiments performed.

Figure 2. Mesothelin-specific immunological profiles of the mice using flow-cytometry analysis and ELISA. (a) Representative figures of mesothelin-specific IFN-γ-secreting CD4+ helper T-cell precursors in various vaccinated groups (A1: PBS; A2: WF-0; A3: WF-0/mock; A4: Meso- VAX alone; A5: AAV-IL-12; A6: WF-0 with AAV-IL-12; A7: WF-0/mock with AAV-IL-12; A8: Meso-VAX with AAV-IL-12). (b) Bar graph depicting the percentages of mesothelin-specific IFN-γ-secreting CD4+ helper T-cell precursors/3.5 × 105 splenocytes (mean ± s.e.m.) [columns indicate 1: PBS; 2: WF-0; 3: WF-0/mock; 4: Meso-VAX; 5: AAV-IL-12; 6: WF-0 with AAV-IL-12; 7: WF-0/mock with AAV-IL-12; 8: Meso-VAX with AAV-IL-12]. Note: Mice vaccinated with Meso-VAX combined with AAV-IL-12 had the highest percentages of mesothelin-specific IFN-γ-secreting CD4+ T-cell precursors compared with the other groups (P = 0.007, by one-way ANOVA). (c) Representative figures of mesothelin-specific IFN-γ- secreting CD8+ cytotoxic T-cell precursors in various vaccinated groups (C1: PBS; C2: WF-0; C3: WF-0/mock; C4: Meso-VAX alone; C5: AAV-IL-12; C6: WF-0 with AAV-IL-12; C7: WF-0/mock with AAV-IL-12; C8: Meso-VAX with AAV-IL-12). (d) Bar graph depicting the percentages of mesothelin-specific IFN-γ-secreting CD8+ cytotoxic T-cell precursors/3.5 × 105 splenocytes (mean ± s.e.m.) [columns indicate 1: PBS; 2: WF-0; 3: WF-0/mock; 4: Meso-VAX alone; 5: AAV-IL-12; 6: WF-0 with AAV-IL-12; 7: WF-0/mock with AAV-IL-12; 8: Meso-VAX with AAV-IL-12]. Note: Mice vaccinated with Meso-VAX combined with AAV-IL-12 had the highest percentages of IFN-γ-secreting CD8+ T-cell precursors among the groups (P = 0.009, by one-way ANOVA). (e) Mesothelin-specific Abs detected by ELISA in various vaccinated groups [columns 1: PBS; 2: WF-0; 3: WF-0/mock; 4: Meso-VAX alone; 5: AAV-IL-12; 6: WF-0 with AAV-IL-12; 7: WF-0/mock with AAV-IL-12; 8: Meso-VAX with AAV-IL-12]. Note: Mice receiving Meso-VAX combined with AAV-IL-12 had the highest titers of anti-mesothelin Abs among the groups. Data from all of the experiments were from one representation of at least two experiments performed.

0.006 ± 0.0007%; and WF-0/mock with AAV-IL-12: 0.005 ± 0.0010%; 0.006 ± 0.0011%; AAV-IL-12: 0.005 ± 0.0005%; WF-0 with AAV-IL-12: Po0.001, by one-way analysis of variance (ANOVA)) (Figure 2b). 0.004 ± 0.0009%; and WF-0/mock with AAV-IL-12: 0.006 ± 0.0015%; The representative figures of mesothelin-specific CD8+ cytotoxic Po0.001, by one-way ANOVA) (Figure 2d). However, the T-cell precursors in the various vaccinated groups by percentages of IFN-γ-secreting CD4+ helper T and CD8+ cytotoxic flow-cytometry analysis (Figure 2c) showed that the mice T cells in the Meso-VAX only group were not higher than those of vaccinated with Meso-VAX and AAV-IL-12 (0.043.0 ± 0.0020%) also the other groups. had the highest percentage of IFN-γ-secreting CD8+ cyto- The antigen-specific humoral immunity generated in the toxic T-cell precursors compared with the other groups various groups was further evaluated by detecting anti- (PBS: 0.003 ± 0.0008%; WF-0: 0.005 ± 0.0007%; WF-0/mock: mesothelin Abs. The mice vaccinated with Meso-VAX and AAV-

Gene Therapy (2016) 38 – 49 © 2016 Macmillan Publishers Limited Antigen-specific cell-based immunotherapy M-C Chang et al 41 IL-12 (1.48 ± 0.21) had the highest titers of anti-mesothelin Ab Meso-VAX vaccine combined with AAV-IL-12 generated tumor than the other groups (PBS: 0.27 ± 0.03; WF-0: 0.51 ± 0.03; WF-0/ protection effects in mice challenged with mesothelin-expressing mock: 0.31 ± 0.02; Meso-VAX: 0.28 ± 0.03; AAV-IL-12: 0.39 ± 0.04; tumor cells WF-0 with AAV-IL-12: 0.42 ± 0.06; and WF-0/mock with AAV-IL-12: The reason why we performed the tumor protection experiments 0.29 ± 0.03; 1:100 dilution; Po0.001, by one-way ANOVA) is because cancer immunotherapy can be clinically applied in (Figure 2e). These results demonstrated that the combination of cancer patients as adjuvant or maintenance therapy after Meso-VAX vaccine and AAV-IL-12 could enhance the mesothelin- conventional therapies such as surgery, irradiation and che- specific cell-mediated and humoral immunity. motherapy to prevent tumor recurrence. The protocol of evaluating the tumor protection effects and immunological responses of the cell-based vaccine with or without AAV-IL-12 in Splenocytes of Meso-VAX combined with AAV-IL-12 had the the mice is shown in Figure 4a. All (100%) of the mice that highest in vitro tumor-specific killing activity received Meso-VAX with AAV-IL-12 remained tumor-free even To further evaluate whether intracellular cytokine staining assays 90 days after WF-3/Luc tumor challenge (Figure 4b). In contrast, all could reflect the tumor killing effect, in vitro tumor killing assays of the mice vaccinated with PBS, WF-0, WF-0/mock, Meso-VAX, were performed. The representative figures of luminescence AAV-IL-12, WF-0 with AAV-IL-12, or WF-0/mock with AAV-IL-12 detected by an IVIS Imaging System are shown in Figure 3a. developed tumors within 42 days of tumor challenge. Mesothelin-specific cytotoxic T lymphocytes from the mice The survival status of the mice in the various groups was vaccinated with Meso-VAX and AAV-IL-12 (9.5 ± 0.8 × 104 (p/s)) also evaluated. All (100%) of the mice that received had the lowest average luminescence of WF-3-LG cells compared Meso-VAX with AAV-IL-12 were alive 90 days after WF-3/Luc with the other treatment groups (for splenocytes: WF-3/LG = 10:1 tumor challenge (Figure 4c); however, none of the mice in the experiments, PBS group 6.7 ± 0.6 × 106 (p/s); WF-0 group other groups survived for longer than 80 days after tumor 6.5 ± 0.7 × 106 (p/s), WF-0/mock group 6.2 ± 0.5 × 106 (p/s), Meso- challenge. VAX group 6.5 ± 0.7 × 106 (p/s), AAV-IL-12 group 5.5 ± 0.5 × 106 (p/ The effects of WF-0 or WF-3 lysates combined with AAV-IL-12 s), WF-0 combined Meso-VAX group 5.9 ± 0.7 × 106 (p/s), WF-0/ were then evaluated. The mice vaccinated with WF-0 or WF-3 mock combined Meso-VAX group 5.7 ± 0.5 × 106 (p/s), P = 0.002, lysates combined with AAV-IL-12 did not express enhanced one-way ANOVA; Figure 3b). These results indicated that Meso- mesothelin-specific Abs (Supplementary Figure 1). These results VAX combined with AAV-IL-12 generated the most potent killing indicated that Meso-VAX combined with AAV-IL-12 could protect effect on WF-3/LG tumor cells. the mice against lethal mesothelin-expressing tumor cells.

Figure 3. In vitro tumor-specific killing activity of splenocytes of various vaccinated groups. (a)Representativefigures of luminescence activity in various vaccinated groups detected by an IVIS Imaging Systems. The ratios of splenocytes to WF-3/LG cells were 2:1, 5:1 and 10:1. (b)Thebarfigure of average luminescence of WF-3/LG cells in various vaccinated groups (columns 1: PBS, 2: WF-0, 3: WF-0/mock, 4: Meso-VAX alone, 5: AAV-IL-12, 6: WF-0 with AAV-IL-12, 7: WF-0/mock with AAV-IL-12, 8: Meso-VAX with AAV-IL-12). Note:The Meso-VAX combined with AAV-IL-12 group had the lowest luminescence activity compared with the other groups (splenocytes: WF-3/LG cells 10:1 ratio, P = 0.002, one-way ANOVA). Data from all of the experiments were from one representation of at least two experiments performed.

Figure 4. In vivo tumor protection experiments of the mice. (a) Diagrammatic representation of the different regimens of cell-based vaccines with or without AAV-IL-12 vaccination. V, vaccination; A, AAV. (b) Percentages of tumor-free mice in the various vaccinated groups. Note: All of the mice (100%) receiving Meso-VAX combined with AAV-IL-12 remained tumor-free 90 days after WF-3/Luc tumor challenge. However, none of the mice in the other groups were tumor-free 42 days after tumor challenge. (c) Percentages of living mice in the various vaccinated groups. Note: All of the mice that received Meso-VAX combined with AAV-IL-12 were alive 90 days after WF-3/Luc tumor challenge. However, none were alive in the other groups more than 80 days after tumor challenge. Data from all of the experiments were from one representation of at least two experiments performed.

Meso-VAX vaccine combined with AAV-IL-12 led to a signiﬁcant Meso-VAX combined with AAV-GM-CSF or recombinant murine reduction in peritoneal tumors in the vaccinated mice IL-12 did not result in the same therapeutic effects as Meso-VAX The protocol of evaluating the therapeutic effects and OS status of combined with AAV-IL-12 the mice in the in vivo tumor treatment experiments is shown in To evaluate whether other cytokines in AAV format could also Figure 5a. The luciferase activities of the WF-3/LG tumor-bearing generate the same potent antitumor effects as Meso-VAX with mice as detected by an IVIS system are shown in Figure 5b, and AAV-IL-12, we further tested Meso-VAX combined with AAV-GM- were found to be the lowest in the Meso-VAX and AAV-IL-12 CSF. The protocol of the vaccine experiments is shown in groups (6.4 ± 1.5 × 104) compared with those in the other groups Figure 6a. The mice that received Meso-VAX with AAV-GM-CSF (Figure 5c) (Po0.001, by one-way ANOVA) (PBS: 1.0 ± 0.1 × 108; developed tumors later and survived for longer than those that WF-0: 7.2 ± 2.2 × 108; WF-0/mock: 1.0 ± 0.2 × 109; Meso-VAX: received PBS or Meso-VAX alone (Figures 6b and c). All of the mice 1.1 ± 0.3 × 108; AAV-IL-12: 3.8 ± 0.7 × 108; WF-0 with AAV-IL-12: vaccinated with Meso-VAX and AAV-GM-CSF had tumors within 1.0 ± 0.1 × 109; and WF-0/mock with AAV-IL-12: 1.1 ± 0.2 × 109; 50 days after tumor challenge (Figure 6b), and none were alive Po0.001, by one-way ANOVA). after 80 days of tumor challenge (Figure 6c). Of the mice that received Meso-VAX with AAV-IL-12, 80% These results demonstrate that Meso-VAX combined with remained tumor-free even at 90 days after WF-3/LG tumor another AAV format such as AAV-GM-CSF did not suppress challenge. However, all of the mice in the other groups developed mesothelin-expressing ovarian cancer as effectively as Meso-VAX tumors within 60 days of tumor challenge (Figure 5d). In terms of combined with AAV-IL-12. survival, all of the mice that received Meso-VAX with AAV-IL-12 To investigate whether IL-12 in protein format had a similar were alive 90 days after WF-3/LG tumor challenge (Figure 5e); antitumor effect to that in AAV format, we evaluated the effect of however, none of the mice in the other groups survived. Meso-VAX combined with recombinant murine IL-12 (rmIL-12).

Figure 5. In vivo tumor treatment experiments of the mice. (a) Diagrammatic representation of the different regimens of cell-based vaccine with or without AAV-IL-12. (b) Representative ﬁgures of luminescence images of mice in various groups at the indicated intervals (columns indicate 1: PBS; 2: WF-0; 3: WF-0/mock; 4: Meso-VAX alone; 5: AAV-IL-12; 6: WF-0 with AAV-IL-12; 7: WF-0/mock with AAV-IL-12; 8: Meso-VAX with AAV-IL-12). (c) Bar graph depicting the luminescence activity of tumor-bearing mice in various vaccinated groups (Columns indicated1: PBS; 2: WF-0; 3: WF-0/mock; 4: Meso-VAX alone; 5: AAV-IL-12; 6: WF-0 with AAV-IL-12; 7: WF-0/mock with AAV-IL-12; 8: Meso-VAX with AAV- IL-12). (d) Percentages of tumor-free mice in the various vaccinated groups. Note: 80% of the mice that received Meso-VAX combined with AAV-IL-12 were tumor-free 90 days after WF-3/Luc tumor challenge. However, all of the mice in the other groups developed tumors within 50 days after tumor challenge. (e) Survival analysis of mice in the various vaccinated groups. Note: 100% of the mice that received Meso-VAX with AAV-IL-12 were alive 90 days after WF-3/Luc tumor challenge; however, none of the mice in the other groups survived 60 days after tumor challenge. Data from all of the experiments were from one representation of at least two experiments performed.

As shown in Figures 6b and c, mice vaccinated with Meso-VAX Representative ﬁgures of luminescence analysis of the IVIS system combined with rmIL-12 developed tumors later and in the Ab depletion experiments are shown in Figure 7a. Mice survived longer than the mice that received Meso-VAX or treated with Meso-VAX and AAV-IL-12 without Ab depletion had rmIL-12 alone. However, all of the mice vaccinated with the lowest luciferase activity on day 35 after WF-3/Luc tumor Meso-VAX and rmIL-12 had tumors within 50 days of tumor challenge compared with the other groups (Figure 7a). The challenge (Figure 6b), and none of them were alive after 80 days luciferase activity in the NK1.1 Ab depletion group was also lower of tumor challenge (Figure 6c). These results showed that the than in the CD4 or CD8 Ab depletion groups and the PBS- Meso-VAX vaccine combined with AAV-IL-12 could generate treated group. potent therapeutic effects in mesothelin-expressing tumor All of the non-depleted vaccinated mice survived for 90 days cells; however, this effect was not seen with a combination of after tumor challenge (Figure 7b). None of the vaccinated Meso-VAX and rmIL-12. mice depleted of CD4+ helper T or CD8+cytotoxic cells survived after 90 days of tumor challenge; however, 90% of the natural CD4+ helper T cells and CD8+ cytotoxic T cells were essential for killer (NK) cell-depleted vaccinated mice survived for 90 days the antitumor effects of the Meso-VAX vaccine combined with after tumor challenge. These results demonstrated that CD4+ AAV-IL-12 helper and CD8+ cytotoxic T cells were essential for the To determine the lymphocyte subsets that are important for the antitumor effects generated by the Meso-VAX vaccine with antitumor effects of the Meso-VAX vaccine combined with AAV- AAV-IL-12, and that NK cells had little to no effect on in vivo IL-12, in vivo antibody depletion experiments were performed. tumor growth.

Figure 6. In vivo tumor treatment experiments of the mice vaccinated with Meso-VAX combined with rmIL-12, AAV-GM-CSF or AAV-IL-12. (a) Diagrammatic representation of the different regimens of cell-based vaccine with AAV-GM-CSF or rmIL-12. V, vaccination; A, AAV; R, rmIL-12. (b) Percentages of tumor-free mice in the various vaccinated groups. Note: 80% of the mice that received Meso-VAX combined with AAV-IL-12 were tumor-free 90 days after WF-3/Luc tumor challenge. However, all of the mice in the other groups developed tumors within 50 days after tumor challenge. (c) Survival analysis of mice in the various vaccinated groups. Note: 100% of the mice that received Meso-VAX with AAV-IL-12 were alive 90 days after WF-3/Luc tumor challenge; however, none of the mice in the other groups survived longer than 80 days after tumor challenge. Data from all of the experiments were from one representation of at least two experiments performed.

Post-vaccination sera were capable of generating mesothelin- including enhancing chemotherapy resistance and inducing specific complement-dependent cell-mediated cytotoxicity tumor invasion have been established,24,26 studies targeting We next determined whether the generation of anti-mesothelin mesothelin as an ovarian cancer antigen are lacking. Several Ab in the sera of the mice vaccinated with Meso-VAX and AAV- clinical trials on mesothelin target therapies are ongoing, most of IL-12 could stimulate antigen-specific complement-dependent which are focused on antibody or adaptive T-cell transfer.27,28 cell-mediated cytotoxicity. The representative figures of lumines- Jaffee et al.29 reported that pancreatic cancer patients show a cence activity of mesothelin-specific complement-dependent cell- mesothelin-specific immunological response when receiving mediated cytotoxicity assays in various groups (Figure 8a) autologous tumor cell transfer. Thomas et al.30 also reported that revealed that the Meso-VAX and AAV-IL-12 (2.5 ± 0.3 × 106) group whole tumor cell vaccinations can lyse mesothelin-expressing had the lowest luciferase activity compared with those of the tumor cells. Thus, mesothelin-target immunotherapy may be other groups (PBS: 7.4 ± 0.6 × 106; WF-0: 7.7 ± 0.4 × 106;WF-0/mock: therapeutic strategy for certain types of cancer. 7.5 ± 0.5 × 106; Meso-VAX: 7.9 ± 0.8 × 106; AAV-IL-12: 6.5 ± 0.7 × 106; In this study, a mesothelin-expressing cell-based vaccine alone WF-0 with AAV-IL-12: 6.6 ± 0.4 × 106; and WF-0/mock with could not generate potent antitumor effects against mesothelin- AAV-IL-12: 6.9 ± 0.5 × 106; Po0.001, by one-way ANOVA) expressing tumor cells (Figures 3 and 4). Mesothelin has been 17,18 (Figure 8b). These results demonstrated that Meso-VAX combined shown to be over-expressed in ovarian cancer, however as a with AAV-IL-12 could generate mesothelin-specific Abs that could TAA, mesothelin cannot generate effective antitumor immunolo- enhance complement-dependent cell-mediated cytotoxicity. gical responses due to immune tolerance to endogenous antigens. A xenogeneic model using human genes that are highly homologous to mice has been developed to generate more DISCUSSION effective antigen-specific immunity, and this may be an innovative Mesothelin is an attractive target for ovarian cancer immunother- strategy for the development of cancer immunotherapy.31,32 apy, as it is predominantly expressed on the cell surface of tumor Therefore, in this study we introduced the human mesothelin cells.24,25 Although the biological functions of mesothelin gene into WF-0 cells (the Meso-VAX cell-based vaccine). However,

Figure 7. In vivo antibody depletion experiments of the mice vaccinated with Meso-VAX combined with AAV-IL-12. (a) Representative figures of luminescence images of the mice vaccinated with Meso-VAX and AAV-IL-12, the respective Ab, and challenged with WF-3/Luc tumor cells at the indicated intervals. (b) Bar graph depicting the luminescence activities of mice in various groups. Note: None of the mice with depleted CD4+ helper or CD8+ cytotoxic T lymphocytes were alive 80 days after WF-3/Luc tumor challenge. However, all (100%) of the mice without Ab depletion and 90% of the mice with depleted anti-NK cells remained alive 90 days after tumor challenge. Data from all of the experiments were from one representation of at least two experiments performed. the Meso-VAX vaccine alone could not generate potent antigen- antitumor effects compared with Meso-VAX or AAV-IL-12 alone specific immunity or potent antitumor effects. These results are (Figures 2–5). consistent with the findings of Chang et al.,25 and thus we IL-12 has also been reported to initiate a T helper type 1 (Th1) hypothesized that the immune tolerance of TAAs may limit the response by stimulating NK cells to secrete IFN-γ.37 More potency of the Meso-VAX cell-based vaccine. importantly, IL-12 has been shown to mediate tumor rejection, IL-12 in AAV format can effectively enhance the potency of TAA- which is dependent on CD4+ helper T cells.38 Furthermore, IL-12 specific cell-based vaccines by breaking the immune tolerance. In has been reported to facilitate the proliferation of CD8+ cytotoxic this study, Meso-VAX combined with AAV-IL-12 enhanced T cells to stimulate the function of immune effector and memory fi 39 stronger mesothelin-speci c antitumor immunity and generated cells. We found that IL-12 had an adjuvant role in reforming – potent antitumor effects (Figures 2 5). IL-12 is naturally produced mesothelin to become potently immunogenic by acting on DCs, by dendritic cells (DCs) and macrophages to enhance the NK cells and T lymphocytes, although the real function of IL-12 in cytotoxic activity of NK cells and T lymphocytes. Previous studies enhancing antitumor immunity may be broader. have demonstrated that a combination of IL-12 with TAA-loaded fi In addition to cellular immunity, humoral immunity including DCs can induce more effective tumor-speci c immune responses complement-dependent cell-mediated cytotoxicity may be an and T-cell stimulation than DCs alone.33,34 We found that the IL-12 important antitumor mechanism generated by a cancer vaccine. concentration in sera was significantly higher in the Meso-VAX Our results imply that Meso-VAX and AAV-IL-12 treatment may and AAV-IL-12 group than in the PBS group from 3 days after lead to an antitumor effect via triggering complement-dependent vaccination. In addition, the IL-12 concentration in sera persisted for at least 14 days after vaccination in the Meso-VAX and AAV- cell-mediated cytotoxicity (Figure 8). Complement-dependent cell- mediated cytotoxicity has been shown to be important for the IL-12 group (Supplementary Figure 2). However, a combination of 40 Meso-VAX and rmIL-12 did not result in as effective an antitumor therapeutic effect of monoclonal antibodies. When tumor- fi effect as with Meso-VAX and AAV-IL-12 (Figures 6b and c). reactive antibodies are generated from tumor-speci c B cells, Nanni et al.35 reported that a combination of systemic IL-12 with they can bind to the surface of cancer cells and then induce an allogeneic antigen-expressing cell vaccine inhibited HER-2/neu- complement-dependent cell-mediated antitumor effects. Tang 41 induced mammary tumorigenesis. Moreover, a phase 2 multi- et al. reported that the human single-domain Ab can generate institutional trial demonstrated that patients with ovarian or complement-dependent cell-mediated cytotoxicity and ADCC peritoneal carcinoma achieved stable disease when injected with activity against mesothelioma and other mesothelin-expressing recombinant IL-12 intra-peritoneally.36 The results of the present tumors.41 Enhancing complement-dependent cell-mediated cyto- study showed that the Meso-VAX cell-based vaccine combined toxicity may therefore be another potential strategy for develop- with AAV-IL-12 enhanced mesothelin-specific immunity and ing cancer immunotherapy.

Figure 8. Mesothelin-specific complement-dependent cell-mediated cytotoxicity assay in various groups of vaccinated mice. (a) Representative figures of luminescence images in various groups. (b) Bar graph depicting the quantification of luminescence in mesothelin-specific complement-dependent cell-mediated cytotoxicity assays after incubation for 18 h (column 1: PBS; 2: WF-0; 3: WF-0/mock; 4: Meso-VAX alone; 5: AAV-IL-12; 6: WF-0 with AAV-IL-12; 7: WF-0/mock with AAV-IL-12; 8: Meso-VAX with AAV-IL-12). Note: Sera from mice vaccinated with Meso-VAX combined with AAV-IL-12 had the lowest luminescence activity compared with the other groups. Data from all of the experiments were from one representation of at least two experiments performed.

Cell-based vaccines can stimulate antigen-specific anti-tumor AAV is a good delivery system to infect non-dividing cells, and it immunity by enhancing T cell-mediated immunity. Our results has the ability to stably integrate into the host cell genome. It has suggest that the Meso-VAX vaccine combined with AAV-IL-12 been reported that the AAV vector can be used to delivery genes could generate both CD4+ helper and CD8+ cytotoxic T-cell into many cells, including skeletal muscles,43 neurons,44 vascular immunity (Figure 2). The in vivo Ab depletion experiments also smooth muscle cells45 and hepatocytes.46 Although AAV can be revealed that both CD4+ helper and CD8+ cytotoxic T cells are an effective vector for gene delivery in many types of animal essential effector cells for the antitumor effect generated by Meso- diseases, there is currently no evidence of which cell types may be 47 VAX and AAV-IL-12 (Figure 7). Consistent with our findings, Chang infected by peritoneal injections. Maciej et al. reported that AAV et al.25 also reported that a mesothelin-specific DNA vaccine can could facilitate gene transferability to cancer cells disseminated in generate both CD4+ helper and CD8+ cytotoxic T-cell immunity. the serosal cavity. Epithelial cells have been reported to be the We hypothesize that enhancing different effector cells against first infective target of AAV-mediated cancer gene therapy after 47 tumor cells may be an important strategy to target TAAs for both peritoneal injection. We hypothesize that AAV-IL-12 infected DNA- and cell-based vaccines. intraperitoneal tumor cells or peritoneal lining cells and mesothe- Our results demonstrated a significant therapeutic effect lial cells in our model. fi against mesothelin-expressing ovarian cancer with Meso-VAX In conclusion, we developed a mesothelin-speci c cell-based fi combined with AAV-IL-12. Mice vaccinated with WF-0 or WF-3 vaccine, Meso-VAX. The ndings support the hypothesis that an lysates combined with AAV-IL-12 did experience an antitumor allogeneic TAA cell-based vaccine combined with IL-12 can be an effect after challenge with WF-3 tumor cells or enhanced effective strategy for cancer immunotherapy by enhancing T-cell fi immunity and controlling mesothelin-expressing tumor cells. mesothelin-speci c Abs (Supplementary Figure 1). As mesothelin fi is an endogenous tumor antigen, self-tolerance could be expected Further studies to evaluate whether a mesothelin-speci c cell- when developing cancer immunotherapy by targeting mesothelin. based vaccine can be a potential strategy for the treatment of different types of cancer that express mesothelin such as Chang et al.25 showed that a DNA vaccine targeting human mesothelioma and pancreatic cancer are warranted. mesothelin could generate potent antitumor effects against human mesothelin-expressing tumors in a murine model. Gritzapis 42 et al. also demonstrated that altered human HER-2/neu MATERIALS AND METHODS peptide ligands could overcome self-tolerance and generate Cell lines antitumor effects against Her-2/neu overexpressing tumors in The generation of WF-0 and WF-3 tumor cells was performed as described mice. A limitation of Meso-VAX is immune tolerance, and to previously. The cells were maintained in RPMI-1640 supplemented with overcome this, the construct of human mesothelin needs to be 10% (volume/volume) fetal bovine serum, 50 U ml − 1 penicillin/streptomy- modified before application in future clinical trials. cin, 2 mM L-glutamine, 1 mM sodium pyruvate, 2 mM non-essential amino

Gene Therapy (2016) 38 – 49 © 2016 Macmillan Publishers Limited Antigen-specific cell-based immunotherapy M-C Chang et al 47 acids and 0.4 mg ml − 1 G418 at 37 °C with 5% carbon dioxide.21 The cell containing 100 μl of AAV-IL-12, AAV-GM-CSF or AAV-GFP viral stocks. The lines were tested for mycoplasma contamination. cells were incubated for 2 h and then supplied with additional growth WF-3/Luc cells were generated by transducing WF-3 cells with the medium. The cells were then maintained for 48–72 h at 37 °C until the lentiviral vector. Briefly, luciferase cDNA was amplified by PCR from pGL2- appearance of the first signs of a cytopathic effect. basic (Promega, Madison, WI, USA) and cloned into the pLKO/AS3.1.EGFP3 Adenoviral particles were collected by alternating the tubes between lentiviral vector (Academia Sinica, Taipei, Taiwan) to generate liquid nitrogen and a 37 °C water bath, with vortexing after each thaw. The pLKO/luciferase/AS3.1.EGFP3, which, together with pCMVΔR8.91 cell debris was collected by centrifugation at 12 000 g for 10 min, and the (Academia Sinica) and pMDG (Academia Sinica), were then transfected supernatants were transferred to new micro-centrifuge tubes and stored at into 293 T cells to assemble the lentivirus. The lentivirus was collected 48 h − 80 °C.34 The virus-containing supernatant was purified using an Adeno-X after transfection. The WF-3 cells were further infected with a lentivirus Maxi Purification kit (Clontech Laboratories, Mountain View, CA, USA) with 8 μg/ml Polybrene (Sigma Chemicals Co., St Louis, MO, USA) for 48 h. according to the manufacturer’s protocol. A single clone was isolated and cultured for further studies. To measure transduction efficiency, HEK293 cells were transfected with AAV-IL-12 or AAV-GM-CSF, and the efficiency was determined using an Adeno-X Rapid Titer Kit (Clontech Laboratories). Briefly, HEK293 cells Plasmid DNA constructs and DNA preparation (5 × 105cells per ml) were seeded in 12-well plates. AAV-IL-12 or AAV-GM- To generate pcDNA3-hMSLN (human mesothelin), mesothelin was first CSF was diluted with PBS from 10–2 to 10–7 diluents. The viral dilution was fi ampli ed with PCR using human placenta cDNA as the template and then added to the 12-well plate for 48 h at 37 °C in 5% CO2. After 5′-CCGGTCTAGAATGGCCTTGCCAACGGCTCG-3′ and 5′-CGCGGATCCTCAGG incubation, the medium was aspirated and fixed in 100% ice-cold CCA;GGGTGGAGGCTAGG-3′ as primers. The amplified product was then methanol. The 12-well plate was rinsed with PBS containing 1% BSA. cloned into the XbaI/BamH1 sites of pcDNA3 vector (Invitrogen Life Anti-hexon Ab (0.5 ml) was then added and the plate was incubated for Technologies, Carlsbad, CA, USA). All of the constructs were verified by another hour at 37 °C. restriction analysis and DNA sequencing. After incubation, the anti-hexon Ab was aspirated and rat anti-mouse Ab (catalog no. 632250, HRP conjugate) dilution (0.5 ml) was added to each Generation of the Meso-VAX vaccine well for an hour at 37 °C. DAB working solution was then added, and the fi brown/black positive cells were counted using a microscope (×20 To generate the Meso-VAX vaccine, pcDNA3-hMSLN was rst transfected objective). The infectious units (ifu)/ml for each well were calculated using into WF-0 cells to form WF-0/hMSLN cells using Lipofectamine 2000 the formula: reagent (Invitrogen Life Technologies) according to the manufacturer’s − instructions. To select stable clones, hygromycin (700 μgml 1) was added ðÞinfected cells=field ´ ðÞfields=well volume virusðÞ ml ´ ðÞdilution factor to the culture medium 48 h after transfection. Hygromycin-resistant clones were individually picked, expanded and used for the subsequent experiments. WF-0 cells were transfected with pcDNA3 alone as WF-0/mock and used as the negative control. Generation of mesothelin peptide and WF-3 lysates fi The WF-0/hMSLN tumor cells were irradiated and de ned as the Meso- The H-2Db-restricted mesothelin peptide-specific cytotoxic T-lymphocyte VAX vaccine. Briefly, WF-0/hMSLN tumor cells (1 × 106cells per ml) were − 1 epitope (amino acid [aa] 406-414, GQKMNAQAI) was synthesized (Kelowna suspended in RPMI complete medium and irradiated at 555.5 Gy;min Inc., Taipei, Taiwan).25 Mesothelin whole protein was prepared from WF-3 using a Gammacell 3000 Elan system (MDS Nordion, Ottawa, ON, Canada) tumor cell lysates. Briefly, cells were suspended in PBS (0.5 ml) and lysed with a cesium-137 source. The Meso-VAX vaccine was used in the with five freeze (liquid nitrogen)-and-thaw (room temperature) cycles. The subsequent experiments. lysate was then centrifuged at 15 000 g (30 min, 4 °C) before the supernatant was collected. The protein extracts were quantified using a Characterization of mesothelin expression in WF-0, WF-0/mock BCA Protein Assay Kit (Pierce, Rockford, IL, USA). and Meso-VAX vaccines The mesothelin expressions in WF-0, WF-0/mock and Meso-VAX vaccines Mice fi were rst characterized by RT-PCR. Total RNA of the cell lines was isolated Six- to 8-week-old female C57BL/6J mice were purchased and kept in the ’ using a TRIzol RNA isolation kit following the manufacturer s instructions animal facility of the School of Medicine of National Taiwan University. All (Invitrogen Life Technologies). Reverse-transcribed cDNA products of the animal procedures were performed according to approved fi fi were ampli ed by PCR with primers speci c for mesothelin and protocols and in accordance with recommendations for the proper use GAPDH. The primers used to detect human mesothelin were and care of laboratory animals. The mice were randomized without ′ ′ ′ ′ 5 -AGCCTCCTGTTCCTGCTCTT-3 and 5 -GGGAGCAGGTCCACATTG-3 . The blinding for the experiments without any specific selection. GAPDH forward primer was 5′-ACCCAGAAGACTGTGGATGG-3′ and the reverse primer was 5′-TGCTGTAGCCAAATTCGTTG3′. The amplification products were separated by 1% agarose gel electrophoresis and visualized Vaccination by ethidium bromide staining. WF-0, WF-0/mock or Meso-VAX was first irradiated as a cell-based vaccine Human mesothelin expressions in WF-0, WF-0/mock and Meso-VAX as described earlier. The mice (n = 10 per group) were subcutaneously vaccines were further detected by flow-cytometry analysis. Briefly, the cells immunized with the respective vaccine (1 × 106 cells/mouse) on day 0 with − were incubated with 2 μgml 1 of anti-mesothelin mAb CAK-1 (catalog no. or without intra-peritoneal injections of 1 × 109 ifu/mouse AAV-GFP, AAV- ab3362, Abcam Biochemicals, Cambridge, MA, USA) diluted with 500 μlof IL-12 or AAV-GM-CSF on day 1, or 2.5 μg/mouse rmIL-12 (R&D Systems, PBS containing 5% bovine serum albumin and 0.02% sodium azide. After Minneapolis, MN, USA) subcutaneously from days 0 to 4 and days 14 to 18 incubation for 1 h at 4 °C, the cells were washed once with PBS and daily in the inguinal region. After 1 week, the mice were boosted with the incubated with a 1:200 dilution of goat anti-mouse FITC-conjugated (Santa same regimens except for rmIL-12 groups. Mice injected with PBS alone Cruz Biotechnology, Dallas, TX, USA) Abs for 1 h. After washing twice with were used as the negative control. PBS, the cells were suspended in 0.5 ml PBS and evaluated using a fl FACS Calibur ow cytometer and analyzed with CellQuest software γ fl (BD Biosciences, San Jose, CA, USA). The rat IgG isotype was used as a Surface markers with intracellular IFN- cytokine staining and ow- control. cytometry analysis Mice (n = 5 per group) were immunized as described earlier. The mice were killed and splenocytes were obtained and prepared 1 week after the last Preparation and generation of AAV containing IL-12 (AAV-IL-12) or immunization. Before intracellular cytokine staining, 2 × 106 pooled GM-CSF cytokine splenocytes of each group were incubated for 16 h with either 1 μgml− 1 An AAV containing mouse interleukin-12 cDNA (AAV-IL-12) or GM-CSF mesothelin peptide (aa 406-414, GQKMNAQAI) containing an MHC class I cDNA (AAV-GM-CSF), or a GFP gene (AAV-GFP) under the control of a CMV epitope25 to detect mesothelin-specific CD8+ cytotoxic T lymphocyte immediate early gene promoter was provided by Prof. BL Chiang (College precursors or 100 μgml− 1 of WF-3 cell lysates to detect mesothelin- of Medicine, National Taiwan University).34 AAV-IL-12, AAV-CM-CSF and specific CD4+ helper T lymphocyte precursors. AAV-GFP were amplified using 293-AD cells. Briefly, 2 × 106 293-AD cells Golgistop (Pharmingen, San Diego, CA, USA) was added 6 h before were plated into a 100-mm culture dish in 5 ml of complete medium harvesting the cells, which were first stained with PE-conjugated CD4

© 2016 Macmillan Publishers Limited Gene Therapy (2016) 38 – 49 Antigen-specific cell-based immunotherapy M-C Chang et al 48 (catalog no. 553049) or CD8 Ab (catalog no. 553032, Pharmingen), and growth using the IVIS system twice a week until they died or were then subjected to intracellular cytokine staining using a Cytofix/Cytoperm sacrificed on day 90. kit according to the manufacturer’s instructions (Pharmingen). Intracellular γ cytokine staining for FITC-conjugated IFN- (catalog no. 554411, Pharmin- fi gen) or immunoglobulin isotype control Ab (rat IgG1) (catalog no. 562026, Antibody-speci c complement-dependent cell-mediated Pharmingen) was then performed.48 Data were collected using a FACS cytotoxicity assay Calibur flow cytometer and analyzed with CellQuest software as described To evaluate whether the Meso-VAX vaccine combined with AAV-IL-12 earlier (BD Biosciences, San Jose, CA, USA).48 could also generate complement-dependent cell-mediated cytotoxicity to tumor cells, complement-dependent toxicity assays were performed as described previously.25 Briefly, WF-3/Luc tumor cells were seeded onto a Anti-mesothelin Abs and IL-12 concentrations in sera by ELISA 96-well plate (5 × 103 per well) overnight, and 0, 10 and 50 μl of sera To detect anti-mesothelin Ab, the mice were immunized as described collected from the various vaccinated groups were then added to the well. earlier. Sera were prepared from the mice 14 days after the last Naive rabbit serum at a final dilution of 1:5 was used for complement immunization. To detect mesothelin-specific Abs in the sera, direct ELISA (Sigma-Aldrich, St Louis, MO, USA) with culture medium in a total volume fi 18 fl was used as described previously with some modi cations. Brie y, a of 100 μl. After incubation for 18 h, cell viability was measured according to μ 96-microwell plate was coated with 0.05 g per well of murine mesothelin bioluminescent activity using the IVIS Imaging System. recombinant protein (Abnova, Taipei, Taiwan) and incubated at 4 °C overnight. The wells were then blocked with PBS containing 20% fetal bovine serum. Sera were serially diluted in PBS, added to the ELISA wells, Statistical analysis and incubated at 37 °C for 2 h. All data were expressed as mean ± s.e.m. and were representative of at After washing with PBS containing 0.05% Tween-20, the plate was least two different experiments. Data for surface marker staining with flow- incubated with a 1:2000 dilution of peroxidase-conjugated rabbit anti- cytometry analysis and tumor treatment experiments were evaluated by mouse IgG antibody (catalog no. MBS674902, Biosource, Camarillo, CA, ANOVA. Comparisons between individual data points were made using the USA) at room temperature for 1 h. The plate was washed, developed with Student’s t-test. Survival curves were generated using the Kaplan–Meier 1-Step Turbo TMB-ELISA (Clinical Science Lab, Mansfield, MA, USA), and method and differences were calculated using the log-rank test. Statistical stopped with 1 M H2SO4.The ELISA plate was read by a standard ELISA significance was set at Po0.05. reader (Molecular Devices, Sunnyvale, CA, USA) at 450 nm. To detect the sera concentrations of IL-12 in the mice vaccinated with Meso-VAX combined with AAV-IL-12, the sera of the vaccinated mice were CONFLICT OF INTEREST collected and prepared at the indicated intervals as described earlier. The The authors declare no conflict of interest. IL-12 concentrations were detected by ELISA (R&D Systems).

In vivo tumor protection experiments ACKNOWLEDGEMENTS For the tumor protection experiments, the mice (n = 10 per group) were This work was supported in part by the 2nd and 7th core laboratory facilities of the vaccinated as described earlier. One week after the last vaccination, the Department of Medical Research of National Taiwan University Hospital. The lentiviral mice were intraperitoneally challenged with 1 × 105 WF-3/Luc tumor cells. vector used to construct WF-3/Luc was obtained from the National RNAi Core Facility Bioluminescent tumor images were used to detect tumor growth, which Platform of the Institute of Molecular Biology/Genomic Research Center, Academia were measured using an IVIS Imaging System Series 200 (Xenogen, Sinica. This study was supported by National Science Committee of Taiwan (NSC 100- Alameda, CA, USA) twice a week until the mice died or were killed at the 2325-B-002-004, MOST 103-2314-B-281-008-MY3) and the National Core Facility indicated time points. Program for Biotechnology Grants of the National Science Committee of Taiwan To detect bioluminescence signals, the mice were first injected with (NSC 100-2319-B-001-002). 300 μlof15mgml− 1 luciferin (Promega) intraperitoneally, and then imaged 10 min later. The bioluminescence signals were then acquired for 3 min. REFERENCES 1 Chiang YC, Chen CA, Chiang CJ, Hsu TH, Lin MC, You SL et al. Trends in incidence In vitro tumor killing activity of splenocytes of various vaccinated and survival outcome of epithelial ovarian cancer: 30-year national population- groups based registry in Taiwan. J Gynecol Oncol 2013; 24:342–351. fl The in vitro tumor killing activity of the splenocytes of various vaccinated 2Peiderer A. Diagnosis and staging of ovarian cancer. J Cancer Res Clin Oncol 1984; – groups on tumor cells was further evaluated as described previously with 107:81 88. some modifications.49 Briefly, splenocytes of the mice in various vaccinated 3 Gonzalez-Diego P, Lopez-Abente G, Pollan M, Ruiz M. Time trends in ovarian – groups were co-cultured with WF-3/LG (10:1, 5:1 or 2:1 ratio) in a 96-well cancer mortality in Europe (1955 1993): effect of age, birth cohort and period – plate (1 × 104 cells/well) for 18 h. Luciferin (Promega) was then added and of death. Eur J Cancer 2000; 36:1816 1824. fl 4 Agarwal R, Kaye SB. Ovarian cancer: strategies for overcoming resistance to total ux (p/s) from each well was measured using the IVIS Imaging System – Series 200. chemotherapy. Nat Rev Cancer 2003; 3:502 516. 5 DiSaia PJ, Bloss JD. Treatment of ovarian cancer: new strategies. Gynecol Oncol 2003; 90: S24–S32. In vivo tumor treatment experiments 6 Greenlee RT, Hill-Harmon MB, Murray T, Thun M. Cancer statistics, 2001. CA Cancer For the in vivo tumor treatment experiments, the mice were intraper- J Clin 2001; 51:15–36. itoneally injected with 1 × 105 WF-3/Luc tumor cells on day 0 as described 7 Ozols RF. Recurrent ovarian cancer: evidence-based treatment. J Clin Oncol 2002; previously.50 Two days after the tumor challenge, the mice were 20:1161–1163. immunized as described earlier. The mice received a booster with the 8 Kruger C, Greten TF, Korangy F. Immune based therapies in cancer. Histol same regimen every 7 days for 4 weeks. Mice receiving PBS were used as Histopathol 2007; 22:687–696. the negative control. Tumor images were then measured using the IVIS 9 Borges VF, Kufe D, Avigan DE. Update on cancer vaccines. Curr Opin Gastroenterol Imaging System Series 200 as described earlier. 2002; 18:723–731. 10 Banchereau J, Palucka AK. Dendritic cells as therapeutic vaccines against cancer. Nat Rev Immunol 2005; 5: 296–306. In vivo antibody depletion experiments 11 Fiander A, Man S, Borysiewicz LK, Wilkinson GW. Therapeutic vaccines for cervical In vivo antibody depletion experiments were performed as described cancer: concept and clinical results. BioDrugs 1997; 8:331–338. previously.51 Briefly, the mice were first immunized and boosted 1 week 12 Mackiewicz A, Mackiewicz J, Wysocki PJ, Wiznerowicz M, Kapcinska M, Laciak M later and then challenged with WF-3/Luc tumor cells 7 days after the last et al. Long-term survival of high-risk melanoma patients immunized with a Hyper- vaccination. Depletion was started 1 week before tumor challenge with the IL-6-modified allogeneic whole-cell vaccine after complete resection. Expert Opin monoclonal antibodies GK1.5, 2.43 and PK136 used for CD4+, CD8+ and Investig Drugs 2012; 21: 773–783. NK1.1 depletion, respectively.48,51 Depletion was terminated on day 70 13 Chiang CL, Kandalaft LE, Coukos G. Adjuvants for enhancing the immunogenicity after the tumor challenge. The mice were monitored for evidence of tumor of whole tumor cell vaccines. Int Rev Immunol 2011; 30:150–182.

Gene Therapy (2016) 38 – 49 © 2016 Macmillan Publishers Limited Antigen-specific cell-based immunotherapy M-C Chang et al 49 14 Ward S, Casey D, Labarthe MC, Whelan M, Dalgleish A, Pandha H et al. Immu- 34 Chang CJ, Chen YH, Huang KW, Cheng HW, Chan SF, Tai KF et al. Combined GM- notherapeutic potential of whole tumour cells. Cancer Immunol Immunother 2002; CSF and IL-12 gene therapy synergistically suppresses the growth of orthotopic 51: 351–357. liver tumors. Hepatology 2007; 45:746–754. 15 Hassan R, Bera T, Pastan I. Mesothelin: a new target for immunotherapy. Clin 35 Nanni P, Nicoletti G, De Giovanni C, Landuzzi L, Di Carlo E, Cavallo F et al. Cancer Res 2004; 10: 3937–3942. Combined allogeneic tumor cell vaccination and systemic interleukin 12 prevents 16 Scholler N, Fu N, Yang Y, Ye Z, Goodman GE, Hellstrom KE et al. Soluble member mammary carcinogenesis in HER-2/neu transgenic mice. J Exp Med 2001; 194: (s) of the mesothelin/megakaryocyte potentiating factor family are detectable in 1195–1205. sera from patients with ovarian carcinoma. Proc Natl Acad Sci USA 1999; 96: 36 Lenzi R, Edwards R, June C, Seiden MV, Garcia ME, Rosenblum M et al. Phase II 11531–11536. study of intraperitoneal recombinant interleukin-12 (rhIL-12) in patients with 17 Cheng WF, Huang CY, Chang MC, Hu YH, Chiang YC, Chen YL et al. High meso- peritoneal carcinomatosis (residual diseaseo1cm) associated with ovarian cancer thelin correlates with chemoresistance and poor survival in epithelial ovarian or primary peritoneal carcinoma. J Transl Med 2007; 5:66–70. carcinoma. Br J Cancer 2009; 100: 1144–1153. 37 Jiao L, Gao X, Joyee AG, Zhao L, Qiu H, Yang M et al. NK cells promote type 1 T cell 18 Huang CY, Cheng WF, Lee CN, Su YN, Chien SC, Tzeng YL et al. Serum mesothelin immunity through modulating the function of dendritic cells during intracellular in epithelial ovarian carcinoma: a new screening marker and prognostic factor. bacterial infection. J Immunol 2011; 187: 401–411. Anticancer Res 2006; 26: 4721–4728. 38 Vom Berg J, Vrohlings M, Haller S, Haimovici A, Kulig P, Sledzinska A et al. 19 Einama T, Homma S, Kamachi H, Kawamata F, Takahashi K, Takahashi N et al. Intratumoral IL-12 combined with CTLA-4 blockade elicits T cell-mediated glioma Luminal membrane expression of mesothelin is a prominent poor prognostic rejection. J Exp Med 2013; 210: 2803–2811. factor for gastric cancer. Br J Cancer 2012; 107:137–142. 39 Starbeck-Miller GR, Xue HH, Harty JT. IL-12 and type I interferon prolong the 20 Parinyanitikul N, Blumenschein GR, Wu Y, Lei X, Chavez-Macgregor M, Smart M division of activated CD8 T cells by maintaining high-affinity IL-2 signaling in vivo. et al. Mesothelin expression and survival outcomes in triple receptor negative J Exp Med 2014; 211:105–120. breast cancer. Clin Breast Cancer 2013; 13:378–384. 40 Beum PV, Lindorfer MA, Beurskens F, Stukenberg PT, Lokhorst HM, Pawluczko- 21 Cheng WF, Hung CF, Chai CY, Chen CA, Lee CN, Su YN et al. Generation and wycz AW et al. Complement activation on B lymphocytes opsonized with ritux- characterization of an ascitogenic mesothelin-expressing tumor model. Cancer imab or ofatumumab produces substantial changes in membrane structure 2007; 110: 420–431. preceding cell lysis. J Immunol 2008; 181: 822–832. 22 Andersen MH, Sorensen RB, Schrama D, Svane IM, Becker JC, Thor Straten P. 41 Tang Z, Feng M, Gao W, Phung Y, Chen W, Chaudhary A et al. A human Cancer treatment: the combination of vaccination with other therapies. Cancer single-domain antibody elicits potent antitumor activity by targeting an Immunol Immunother 2008; 57: 1735–1743. epitope in mesothelin close to the cancer cell surface. Mol Cancer Ther 2013; 12: 23 de Gruijl TD, van den Eertwegh AJ, Pinedo HM, Scheper RJ. Whole-cell cancer 416–426. vaccination: from autologous to allogeneic tumor- and dendritic cell-based vac- 42 Gritzapis AD, Mahaira LG, Perez SA, Cacoullos NT, Papamichail M, Baxevanis CN. cines. Cancer Immunol Immunother 2008; 57: 1569–1577. Vaccination with human HER-2/neu (435-443) CTL peptide induces effective 24 Chang MC, Chen CA, Chen PJ, Chiang YC, Chen YL, Mao TL et al. Mesothelin antitumor immunity against HER-2/neu-expressing tumor cells in vivo. Cancer Res enhances invasion of ovarian cancer by inducing MMP-7 through MAPK/ERK and 2006; 66:5452–5460. JNK pathways. Biochem J 2012; 442: 293–302. 43 Manno CS, Chew AJ, Hutchison S, Larson PJ, Herzog RW, Arruda VR et al. AAV- 25 Chang CL, Wu TC, Hung CF. Control of human mesothelin-expressing tumors by mediated factor IX gene transfer to skeletal muscle in patients with severe DNA vaccines. Gene Ther 2007; 14: 1189–1198. hemophilia B. Blood 2003; 101: 2963–2972. 26 Chang MC, Chen CA, Hsieh CY, Lee CN, Su YN, Hu YH et al. Mesothelin inhibits 44 Marr RA, Addison CL, Snider D, Muller WJ, Gauldie J, Graham FL. Tumour paclitaxel-induced apoptosis through the PI3K pathway. Biochem J 2009; 424: immunotherapy using an adenoviral vector expressing a membrane-bound 449–458. mutant of murine TNF alpha. Gene Ther 1997; 4:1181–1188. 27 Hassan R, Bullock S, Premkumar A, Kreitman RJ, Kindler H, Willingham MC et al. 45 Richter M, Iwata A, Nyhuis J, Nitta Y, Miller AD, Halbert CL et al. Adeno-associated Phase I study of SS1P, a recombinant anti-mesothelin immunotoxin given as a virus vector transduction of vascular smooth muscle cells in vivo. Physiol Genomics bolus I.V. infusion to patients with mesothelin-expressing mesothelioma, ovarian, 2000; 2:117–127. and pancreatic cancers. Clin Cancer Res 2007; 13: 5144–5149. 46 Koeberl DD, Alexander IE, Halbert CL, Russell DW, Miller AD. Persistent expression 28 Hassan R, Ebel W, Routhier EL, Patel R, Kline JB, Zhang J et al. Preclinical evaluation of human clotting factor IX from mouse liver after intravenous injection of adeno- of MORAb-009, a chimeric antibody targeting tumor-associated mesothelin. associated virus vectors. Proc Natl Acad Sci USA 1997; 94: 1426–1431. Cancer Immun 2007; 7:20–29. 47 Malecki M, Proczka R, Chorostowska-Wynimko J, Swoboda P, Delbani A, Pachecka J. 29 Jaffee EM, Hruban RH, Biedrzycki B, Laheru D, Schepers K, Sauter PR et al. Novel Recombinant adeno-associated viruses (rAAV2) facilitate the intraperitoneal gene allogeneic granulocyte-macrophage colony-stimulating factor-secreting tumor delivery to cancer cells. Oncol Lett 2010; 1:177–180. vaccine for pancreatic cancer: a phase I trial of safety and immune activation. 48 Cheng WF, Chang MC, Sun WZ, Lee CN, Lin HW, Su YN et al. Connective tissue J Clin Oncol 2001; 19: 145–156. growth factor linked to the E7 tumor antigen generates potent antitumor 30 Thomas AM, Santarsiero LM, Lutz ER, Armstrong TD, Chen YC, Huang LQ et al. immune responses mediated by an antiapoptotic mechanism. Gene Ther 2008; Mesothelin-specific CD8(+) T cell responses provide evidence of in vivo cross- 15: 1007–1016. priming by antigen-presenting cells in vaccinated pancreatic cancer patients. 49 Huang CT, Chang MC, Chen YL, Chen TC, Chen CA, Cheng WF. Insulin-like growth J Exp Med 2004; 200: 297–306. factors inhibit dendritic cell-mediated anti-tumor immunity through regulating 31 Steitz J, Bruck J, Steinbrink K, Enk A, Knop J, Tuting T. Genetic immunization of ERK1/2 phosphorylation and p38 dephosphorylation. Cancer Lett 2015; 359: mice with human tyrosinase-related protein 2: implications for the immu- 117–126. notherapy of melanoma. Int J Cancer 2000; 86:89–94. 50 Chen YL, Chang MC, Chen CA, Lin HW, Cheng WF, Chien CL. Depletion of reg- 32 Lu Y, Wei YQ, Tian L, Zhao X, Yang L, Hu B et al. Immunogene therapy of tumors ulatory T lymphocytes reverses the imbalance between pro- and anti-tumor with vaccine based on xenogeneic epidermal growth factor receptor. J Immunol immunities via enhancing antigen-specific T cell immune responses. PLoS One 2003; 170: 3162–3170. 2012; 7: e47190. 33 Okada N, Iiyama S, Okada Y, Mizuguchi H, Hayakawa T, Nakagawa S et al. 51 Chen CA, Ho CM, Chang MC, Sun WZ, Chen YL, Chiang YC et al. Metronomic Immunological properties and vaccine efficacy of murine dendritic cells simul- chemotherapy enhances antitumor effects of cancer vaccine by depleting reg- taneously expressing melanoma-associated antigen and interleukin-12. Cancer ulatory T lymphocytes and inhibiting tumor angiogenesis. Mol Ther 2010; 18: Gene Ther 2005; 12:72–83. 1233–1243.

Supplementary Information accompanies this paper on Gene Therapy website (http://www.nature.com/gt)