Pathogen-free C57BL/6 (B6) mice were purchased from Charles River at 6 to 8 weeks of age. Rag1 ^-/- (B6.129S7-Rag1^tm1Mom /J) mice were purchased from the Jackson Laboratory. CD1d^-/- mice with B6 genetic background were kindly provided from L. Van Kaer (Vanderbilt University, Nashville, TN). Ly5.1 congenic OT-1 mice were generated by cross/backcross breeding of OT-1 with B6. Ly5.1 mice. All mice were maintained under specific pathogen-free conditions at the RIKEN animal facility, and all procedures were performed in compliance with the protocols approved by the Institutional Animal Care Committee at RIKEN (AEY2022-020). NIH3T3 cells were obtained from the RIKEN Cell Bank. NKT hybridoma (1.B2) and B16 cell line were kindly provided by Dr. M. Kronenberg (La Jolla Institute) and Dr. RM Steinman (Rockefeller University) respectively. All cell lines were maintained and treated according to the supplier’s recommendations. B16-PSMA, B16-PSA, and B16-PAP cells were generated by retroviral transduction of human PSMA, PSA, or PAP genes into B16 cells.

The spleen cells were preincubated with anti-CD16/CD32 (BioLegend) in PBS with 2% heat-inactivated FBS and then stained with surface antibodies and viability dye for 30 min on ice. A Fixable Aqua or Violet-Dead Cell Stain Kit (Invitrogen) was used to remove the dead cells. The cytokine production was analyzed by intracellular staining using a BD Intracellular Cytokine Staining Kit (BD Biosciences). For these, LSRFortessa X-20 or Canto II instrument and CELLQuest, FACSDiva (BD Biosciences), or the FlowJo software (v10.3B2) were used.

Murine CD1d and OVA plasmids used in this study was previously described (29). Coding sequences for PSMA (GenBank Accession Number: NM_004476), PSA (GenBank Accession Number: NM_001648), and PAP (GenBank Accession Number: NM_001099) were generated via gene synthesis (Takara) and cloned into the HindIII and BamHI, HindIII and EcoRI, and HindIII and BamHI sites of the pGEM-4Z vector (Promega), respectively. The resultant plasmid was then linearized with BamHI and EcoRI and purified using the QIAquick PCR Purification Kit (Qiagen). In vitro transcription was conducted using the mMessage mMachine T7 Ultra Kit (Ambion), and then RNA was purified using the RNeasy Mini/Midi Kit (Qiagen). aAVCs were prepared as previously described (27). Briefly, NIH3T3 cells resuspended in OptiMEM and RNA were transferred simultaneously to a cuvette. This cell suspension was pulsed in an ECM 830 Square Wave Electroporation System (Harvard Apparatus). After electroporation, the cells were transferred to a culture medium and cultured in 500 ng/mL of α-GalCer. Protein expression in the transfected cells was assessed by flow cytometry for CD1d, ELISA (ITEA) for OVA and western blot analysis for antigen proteins, such as PSMA, PSA and PAP.

To quantify the protein levels of prostate antigens (PSMA, PSA, and PAP), western blotting was used as previously described (35). Briefly, the relevant antigen-expressing aAVC was lysed in Laemmli sample buffer (Bio-Rad) and boiled at 95°C. Samples were loaded onto a polyacrylamide gel and electrotransferred to a PVDF membrane. Membranes were blocked and incubated with primary anti-PSMA, anti-PSA, or anti-PAP antibodies overnight at 4°C with gentle shaking. The membranes were incubated with horseradish peroxidase (HRP)-conjugated goat anti-mouse antibody (R&D Systems) at room temperature. Chemiluminescence images were obtained using a luminescence image analyzer (LAS 1000) and Image Gauge software (Fujifilm Co.). Protein levels were evaluated using the Image Gauge software, and image processing (resizing, cropping, and merging) was conducted using Adobe Photoshop (Adobe Systems).

Ten thousand mouse aAVCs were co-cultured with 1 × 10⁵ NKT hybridoma (1.B2) for 24 h in 96-well round-bottom plates. Culture supernatants were analyzed for mouse IL-2 production using ELISA for cytokines with a mouse IL-2 Duo set (R&D Systems).

Murine IFN-γ-secreting CD8⁺ T cells were performed by ELISPOT assays as previously described (27). Ninety six well filtration plates were coated with rat anti-mouse IFN-γ (R4-6A2; BD) overnight at 10 µg/mL. CD8⁺ T cells were isolated from the spleens of naïve or aAVC-treated mice by using CD8 MACS beads (Miltenyi Biotec). To prepare for antigen-presenting cells (APCs), splenic DCs from naïve mice were isolated using CD11c MACS beads. Additionally, CD8⁺ T cells (5 × 10⁵/well) were co-cultured with DCs (1 × 10⁵/well) pulsed with or without PepTivator (Miltenyi Biotec) for 24 h. After the culture, the plates were incubated with biotinylated anti-mouse IFN-γ (XMG1.2; BD) (2 µg/mL) for 2 h. Finally, IFN-γ-secreting spots were developed with streptavidin-HRP (BD) and stable DAB substrate (Research Genetics). IFN-γ SFCs were quantified using microscopy.

For the prophylactic model for each antigen, female C57BL/6 mice were immunized intravenously with 5 × 10⁵ aAVC expressing the relevant antigen and inoculated subcutaneously with 1 × 10⁵ B16 cells expressing the relevant antigen after two weeks. For the therapeutic model, C57BL/6 mice were injected subcutaneously with 5 × 10⁵ tumor cells and treated with 5 × 10⁵ aAVC when tumor size reached around 50mm³. Tumor diameters were measured using a Vernier caliper every 2–3 days, and the tumor volume was calculated using the following equation: V = L × W2 × 0.52, where V is the volume, L is the length, and W is the width. Mice with high tumor burden (volume > 2000 mm³) were sacrificed under anesthesia.

StatMate V software (Nihon 3 B Scientific Inc., Japan) was used for the statistical analyses. A student’s t-test (two-tailed) was used for the analysis of differences between the two groups. One-way analysis of variance (ANOVA) and post-hoc Dunnett’s tests were used for multiple comparisons. P values less than 0.05 were considered significant (*P<0.05; **P<0.01; ***P<0.001).

Mouse aAVCs were created by co-transfecting α-GalCer-loaded, mCD1d mRNA NIH3T3 cells with three human prostate antigen mRNAs (PSMA-mRNA, PSA-mRNA, or PAP-mRNA) and denoted as aAVC-PSMA (Figures 1A, B), aAVC-PSA (Figures 1C, D), and aAVC-PAP (Figures 1E, F), respectively. After transfection, we verified the expression of each protein using western blotting (Figures 1A, C, E). aAVC-PSMA, aAVC-PSA, and aAVC-PAP expressed the corresponding proteins at 1546 ± 202, 564 ± 97, and 172 ± 34 ng/10⁶ cells, respectively. We verified that the expression rate of mCD1d in each aAVC was > 90% (Figures 1B, D, F).

After verifying the establishment of each aAVC type, we evaluated whether aAVC-PSMA, aAVC-PSA, and aAVC-PAP induced iNKT cell activation. We examined the iNKT response by culturing these aAVCs with iNKT cell hybridomas in vitro. After culturing aAVC-PSMA, aAVC-PSA, and aAVC-PAP with iNKT cells, IL-2 was produced by iNKT in the presence of aAVC, but not parental NIH cells in all groups (Figures 2A, C, E). Next, we investigated the capacity of each aAVC to expand iNKT cells in vivo. The percentage of iNKT cells in the spleen was examined three days after administering each aAVC. The iNKT cell frequency in all aAVC-PSMA-, aAVC-PSA-, and aAVC-PAP-injected mice was five times higher than that in the naïve mice (Figures 2B, D, F). Therefore, aAVC-PSMA, aAVC-PSA, and aAVC-PAP similarly stimulated iNKT cells in vivo, with aAVC-PSMA demonstrating the highest potency as an activator of iNKT cells in this context.

Owing to its resistance to ICB, B16 melanoma is widely recognized as a cold tumor. To evaluate whether the three types of prostate antigen-expressing aAVCs exhibit antitumor effects, we established three permanent B16 cell lines: B16 expressing PSMA, PSA, or PAP (Figures 3A, C, E). C57BL/6 mice were intravenously administered aAVC-PSMA, aAVC-PSA, or aAVC-PAP individually and then inoculated with PSMA-, PSA-, or PAP-expressing melanoma cell lines (B16-PSMA, B16-PSA, and B16-PAP) respectively after 14 days. Compared with the mice in the untreated control group, the aAVC-PSMA-, aAVC-PSA-, or aAVC-PAP-vaccinated mice were protected from the growth of B16-PSMA, B16-PSA, and B16-PAP (Figures 3B, D, F). Vaccination with monovalent prostate antigen-expressing aAVC exhibited substantial antitumor effects.

Next, we examined the relationship between iNKT cell frequency and antigen-specific T cell immunity in aAVC therapy. For this, we transferred different number of iNKT cells to Rag1^-/- mice. We harvested B cell-depleted MNCs from spleen and liver of WT mice and CD1d^-/- mice that are deficient in iNKT cells at the ratio of 100:0(%) or 50:50 (%) to Rag1^-/- mice (Figures 4A, B). We verified that the number of iNKT cells (1^st; 0.74x 10⁶ cells/mouse, 2^nd 0.53 x 10⁶ cells/mouse) included in MNCs from WT mice and CD1d^-/- mice 50:50 (%) was half the number of iNKT cells included n MNCs from WT mice (1^st; 1.41x 10⁶ cells/mouse, 2^nd 1.05x 10⁶ cells/mouse). Then, to trace and measure the antigen-specific T cell response, we administered low number of OT-1 cells (1x 10⁵ cells/mouse) and subsequently aAVC-OVA cells the following day. Despite the reduced frequency of iNKT cells, the OVA-specific CD8⁺ T cell response in mice given iNKT cells from WT mice and CD1d^-/- mice wassimilar to those from WT mice-derived iNKT cells transferred Rag1^-/- mice (Figures 4C, D). Thus, although the activation of iNKT cells is important factor, even if low frequent iNKT cells in vivo can show the adjuvant activity for T cell priming sufficiently in aAVC therapy.

Next, we investigated whether aAVC therapy had a therapeutic effect on prostate antigen-expressing tumor-bearing mice. B16-PSMA, B16-PSA, and B16-PAP cells were inoculated subcutaneously into mice individually. When we observed tumors measuring around 50 mm³ in all tumor-bearing mice, they were treated with aAVC-PSMA (Figure 5A), aAVC-PSA (Figure 5B), or aAVC-PAP (Figure 5C), respectively. Each aAVC exhibited tumor abrogation individually (Figures 5A–C). Among these, aAVC-PSA appears to have the most potent therapeutic effect.

Next, we investigated whether aAVC therapy induces antigen spreading in a therapeutic model. We established three antigen-expressing tumor cell lines by introducing DNA of PSMA, PSA, and PAP into B16 cells (hereafter, B16-PSMA/PSA/PAP) (Figure 6A). We verified that these cells expressed each antigen using RT-PCR and western blot analyses (Figure 6B). B16-PSMA/PSA/PAP-bearing mice were treated with monovalent prostate cancer antigen expressing aAVC, that is, either aAVC-PSMA, aAVC-PSA, or aAVC-PAP on day 7. We found that immunization with aAVC-PSMA, aAVC-PSA, or aAVC-PAP abrogated the tumors (Figures 6C, E, G). After verifying the antitumor effect on day 16, we sacrificed the mice. We analyzed the antigen-specific CD8⁺ T cell response: PSMA-specific T cell response in aAVC-PSMA-treated mice (Figure 6D), PSA-specific T cell response in aAVC-PSA-treated mice (Figure 6F), and PAP-specific T cell response in aAVC-PAP-treated mice (Figure 6H). We detected a higher antigen-specific CD8⁺ T cell response following treatment with the relevant antigen-expressing aAVC. We found the antitumor effects through the therapy with aAVC-PSMA, aAVC-PSA or aAVC-PAP respectively (Figures 6C, E, G). In addition, intriguingly, we found that monovalent-aAVC induced not only CD8⁺ T cell responses against the target antigen, but also elicited antigen-specific CD8⁺ T cell responses against other tumor antigens simultaneously. This discrepancy might be explained by the capacity of aAVC therapy to induce significant epitope spreading from the tumor to CD8⁺ T cells.

Subsequently, we generated a multivalent aAVC expressing the three prostate cancer antigens (Figure 7A). After simultaneous transfection with PSMA, PSA, and PAP mRNA in addition to murine CD1d mRNA, we verified the expression of PSMA, PSA, and PAP protein by western blot analysis (Figure 7B). The trivalent cancer antigen aAVC, denoted as aAVC-PROS, effectively expressed PSMA (483 ± 31ng/10⁶ cells), PSA (59.2 ± 5.0 ng/10⁶ cells), and PAP (17.9 ± 1.2 ng/10⁶ cells). CD8⁺ T cells were isolated from the spleen cells of vaccinated mice that had been administered aAVC-triPROS to analyze the antigen-specific T cell response. At day 7, they were cocultured with splenic DCs from naïve mice in the presence or absence of PSMA, PSA or PAP peptivators. The antigen-specific CD8⁺ T cells produced in response to the peptides were determined using the ELISpot assay (Figure 7C). Although the amounts of each protein expressed in aAVC-PROS were different, the T cell response induced by aAVC-PROS was higher in PSA-specific T cells, followed in order by PAP-specific T cells and PSMA-specific T cells. The amount of protein expressed in the aAVC did not always correlate with the T cell immune response. T cell immune response depends on the amount of antigen present and antigenicity of the specific expressed protein. Our results revealed that the response depends on the antigen quantity when the antigens are identical. However, when dealing with different antigens, the response relies more on antigenicity than the amount of antigen present in the aAVCs. This suggests that the T cell response may depend on the immunogenicity of the antigen involved in aAVC when we use multifunctional epitope-expressing aAVC.

Having determined that monovalent antigen-expressing aAVC therapy is effective against multiple antigen-expressing tumors and that this effect was caused by epitope spreading, we subsequently investigated whether divalent or trivalent antigen-expressing aAVC generated antitumor effects more effectively. Because PSMA and PSA were more effective than PAP, suggesting that combining these two antigens would be more potent (Figure 6), we established a divalent aAVC expressing PSMA and PSA. We compared the antitumor effects of divalent aAVC-PSMA/PSA and trivalent aAVC-PROS against B16-PSMA/PSA/PAP (Figure 8A). Both vaccinations showed an abrogation of tumor cells similarly. Further analysis indicates that divalent aAVC is better than monovalent aAVC; however, there was no difference in efficacy between divalent and trivalent aAVCs. Finally, we compared the antitumor effects of the three vaccines during the early and late tumor phases. When we assessed the tumor size on days 10 (early) and 19 (late), we did not observe any difference between divalent and trivalent aAVC vaccination until day 10, but found that the antitumor effect of trivalent aAVC was superior to that of monovalent or divalent aAVC on day 19 (Figure 8B). These findings imply that even monovalent aAVC show the antitumor effect sufficiently during the early phase. Notably, only trivalent antigen-expressing aAVC exhibits increased potency in the late phase. Thus, multivalent antigen-expressing aAVC can elicit multiple T cell responses more effectively than monovalent antigen-expressing aAVC for the long period.