Mice and study approval. Five- to 6-week-old male C57BL/6 mice were purchased from The Jackson Laboratory and were allowed to acclimate for 1 week prior to the experiments.

Cell cultures. The murine pancreatic cancer cell line, Panc02, was maintained in DMEM supplemented with 10% FBS (Gemini Bio Products), 1% L-Glutamine (Thermo Fisher Scientific, catalog 25030-081), 0.5% penicillin/streptomycin (Thermo Fisher Scientific, catalog 15140-122) at 37°C in 10% CO2. T2-H2-Kb, T2-H2-Db, and T2-IAb cells were maintained in RPMI supplemented with 10 % FBS, 1% L-Glutamine (Thermo Fisher Scientific), 1% penicillin/streptomycin (Invitrogen), 1% MEM Non-Essential Amino Acids (Thermo Fisher Scientific, catalog 11140-050), and 1% sodium pyruvate (Thermo Fisher Scientific, catalog 11360-070). T2-H2-Kb and T2-H2-Db cell lines and the T2-IAb cell line were subject to selection in 1 mg/mL and 0.25 mg/mL Geneticin (Thermo Fisher Scientific, catalog 10131-027), respectively. All cell lines were confirmed to remain free of Mycoplasma through regular testing at the Johns Hopkins University Genetic Resources Core Facility.

Primary murine splenocytes or TILs were maintained in RPMI supplemented with 10% FBS, 0.5% L-glutamine, 1% penicillin/streptomycin, and 1× BME (Thermo Fisher Scientific, catalog 21985-023) at 37°C in 5%CO2.

Primary murine BMDCs were prepared by isolating bone marrow from non–tumor-bearing C57BL/6 mice per the Jove protocol (42). Isolated bone marrow was treated with ACK lysis buffer, washed with complete lymphocyte media, and then resuspended at 0.5 × 106 cells/mL in media supplemented with 20 ng/mL murine GM-CSF (PeproTech, catalog 315-03). Cells were seeded at 1 × 106 cells/well in 24-well plates. Media were replenished with fresh GM-CSF–containing media on day 3. BMDCs were matured with 100 ng/mL LPS on days 4 and 5, collected on the next day, and analyzed by flow cytometry to confirm expression of CD11c, MHCII, and CD86 expression.

MHCII epitope prediction. We previously screened and generated SLPs for 245 neoantigens identified in the Panc02 line that were predicted to bind 2 MHCI alleles expressed in Panc02 cells, namely H2-Kb or H2-Db, with a binding affinity of < 1,000 nM (14). To build on this, we used IEDB to rank the binding affinity of these 245 neoantigens to I-Ab, the MHCII allele expressed by Panc02. Neoepitopes that displayed a binding affinity greater than the 50th percentile for I-Ab were tested for in vivo immunogenicity.

Peptide vaccines. Peptides were synthesized by Peptide 2.0 at 95% purity. Lyophilized peptides were stored at –20°C with CaSO4 desiccant (Drierite). Peptides were dissolved in DMSO at 50 mg/mL, aliquoted, and stored at –80°C. Peptide vaccines were prepared by diluting AddaVax at a 1:1 vol/vol ratio with peptides and 2’3’-c-di-AMP(PS)2 (Rp,Rp), an analog of c-di-AMP STING agonist (5 μg/dose, InvivoGen, catalog vac-nacda2r) in PBS, vortexed 3 times, and incubated for 10 minutes. Peptides were added to the Addavax/STING/PBS mix at a final concentration of 50 μg/peptide and vortexed 3 times. Peptide vaccine composition was as follows: CD4 vaccine (peptides 43, 53, 133, 197, 230, and 239), CD8 vaccine (peptides 20, 23, 44, 66, 77, 84, 94, 175, 219, and 237), and PancVAX2 (peptides 20, 23, 43, 44, 53, 66, 77, 84, 94, 133, 175, 197, 219, 230, 237, and 239). Sequences shown in Supplemental Tables 1 and 2.

In vivo immunogenicity studies. Six- to 7-week-old non–tumor-bearing C57BL/6 mice were vaccinated with CD4 vaccine, CD8 vaccine, or PancVAX2 or PBS twice, 7 days apart. Mice were injected with 100 μL of vaccine s.c. at the base of the tail. Seven days following the second vaccination, spleens were harvested for downstream analysis of T cells.

Lymphocyte isolation from spleens. Freshly harvested spleens were processed into a single-cell suspension by passing the spleen through a 40 μm filter (CELLTREAT, catalog 229481) utilizing a pestle (CELLTREAT, catalog 229480). Single-cell suspensions were centrifuged at 428 g for 5 minutes, supernatant aspirated, and then resuspended in 1 mL ACK lysis buffer for 2 minutes at room temperature. In total, 10 mL splenocyte medium was added to quench ACK lysis buffer. CD8+ or CD4+ T cells were negatively isolated using EasySep Mouse CD8+ or CD4+ T cell Isolation Kits (STEMCELL Technologies; CD8 Kit, catalog 19853; CD4 Kit catalog 19852) per manufacturer’s protocol. Purified CD4+ and CD8+ T cell populations were verified by flow cytometry for CD4 and CD8 expression.

In vivo tumor studies. Five- to 6-week-old C57BL/6 mice were implanted s.c. with 3 × 106 Panc02 cells with a 30 gauge insulin needle. For tumor immune cell analysis, tumors were allowed to establish for 14–28 days. Mice were then vaccinated with CD4 vaccine, CD8 vaccine, or PancVAX2 or PBS s.c. twice, 7 days apart. Seven days after the second vaccine dose, mice were euthanized and tumors were isolated for downstream analysis. For survival studies, tumor-bearing mice were vaccinated s.c. on days 3 and 10. Tumor growth was measured by calipers, and mice bearing tumors that reached 10 mm by any measurement were euthanized.

For PancVax with combination ICI tumor growth and survival studies, 6- to 7-week-old C57BL/6 mice were implanted s.c. with 3 × 106 Panc02 cells with a 30 gauge insulin needle. Mice were given 1 of the following treatment regimens: isotype antibodies, anti–CTLA-4 (InVivoMAb BioXCell, catalog BE0131) + anti–PD-1 (InVivoMAb BioXCell, catalog BE0146) antibodies, PancVAX + isotype antibodies, or PancVAX + anti–CTLA-4 + anti–PD-1. Tumor-bearing mice were vaccinated s.c. on day 12 and 19 with PancVAX or PBS. PancVAX includes peptides 20, 23, 44, 66, 77, 84, 94, 175, 218, 219, 230, and 237. Vaccines for the tumor growth and survival with PancVAX and ICI were prepared as detailed above. Vaccines for the single-cell RNA-seq experiment were prepared as detailed above, with ADU V16 as the sting agonist at a dose of 5 μg/mouse. Mice were dosed with InVivoMAb anti-mouse (100 μg, BioXCell, catalog BE0146) PD-1 or InVivoMAb rat IgG2a isotype control (100 μg, BioXCell, catalog BE0090) starting at day 12. Mice were dosed with InVivoMAb anti–mouse CTLA-4 antibodies (100 μg, BioXCell, catalog BE0131) or isotype InVivoMAb polyclonal Syrian Hamster IgG (100 μg, BioXCell, catalog BE0087) were dosed on day 15 and day 22. Tumor growth was measured by calipers, and mice bearing tumors that had ulcerated or reached 10 mm by any measurement were euthanized.

TIL isolation. Tumors were processed to a single-cell suspension using a mouse tumor dissociation kit (Miltenyi Biotec, catalog 130-096-730) and gentleMACS dissociator. Briefly, tumors were collected and weighed before being chopped into small fragments and then transferred to a Miltenyi gentleMACS C tube (Miltenyi Biotec, catalog 130-093-237). Media containing diluted tumor dissociated kit enzymes were added to the tube. Tumors were digested in the gentleMACS dissociator with the “soft tumor” setting for 40 minutes. Once dissociated, tumors were filtered first using a 100 μm filter (CELLTREAT, catalog 229485), and then using a 40 μm filter (CELLTREAT, catalog 229481). Single-cell suspensions were centrifuged at 428 g for 5 minutes; supernatant was aspirated and then resuspended in 1 mL ACK lysis buffer for 1 minute at room temperature. In total, 10 mL of the splenocyte medium was added to quench ACK lysis buffer. Cells were centrifuged at 428 g for 5 minutes; supernatant was aspirated and resuspended in splenocyte media. For whole TIL analysis, 15 or 30 mg of tumor was plated.

CD8+ and CD4+ T cells were successively isolated from processed tumors. First, mouse CD8+ T cells were positively isolated using CD8+ TIL microbeads (Miltenyi Biotec, catalog 130-116-478) per the manufacturer. Flow through was collected, and CD4+ T cells were positively isolated using CD4+ TIL microbeads (Miltenyi Biotec, catalog 130-116-475) per manufacturer. Purified CD4+ and CD8+ T cell populations were verified by flow cytometry for CD4 and CD8 expression.

Single-cell RNA-Seq. Mice were implanted with tumors and treated following the same PancVax treatment regimen previously described above (14). Mice were sacrificed at 38 days after Panc02 implantation. Tumors were dissociated as described above. Samples were sequenced in 2 batches, one composed of single-cell suspensions from dissociated tumors pooled within the same treatment groups and one immune-enriched single-cell suspensions from pooled dissociated tumor cells using Percoll (MilliporeSigma) density separation. Library preparation was conducted using Chromium Single Cell 5′ Library and Gel Bead Kit followed by NextSeq 500 sequencing (Illumina). Sequence alignment and counting was conducted with CellRanger (v 4.0.0). Data analysis was carried out in R (v. 4.2.0) using Seurat (v. 4.1.1). Genes were filtered to include only genes expressed in at least 3 cells. Cells were filtered to include only cells with between 200 and 3,000 genes detected, between 1,000 and 10,000 unique molecular identifiers (UMI), less than 10% of UMIs from mitochondrial genes, and no expression of hemoglobin genes. Cell cycle G2M and S phase scores were calculated using the Seurat module score function and mouse orthologs of the human cell cycle-associated genes from Kowalczyk et al. (43) Expression counts were normalized by log transformation with a scaling factor of 10,000 and scaled with regressing out of G2M and S phase scores. Batch correction of the low-dimensional embeddings of cells with regression of the 2 sequencing batches was carried out using Harmony (v. 0.1.0). (44) Neighborhoods of cells were detected using the Seurat FindNeighbors function based on the first 25 Harmony dimensions. Uniform manifold approximation projection (UMAP) embeddings were calculated based on the first 25 dimensions from Harmony with embedding parameters a = 0.9922, b = 1.112. Clusters of transcriptionally similar cells were detected with Louvain clustering at resolution 0.6. Clusters of cells were annotated based on expression of canonical marker genes. Louvain clustering at resolution 1.0 was used to distinguish NK cells from naive CD8+ T cells. Differential gene expression between treatment groups was assessed using the MAST test wrapper of the Seurat FindMarkers function controlling for batch. Tests considering all CD8+ T cells used cells annotated as cycling_CD8_T, cytotoxic_CD8_T, exhausted_CD8_T, or naïve_CD8_T. Results of MAST differential gene expression tests were plotted using EnhancedVolcano (v. 1.14.0). T cell phenotype proportions were calculated for each treatment group as the number of cells annotated as the T cell phenotype divided by the total number of T cells within the treatment group. Stacked bar plots were rendered using ggplot2 (v. 3.4.1).

ELISpot. Capture antibody, anti–mouse IFN-γ (clone AN18, Mabtech, catalog 3321-3-1000), was diluted (1:100 in PBS) and added (100 μL) to each well of the filter plate (MilliporeSigma, catalog MSHAS4B10). Plates were covered with parafilm and incubated overnight at 4°C. The following morning, plates were washed 3 times with PBS and blocked with complete lymphocyte media for at least 2 hours at 37°C. T2-H2-Kb, T2-H2-Db, T2-I-Ab, or BMDCs were used as APCs, which were collected, washed with PBS, counted, and resuspended in lymphocyte media at 1 × 106 cells/mL. Individual peptides were added at a final concentration of 2.5 μg/mL and shaken for 2 hours at room temperature. CD8+ or CD4+ T cells, isolated as described above, were resuspended in lymphocyte media at 1 × 106 cells/mL. In total 1 × 105 peptide-pulsed APCs and 1 × 105 T cells were added in a total volume of 200 μL to the coated and blocked ELISpot plate and incubated overnight at 37°C. The following day, plates were washed 6 times with PBST and incubated with 100 μL secondary anti–mouse IFN-γ biotin (clone R4-6A2-biotin, Mabtech, catalog 3321-6-1000) diluted 1:1,000 in PBS for 2 hours at room temperature. Thirty minutes before incubation with the secondary antibody was complete, Vectastain ABC-HRP Kit (Vector Laboratories, catalog PK-4000) was prepared by diluting reagent A (1:100) and reagent B (1:100) in PBS, followed by vortexing. After a 30-minute incubation, plates were washed 6 times with PBST and 100 μL of prepared Vectastain reagent was added to each well and incubated for 1 hour at room temperature. Plates were washed 3 times with PBST and PBS in succession. AEC substrate (Vector Laboratories, catalog SK-4200) was prepared by adding 144 μL reagent 1, 180 μL reagent 2, and 160 μL reagent 3 to 10 mL deionized (DI) water (for each plate), followed by vortexing between reagent addition. In total, 100 μL AEC substrate was added to each well and incubated for 15 minutes. Plates were washed 6 times with DI water and patted on paper towels to dry overnight at room temperature, with plastic bottoms removed. Spots were quantified in the immune-monitoring core at Johns Hopkins University.

Antibodies. The following antibodies were used for flow cytometry analysis: anti–CD45-AF700 (clone 30-F11, 1:500, BD Biosciences, catalog 560510); anti–CD11c-BV605 (clone N418, 1:250, BioLegend, catalog 117334); anti–IA/IE-BV650 (clone M5/114.15.2, 1:500, BioLegend, catalog 107641); anti–CD11b-FITC (clone M1/70, 1:250, BD Biosciences, catalog 553310); anti–F4/80-PECy7 (clone BM8, 1:500, Invitrogen, catalog 25-4801-82); anti–Ly6G-BV421 (clone IA8, 1:250, BD Biosciences, catalog 562737); anti–Ly6C-PECy5.5 (clone HK1.4, 1:250, BioLegend, catalog 128012); anti-CD86 (clone GL-1, 1/500, BioLegend, catalog 105012); anti–CD206-PE/Dazzle (clone C068C2, 1:100, BioLegend, catalog 141732); anti–PD-L1–PE (clone M1H5, 1:100, BD Biosciences, catalog 558091); anti–CD3-BV785 (clone 17A2, 1:250, BioLegend, catalog 100232); anti–CD19-BV785 (clone 6D5, 1:250, BioLegend, catalog 115543); anti–CD3-PE/Dazzle (clone 17A2, 1:500, BioLegend, catalog 100246); anti–CD8–PacBlue (clone 53-6.7, 1:500, BD Biosciences, catalog 558106); anti–CD4-BV605 (clone RM4-5, 1:500, BioLegend, catalog 100547); anti–NK1.1-BV650 (clone PK136, 1:100, BioLegend, catalog 108736); anti–CD44-APC (clone IM7, 1:250, BioLegend, catalog 103012); anti–PD-1–FITC (clone J43, 1:100, Invitrogen, catalog 11-9985-85); anti–CD137-PE (clone 17B5, 1:250, BioLegend, catalog 106105); anti–GzmB-PECy7 (clone NGZB, 1:100, Invitrogen, catalog 25-8898-82); anti–IFN-γ–PerCP/Cy5.5 (clone XMG1.2, 1:100, Invitrogen, catalog 45-7311-82); anti–IL-2–APC (clone JES6-5H4, 1:100, Invitrogen, catalog 17-7021-81); anti–TNF-α–AF700 (clone MP6-XT22, 1:100, BioLegend, catalog 506338); anti–CD4-PacBlue (clone RM4-5, 1:250, BioLegend, catalog 100531); anti–CD137-APC (clone 17B5, 1:250, BioLegend, catalog 106109); anti–IL-2–BV605 (clone JES6-5H4, 1:100, BioLegend, catalog 503829); anti–IL-10–AF488 (clone JES5-16E3, 1:100, BioLegend, catalog 505006), anti–IL-4–PE (clone 11B11, 1:100, BD Biosciences, catalog 554435), anti–GzmB-PE eFluor610 (clone NGZB, 1:100, Invitrogen, catalog 61-8898-82), anti–IL-2–BV605 (clone JES6-5H4, 1:100, BioLegend, catalog 503829), anti–IL-17A-BV650 (clone TC11-18H10.1, 1:100, BioLegend, catalog 506930), anti–FOXP3-AF488 (clone FJK-16s, Invitrogen, catalog 53-5773-82), anti–T-bet–PE/Cy7 (clone 4B10, BioLegend, catalog 644824), and anti–Gata-3–eFluor450 (clone TWAJ, Invitrogen, catalog 48-9966-41).

Flow cytometry. Single-cell suspensions were plated in 96-well U-bottom plates. For samples with cytokine detection, cells were treated with 1× Protein Transport Inhibitor Cocktail (eBioscience, catalog 00-4980-03) for at least 5 hours prior to collection and staining. Cells were washed twice with PBS and stained for viability in 100 μL of Zombie NIR fixable live/dead stain (1:1500, BioLegend, catalog 423106) and Mouse BD Fc Block (1:50, BD Biosciences, catalog 553141) in PBS for 15 minutes at room temperature. Cells were washed twice with FACS buffer (1× HBSS, 2% FBS, 0.1% Sodium Azide) and then stained with 100 μL extracellular antibody mix diluted in FACS buffer for 20 minutes at 4°C. The cells were then washed twice with FACS buffer. For intracellular cytokine detection, BD Cytofix/Cytoperm Fixation/Permeabilization Kit was used (BD Biosciences, catalog 554714). For transcription factor detection, eBioscience FOXP3/Transcription Factor Staining Buffer Set was used (Invitrogen, catalog 00-5523-00). Briefly, cells were permeabilized in 100 μL fix/perm for 20 minutes at 4°C, washed twice in 1× perm/wash buffer, and then stained with 100 μL intracellular antibody mix diluted in 1× perm/wash buffer for 30 minutes at 4°C. Cells were then washed twice with 1× perm/wash buffer and were immediately run on a BD Cytoflex. All flow cytometry data were analyzed using FlowJo version 10.8.1.

Statistics. Data were graphed and analyzed by GraphPad Prism (version 9.5.1). Multiple comparisons were analyzed by 1-way or 2-way ANOVA, followed by Hidak-Holms or Dunnett’s test when significance was found. Survival studies were analyzed by a log-rank test. Data are represented as mean ± SD. Tumor growth plots are represented as mean ± SEM. P ≤ 0.05 is considered significant.

Study approval. All animal studies were approved by the IACUC at Johns Hopkins University.

Data availability. All investigators are committed to timely distribution of data obtained in this research work. The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus (45) and are accessible through GEO Series accession no. GSE244992. All programming code used in analysis and computational tool development will be shared as open-source software via GitHub and Bioconductor at time of publication at https://github.com/FertigLab/PacVAX2_pancvaxICI_scRNAseq, commit ID 69053b3. Values for all data points in graphs are reported in the Supporting Data Values file.