A DNA-Launched Nanoparticle Vaccine Elicits CD8⁺ T-cell Immunity to Promote In Vivo Tumor Control

Cytolytic T cells (CTL) are an extremely important component of the adaptive immune system, which can selectively kill target cells through release of cytokines, granzyme, and perforin, as well as mediate target cell apoptosis through Fas and Fas-ligand interactions (1). As such, induction of broad endogenous antigen-specific CTLs through vaccination may be an important approach to treat cancer (2). However, induction of CTLs through vaccination can be challenging, as it requires antigen-presenting cells (APC) to present HLA-restricted epitopes to MHC I and provide costimulatory signals (3). Vaccinating with viral vectors encoding target antigens is an approach to drive CTL responses, potentially through direct transduction of target cells (4), and special adjuvants can be used to facilitate cross-presentation of target antigens, which to varying degrees promote phago-lysosomal escape and retro-translocation of target antigens into the cytosol (5). Alternatively, dendritic cells (DC) can be loaded ex vivo with saturating concentrations of peptides and adoptively transferred back into the host to prime CD8⁺ T cells (6). However, each approach has certain drawbacks: (i) relatively few adjuvants have been demonstrated to help prime CD8⁺ T-cell responses in clinic (7), (ii) viral-vectored vaccines can be limited by preexisting immunity against the viral vector (8), and (iii) there are manufacturing issues with large-scale production of DC-based vaccines.

DNA vaccination has also been observed to elicit CD8⁺ T-cell responses both in preclinical animal models and in clinical trials (9), plausibly through a combination of direct transfection of target cells (10) and cross-presentation of DNA-encoded antigens (11). We previously reported a new strategy to use DNA to launch in vivo–produced nanoparticle vaccines (DLnano-vaccines), which result in significant improvements in the rate of seroconversion, magnitude of humoral response, dose sparing, and protection from viral challenge (12). We have also observed that DLnano-vaccines induce significantly improved CD8⁺ T-cell responses compared with DNA-encoding monomeric antigens (12). In this work, we compared induction of CTL responses by DLnano-vaccines and the corresponding protein nano-vaccines, DNA-encoded monomeric vaccines, and CpG-adjuvanted peptide vaccines in mice. Mechanistic studies were pursued to determine the role of electroporation, transient tissue apoptosis, and APC infiltration in the priming of CTL responses by DLnano-vaccines. Finally, DLnano-vaccines capable of eliciting epitope-specific CTL responses were designed and evaluated in immunogenicity studies and a prophylactic B16-F10 melanoma challenge model. This work provides a demonstration of the capability of DLnano-vaccines to drive CTLs against desired antigens and should be explored against additional cancer targets.

DNnano_LS_GT8 and DLnano_LS_HA(NC99) were developed in prior work (12). The self-assembling nanoparticles scaffolding Trp2 and Gp100 peptides were engineered using structure-guided design. In our preliminary experiments, lumazine synthase did not express well in mammalian cell lines on its own, and we hypothesized that it could not be used on its own to scaffold antitumor peptide antigens. We engineered a new version of DLnano_LS_GT8 capable of displaying peptide antigens. In this construction, the heavily glycosylated GT8 domain facilitated solubilization and secretion of designed nanoparticles and could potentially be replaced by other heavily glycosylated domains. To ensure epitope accessibility and homogeneous nanoparticle assembly, N-linked glycans in proximity to the C-terminus of GT8 were removed by mutations in the PNGS sequence (N112D, N165D, and N170D), and a nine amino acid linker (GGSGSGGGS) was incorporated to the C-terminus of GT8 upstream of the scaffolded antitumor peptides (Gp100₂₅: EGPRNQDWL and Trp2₁₈₈: SVYDFFVWL).

Protein sequences for IgE leader sequence and eOD-GT8-60mer were as reported previously (Supplementary Table S1; refs. 13, 14). Protein sequence for HA1_NC99 was obtained from GenBank (accession no. AY289929.1; Supplementary Table S1). DNA-encoding protein sequences were codon and RNA optimized as described previously (13). The optimized transgenes were synthesized de novo (GenScript) and cloned into a modified pVAX-1 Backbone (Invitrogen, catalog no. V26020) under the control of the human cytomegalovirus promoter and bovine growth hormone polyadenylation signal.

A total of 2 × 10⁹ Expi293F Cells (Invitrogen) in 1 L Expi Expression Medium (Invitrogen) were transfected with 300 μg pVAX-1 plasmid vector encoding the eOD-GT8-60mer, HA(NC99)-60mer, LS_ Trp2188-60mer, or LS_Gp10025-60mer with PEI (Sigma-Aldrich)/OPTI-MEM (Invitrogen) and harvested 6 days posttransfection. Transfection supernatant was first purified with affinity chromatography using the AKTA Pure 25 System and gravity flow columns filled with GNL Lectin Beads (GE Healthcare). The eluate fractions from the affinity purification were pooled, concentrated with Amicon Ultra-15 Centrifugal Filter Unit with 30 kDa cutoff (Millipore), and dialyzed into 1 × PBS buffer before being loaded onto the Superose 6 Increase 10/300 GL Size Exclusion Chromatography (SEC) Column (GE Healthcare) for purification. Identified eluate fractions were then collected and concentrated to 1 mg/mL in PBS as described previously (12).

The study does not involve any human subjects requiring institutional review board approval. All animal experiments were carried out in accordance with animal protocols 201214, 201115, and 201221 approved by the Wistar Institute Institutional Animal Care and Use Committee.

For DNA-based immunization, 6- to 8-week-old female C57BL/6, BALB/c, or B6.129S(c)-Batf3tm1Kmm/J (BATF3-KO) mice (The Jackson Laboratory) were immunized with DNA vaccines via intramuscular injections into the tibialis anterior muscles, coupled with or without intramuscular electroporation with the CELLECTRA 3P Device (Inovio Pharmaceuticals), as described previously (15). For DNA immunizations using DLnano_LS_GT8 and DLnano_LS_HA (NC99) as antigens (except in dose comparison study; Supplementary Fig. S1E–S1H), 25 μg of plasmid DNA was used, a standard DNA dose that has been utilized in prior studies (12, 16). For DNA immunizations involving DLnano_LS_Trp2₁₈₈, DLnano_LS_Gp100₂₅, DLmono_Trp2₁₈₈, and DLmono_Gp100₂₅, 10 μg of individual plasmid DNA was used either alone or in combination. For vaccinations involving recombinant protein, a high protein dose was used in this study compared with prior studies (17). Six- to 8-week-old female BALB/c were immunized subcutaneously with 10 μg of recombinant eOD-GT8-60mer protein coformulated with either 50 μL Sigma Adjuvant System (RIBI; Sigma-Aldrich), 50 μg poly(I:C) (InvivoGen), or 20 μg CpG ODN 1826 VacciGrade (InvivoGen), with or without electroporation. Alternatively, BALB/c or C57BL/6 mice were immunized with 10 μg protein HA(NC99)-60mer and 4 μg of LS_Trp2₁₈₈-60mer and LS_Gp100₂₅-60mer in 50 μL PBS coformulated with 50 μL Sigma Adjuvant System (Sigma-Aldrich). For the high dose protein versus DNA comparison study, reported in Supplementary Fig. S1E–S1H, 50 μg DLnano_LS_GT8 was administered with electroporation, or 50 μg protein eOD-GT8-60mer coformulated with RIBI was administered without electroporation. For peptide vaccinations, C57BL/6 mice received intramuscular vaccinations of 10 μg Trp2₁₈₈ or 10 μg Gp100₂₅ peptides (GenScript) coformulated with 20 μg CpG ODN 1826 VacciGrade in sterile saline. For all immunizations, reported in Fig. 1 and Supplementary Fig. S1, the mice were immunized twice, 3 weeks apart and euthanized 2 weeks after the second vaccination for spleen collection. The sera were collected on days 0, 7, 14, 21, and 35 days after the first vaccination by bleeding through the submandibular vein. For immunizations reported in Fig. 2H–I and Supplementary Fig. S2K, the mice were immunized once and euthanized 2 weeks after the vaccination for spleen collection. For all immunogenicity studies in Fig. 3C and D and Supplementary Fig. S3D and S3F–S3I, the mice were immunized twice 2 weeks apart and euthanized 2 weeks after the second vaccination for spleen collection.

C57BL/6 mice received intramuscular injections of 80 μg of DLnano_LS_GT8 coformulated with 12 U hyaluronidase (Sigma-Aldrich) with or without intramuscular electroporation or 20 μg of protein eOD-GT8-60mer coformulated 1:1 with Sigma Adjuvant System with or without intramuscular electroporation (Fig. 2; Supplementary Fig. S2). For Supplementary Fig. S2C and S2D, mice sera were collected by cheek bleeds 0, 1, 4, and 10 days postinjection (d.p.i.) for measurements of muscle enzymes in the sera. For in vivo macrophage depletion, 6- to 8-week-old female C57BL/6 mice received one or three intravenous injections (on days −3, 0, and 3 with respect to DNA vaccination) of 750 μg clodronate disodium formulated in clodrosome (150 μL injection, Encapsula Nanoscience) via retro-orbital injections and control mice each received 150 μL of encapsosome (Encapsula Nanoscience) intravenously at the same time points. For the one injection study (Supplementary Fig. S2H–S2J), spleens were collected 1 day after clodrosome/encapsosome treatments, and for the three injections study, muscles were collected 4 days after DNA vaccination for Fig. 2G and spleens were collected from a different set of mice 14 days after DNA vaccination for Fig. 2H. For Fig. 2I and Supplementary Fig. S2K, the BATF3-KO mice or the control wild-type C57BL/6 mice received single 25 μg DLnano_LS_GT8 vaccination and were euthanized for sera and spleen collection 14 days after DNA vaccination.

B16-F10 cells (the ATCC, catalog no. CRL-6475; authenticated by the ATCC) were maintained in 10% FBS (Lampire Biologicals)/DMEM (Corning) under low passage (less than 10) and B16-F10-Luc2 cells (the ATCC, catalog no. CRL-6475-LUC2; authenticated by the ATCC) were maintained in 10% FBS/DMEM enriched with 10 μg/mL Blasticidin (Gibco) with routine testing for Mycoplasma contaminations and other mouse pathogens. Cells were trypsinized and strained through 70-μm strainer to generate single-cell suspensions. They were then administered subcutaneously to mice (10⁵ cells to each mouse in 100 μL PBS). The tumor size was measured every 2 days with a digital caliper, and tumor volume was determined with the formula V = 0.5W²L (V, tumor volume; W, tumor width; and L, tumor length). Mice with tumor volumes greater than 2,000 mm³ or with any dimension exceeding 2 cm were euthanized for humane purposes or until 80 days after tumor inoculation (experimental endpoint). For recombinant anti–PD-1 (RMP1–14, BioXCell, catalog no. BE0146) administration in the melanoma therapeutic treatment model, 200 μg of antibody was injected intraperitoneally in 100 μL PBS to each mouse weekly, starting 3 days after tumor inoculation. For mechanistic study, to determine the role of CD8⁺ T cells in mediating in vivo control of tumor growth (Fig. 3M; Supplementary Fig. S3I), prophylactically vaccinated mice received 200 μg anti-mouse CD8α (2.43, BioXCell, catalog no. BE0061) or 200 μg Rat IgG2b isotype control (LTF-2, BioXCell, catalog no. BE0090) in 100 μL sterile PBS 1 day prior to tumor inoculation.

Ninety-six–well half area plates (Corning) were coated at room temperature for 8 hours with 1 μg/mL MonoRab anti-His antibody (GenScript, catalog no. A01857), followed by overnight blocking with blocking buffer containing 1 × PBS, 5% Skim Milk (Sigma), 10% Goat Serum (Millipore), 1% BSA (Sigma), 1% FBS (Lampire Biologicals), and 0.2% Tween-20 (Sigma). The plates were then incubated with 2 μg/mL of his-tagged GT8-monomer at room temperature for 2 hours, followed by addition of mice sera serially diluted with PBS containing 1% FBS and 0.1% Tween and incubated at 37°C for 2 hours (diluted 100- to 7,812,500-fold). BALB/c mice sera from 0, 7, 14, 21, and 35 d.p.i. were analyzed (Fig. 1), and BATF3-KO mice sera from 0 and 14 d.p.i. were analyzed (Supplementary Fig. S2K). The plates were incubated at room temperature for 1 hour with anti-mouse IgG H+L HRP (Bethyl) at 1:20,000 dilution, followed by addition of TMB substrates (Thermo Fisher Scientific), and then quenched with 1 mol/L H₂SO₄. Absorbance at 450 and 570 nm was recorded with Synergy 2 Plate Reader (BioTek). Endpoint titer was defined as the highest dilution at which the optical density (OD) of the post-immune sera exceeded the cutoff (mean OD of naïve animals plus SDs of the OD in the naïve sera multiplied with SD multiplier f at the 99% confidence level).

Ninety-six–well half area plates were coated at 4°C overnight with 2 μg/mL of recombinant HA(ΔTM)(H1N1/A/New Caledonia/20/1999; Immune Technology) and blocked at room temperature for 2 hour with the buffer as described above. The plates were subsequently incubated with serially diluted mouse sera at 37°C for 2 hours (diluted 100- to 7,812,500-fold). BALB/c mice sera from 0, 7, 14, 21, and 35 d.p.i. were analyzed (Fig. 1). This was followed by 1-hour incubation with anti-mouse IgG H+L HRP (Bethyl, catalog no. A90–116P) at 1:20,000 dilution at room temperature and developed with TMB substrate. Absorbance at 450 and 570 nm was recorded with BioTek Plate Reader and endpoint titers were calculated as above.

Corning half area 96-well plates were coated with 2 μg/mL of GT8-monomer, eOD-GT8-60mer, Trp2-60mer, and Gp100-60mer at 4°C overnight. The plates were then blocked with the buffer as described in the “GT8-binding ELISA” section for 2 hours at room temperature, followed by incubation with serially diluted broadly neutralizing human immunodeficiency virus (HIV) antibody, VRC01 (NIH AIDS reagent), from 1,000 ng/mL to 0.5 ng/mL at room temperature for 2 hours. The plates were then incubated with anti-human Fc (cross-adsorbed against rabbits and mice; Jackson ImmunoResearch) at 1:10,000 dilution for 1 hour, followed by addition of TMB substrate for detection. Absorbance at 450 and 570 nm was recorded with BioTek Plate Reader.

BALB/c mice sera at 35 d.p.i. were treated with receptor-destroying enzyme (RDE, 1:3 ratio; Seiken) at 37°C overnight for 18 to 20 hours, followed by complement and enzyme inactivation at 56°C for 45 minutes. RDE-treated sera were subsequently cross-adsorbed with 10% Rooster Red Blood Cells (Lampire Biologicals) in PBS at 4°C for 1 hour. The cross-adsorbed sera were then serially diluted (16- to 2,048-fold) with PBS in a 96-well V-Bottom Microtiter Plate (Corning). Four hemagglutinating doses of A/New Caledonia/20/99 (BEI) were added to each well, and the serum–virus mixture was incubated at room temperature for 1 hour. The mixture was then incubated with 50 μL 0.5% v/v Rooster Red Blood Cells (Lampire Biologicals) in 0.9% saline for 30 minutes at room temperature. The hemagglutination inhibition (HAI) antibody titer was scored with the dot method, and the reciprocal of the highest dilution that did not cause agglutination of the rooster red blood cells was recorded.

Spleens were collected from immunized BALB/c mice at 35 d.p.i. or C57BL/6 mice at 28 d.p.i. and homogenized into single-cell suspensions with a tissue stomacher in 10% FBS/1% Penicillin–Streptomycin (Sigma) in RPMI1640. Red blood cells were subsequently lysed with ACK Lysing Buffer (Thermo Fisher Scientific), and the percentage of viable cells was determined with trypan blue exclusion using Vi-CELL XR (Beckman Coulter). A total of 2 × 10⁵ cells were then plated in each well of mouse IFNγ ELISpot Plates (MabTech), followed by addition of peptide pools that span both the lumazine synthase, GT8, or HA domains, or individual Trp2 (SVYDFFVWL) and Gp100₂₅ peptide (EGPRNQDWL) at 5 μg/mL of final concentration for each peptide (GenScript). The cells were then stimulated at 37°C for 16 to 18 hours, followed by development according to the manufacturer's instructions. Spots for each well were then imaged and counted with ImmunoSpot Macro Analyzer.

Single-cell suspension from spleens of immunized animals (BALB/c mice 35 d.p.i. or C57BL/6 mice 28 d.p.i.) was prepared as described before and stimulated with 5 μg/mL of peptide pools that span both the lumazine synthase, GT8, or HA domains, or individual Trp2 (SVYDFFVWL) and Gp100₂₅ peptide (EGPRNQDWL) (GenScript) for 5 hours at 37°C in the presence of 1:500 Protein Transport Inhibitor (Thermo Fisher Scientific) and anti-mouse CD107a-FITC (Thermo Fisher Scientific). The cells were then incubated with live/dead Fixable Violet Dead Cell Stain Kit (for 405 nm excitation) for 10 minutes at room temperature, and surface stained (anti-mouse CD4-BV510, BioLegend, catalog no. 100559 and anti-mouse CD8–APC-Cy7, BioLegend, catalog no. 100714) at room temperature for 30 minutes. The cells were then fixed and permeabilized according to the manufacturer's instructions for BD Cytoperm Cytofix Kit and stained with anti-mouse IL2–PE-Cy7 (BioLegend, catalog no. 503832), anti-mouse IFNγ-APC (BioLegend, catalog no. 505810), anti-mouse CD3e–PE-Cy5 (BioLegend, catalog no. 100310), and anti-mouse TNFα-BV605 (BioLegend, catalog no. 506329) at 4°C for 1 hour. The cells were subsequently analyzed with LSR II 18-color flow cytometer. The data were analyzed with FlowJo V10.6.1. Intracellular cytokine staining (ICS) gate for CD8⁺IFNγ⁺ population was set as in Fig. 1C and Supplementary Fig. S4.

To detect presence of cleaved caspase-3 using immunofluorescence, 4 days after C57BL/6 mice were intramuscularly immunized with DLnano_LS_GT8 or protein eOD-GT8-60mer, tibialis anterior muscles of the mice were harvested and preserved in 4% paraformaldehyde (PFA)/PBS for 4 hours at room temperature and then stored overnight in 70% EtOH/H₂O at 4°C. The tissues were then serially dehydrated for paraffin embedding, blocked in 3% BSA/PBS for 1 hour at room temperature, and permeabilized with 0.5% Triton X-100 (Thermo Fisher Scientific)/PBS, followed by overnight staining with rabbit anti-cleaved caspase 3 (1:200, Cell Signaling Technology, catalog no. 9664S). The sections were then washed and stained with anti-Rabbit Alexa Fluor 594 Antibody (1:200, Invitrogen, catalog no. A11037) for 1 hour at room temperature, then counterstained with 0.5 μg/mL DAPI, and imaged with Leica SP5 Confocal Microscopes (with the following settings: 3% laser power for 405 nm laser and 1.25% laser power for 488 nm laser; line accumulation, 3; frame average, 3; and instrument gain, 100%).

Mouse serum creatine kinase and lactose dehydrogenase were diluted 10-fold in PBS and the enzyme activities were measured 0, 1, 4, 7, and 10 days following intramuscular administrations of 25 μg DLnano_LS_GT8 or 10 μg RIBI-adjuvanted protein eOD-GT8-60mer, or from untreated control mice, using colorimetric assays according to the manufacturer's protocol (Abcam, Ab 155901 and Ab 65393). Changes in absorbance at 450 nm were measured at 37°C under kinetics mode over time using a BioTek Plate Reader.

A terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay was performed on fixed muscle specimen according to the manufacturer's protocol (Abcam, catalog no. ab206386). Briefly, 4 days after treatment with DNA or protein vaccine, tibialis anterior muscles were harvested and fixed in 4% PFA/PBS for 3 hours at 4°C. The tissues were then dehydrated and paraffin embedded, and the sample sections were then rehydrated by serial transfer of the tissues from 100% ethanol to 70% ethanol/water. The specimens were then permeabilized by treatment with 50 μL Proteinase K at 20 μg/mL (Abcam) at room temperature for 20 minutes. Endogenous peroxidase was then quenched by treatment of tissue sections with 3% H₂O₂ in methanol. The sections were then incubated with 40 μL TdT Enzyme (diluted 1:40 from the TdT Enzyme Stock, Abcam) in equilibration buffer at room temperature for 1.5 hours in a humidified chamber, and the reactions were stopped with the stop solution provided in the kit. The sections were then blocked with the provided blocking buffer at room temperature for 10 minutes and labeled with provided detection conjugate diluted 1:25 in blocking buffer at room temperature for 30 minutes. The sections were then developed with DAB substrate at room temperature for 15 minutes and counterstained with methyl green before they were imaged with a Nikon Eclipse 80i Microscope under 20× and 60× magnification (with autoexposure settings).

Following intramuscular protein or DNA vaccinations in C57BL/6 mice at 4 or 7 d.p.i. (as described above), transfected muscles were harvested. Cells were extracted from muscles using a 30-minute digestion in Hanks' Balanced Salt Solution (Gibco) supplemented with 0.5% Type IV Collagenase (Life Technologies), 0.2% BSA, and 0.025% Trypsin (Corning) at 37°C. The cells were then stained with ultra-violet reactive (live/dead dye, 1:333 in PBS), and then with a 1:200 dilution of anti-CD11b–PerCPCy5.5 (BioLegend, catalog no. 101228), anti-CD11c–PE (BioLegend, catalog no. 301606), anti-MHC Class II–APC-Cy7 (BioLegend, catalog no. 107628), anti-F4/80–PE-Cy7 (BioLegend, catalog no. 123114), and anti-CD206–APC (BioLegend, catalog no. 141708) in 2% FBS/PBS at room temperature for 1 hour before they were permeabilized with BD CytoPerm/CytoFix Buffer for 20 minutes at 4°C and subsequently labeled with FITC-lightning link labeled VRC01 (Expedon) at 4°C for 1 hour. The cells were then resuspended in 1 × BD Permeabilization Wash Buffer for analyses on LSR II 18-color flow cytometer. The data were analyzed with FlowJo V10.6.1. ICS gate for macrophages was set as in Fig. 2C and Supplementary Fig. S2E.

The nanoparticles produced in Expi293 cells were purified using agarose-bound lectin beads (Agarose Galanthus Nivalis Lectin, Vector Laboratories) followed by SEC (GE Healthcare) using the Superose 6 Increase 10/300 GL column. The proteins were further dialyzed into TBS. A total of 3 μL of purified proteins were adsorbed onto glow-discharged carbon-coated Cu400 Electron Microscopy (EM) Grids (Electron Microscopy Sciences). The grids were then stained with 3 μL of 2% uranyl acetate, blotted, and stained again with 3 μL of the stain followed by a final blot. Image collection and data processing were performed on a FEI Tecnai T12 microscope equipped with a Oneview Gatan camera at 90,450× magnification at the camera and a pixel size of 1.66 Å.

For B16-F10-Luc2 challenge study, tumor-challenged mice received 150 mg/kg administration of VivoGlo Luciferin (Promega) at each time point (0, 5, 10, and 14 days after tumor inoculation) formulated in sterile PBS, and then imaged with an IVIS spectrum CT for bioluminescence with the auto-exposure settings (or for 60 seconds, whichever was shorter) 10 minutes after injection.

Mice in the experiments were randomly assorted into cages by Animal Facility staff and not further randomized. Data acquisition and analysis were not blinded. Power analysis was performed with R on the basis of our preliminary data to determine the smallest sample size that would allow us to achieve a power of 0.9 with a preset α value of 0.05. All statistical analyses were performed with PRISM V8.2.1 and R V3.5.1. Each individual data point was sampled independently. Two-tailed Mann–Whitney rank tests were used to compare differences between groups. Log-rank test was used to compare survival between two groups in challenge survival studies. Bonferroni corrections were used to adjust for multiple comparisons.

First, we compared adaptive immune responses induced by DLnano-vaccines and protein nano-vaccines. We utilized two model antigens: a priming HIV antigen, eOD-GT8-60mer, and an influenza hemagglutinin (H1 A/NewCaledonia/20/1999)60mer antigen scaffolded by lumazine synthase (LS; refs. 18, 19). DNA-launched nanoparticle lumazine synthase decorated with an anti–HIV-1 immunogen, eOD-GT8 (DLnano_LS_GT8; Supplementary Fig. S1A) and RIBI-adjuvanted protein, eOD-GT8-60mer, induced similar antibody titers in BALB/c mice after two immunizations (Fig. 1A). Epitope mapping conducted in prior studies identified that CD4⁺ T-cell responses elicited by LS-scaffolded DLnano-vaccines were predominantly directed to the LS domain (20). In this study, CD4⁺ T-cell responses to the LS domain after two vaccinations, measured by ICS, were similar between DLnano_LS_GT8 and protein eOD-GT8-60mer (Fig. 1B). We also observed robust induction of CD8⁺ T-cell responses to the GT8 domain in mice immunized with DLnano_LS_GT8, but not in those vaccinated with protein eOD-GT8-60mer (Fig. 1C–F).

We further determined whether this observation was limited by our choice of RIBI as the protein adjuvant and performed a head-to-head comparison of 25 μg DLnano_LS_GT8 versus 10 μg protein eOD-GT8-60mer adjuvanted with either RIBI, 50 μg Poly(I:C), or 20 μg CpG ODN (21, 22). We observed, only DLnano_LS_GT8, but not protein eOD-GT8-60mer administered with any form of adjuvant, was capable of inducing CD8⁺ T-cell responses (Supplementary Fig. S1B and S1C), even though both DLnano_LS_GT8 and protein eOD-GT8-60mer induced CD4⁺ T-cell responses to varying degrees (Supplementary Fig. S1D). We next determined whether this observation might be dose limited and compared DLnano_LS_GT8 and RIBI-adjuvanted protein eOD-GT8-60mer vaccinations both at 50 μg doses. Only DLnano_LS_GT8, but not protein eOD-GT8-60mer, was observed to induce CTL responses (Supplementary Fig. S1E and S1F), despite induction of CD4⁺ T-cell and humoral responses by both routes of vaccination (Supplementary Fig. S1G and S1H).

We also determined the role of electroporation in mediating CD8⁺ T-cell responses by vaccinating mice with DLnano_LS_GT8 or RIBI-adjuvanted protein eOD-GT8-60mer with or without electroporation. Although electroporation increased humoral responses for DLnano_LS_GT8 and protein eOD-GT8-60mer (Supplementary Fig. S1I and S1J), electroporation could only adjuvant CD8⁺ T-cell responses in the DLnano_LS_GT8 groups, and no responses were observed in the protein eOD-GT8-60mer groups with or without electroporation (Supplementary Fig. S1K). Antigens supplied by protein versus DNA-launched eOD-GT8-60mer existed in different forms. Protein was given exogenously as a depot and existed as soluble factors (23). DNA-launched nano-vaccines, however, were expressed by hosts' own myocytes and were cell associated (12). This finding suggested that induction of CD8⁺ T-cell immunity, in this context, required cell-associated antigens expressed by host cells from DNA cassettes.

We determined whether this observation may apply to an alternative nanoparticle-scaffolded influenza antigen. Indeed, two vaccinations of DLnano_LS_HA(NC99) and RIBI-adjuvanted protein HA(NC99)_60mer induced similar binding antibody titers to recombinant NC99 hemagglutinin (Fig. 1G), as well as hemagglutination inhibition titers against the autologous A/NewCaledonia/20/1999 virus (Fig. 1H). CD4⁺ T-cell responses were similarly induced by DLnano_LS_HA(NC99) and protein HA(NC99)_60mer (Supplementary Fig. S1L). However, only DLnano_LS_HA(NC99), but not protein HA(NC99)_60mer, induced CD8⁺ T-cell responses to HA (Fig. 1I; Supplementary Fig. S1M and S1J).

We hypothesized that the distinct modality of antigen uptake and presentation for DLnano-vaccines compared with protein nano-vaccines may attribute to unique induction of CTL responses by DLnano-vaccines. Previous studies observe that coadministration of DNA vaccines with DNA cassettes encoding proapoptotic genes lead to significantly improved CTL responses (24). We, therefore, decided to examine tissue apoptosis and APC infiltration following vaccination. We first assessed the extent of muscle tissue damage upon vaccination using immunofluorescence staining against cleaved caspase-3 (25). In C57BL/6 mice, 4 d.p.i., hypercellularity and expression of cleaved caspase-3 were only observed in the muscles of mice immunized with DLnano_LS_GT8 combined with electroporation, but not in those vaccinated with protein eOD-GT8-60mer coformulated in RIBI without electroporation or in naïve mice (Fig. 2A). Alternatively, using a TUNEL assay (26) 4 d.p.i., brown nuclei, suggestive of double-stranded DNA breaks and cellular apoptosis, were only observed in muscle tissues of mice treated with DLnano_LS_GT8 and electroporation, but not in muscle sections from naïve mice or those treated with RIBI-formulated protein eOD-GT8-60mer without electroporation (Fig. 2B). Electroporation was determined to be instrumental to the induction of myocyte apoptosis, as cleaved caspase-3 expression was observed 4 d.p.i. when either protein eOD-GT8-60mer or DLnano_LS_GT8 was delivered with electroporation (Supplementary Fig. S2A). Without electroporation, low cleaved caspase-3 expression was observed only for DLnano_LS_GT8, but not for protein eOD-GT8-60mer.

We performed a time–course experiment and observed that DLnano_LS_GT8–induced tissue apoptosis was transient, peaking at 6 d.p.i. and was fully resolved by 14 d.p.i. (Supplementary Fig. S2B). We also observed that tissue apoptosis was limited locally to the injection sites, as indicators of tissue damage such as serum lactose dehydrogenase and creatine kinase were not elevated in mice immunized with DLnano_LS_GT8 or protein eOD-GT8-60mer relative to untreated control mice (Supplementary Fig. S2C and S2D).

We next examined the consequences of induced tissue apoptosis on tissue APC infiltration. At 7 d.p.i., we observed increased influx of CD11b⁺F4/80⁺ macrophages into muscles of mice treated with DLnano_LS_GT8 and electroporation compared with protein eOD-GT8-60mer formulated in RIBI without electroporation (Fig. 2C and D). Infiltrating macrophages were observed to be predominantly proinflammatory CD11c⁺CD206⁻ M1 macrophages rather than anti-inflammatory CD11c⁻CD206⁺ M2 macrophages (Supplementary Fig. S2E and S2F). Increased influx of CD11c⁺MHCII⁺ DCs was also observed for mice treated with DLnano_LS_GT8 compared with those treated with protein eOD-GT8-60mer. However, infiltrating DCs were significantly less abundant than macrophages (Fig. 2E). Staining of APCs with GT8-specific VRC01 demonstrated that infiltrating macrophages induced by DLnano_LS_GT8, but not those induced by protein eOD-GT8-60mer, had taken up the GT8 antigen (Fig. 2F; Supplementary Fig. S2G).

To determine the functional roles played by macrophages in priming of CD8⁺ T-cell responses by DLnano-vaccines, we systemically depleted the macrophage population in C57BL/6 mice prior to DNA vaccination with intravenous injection of clodrosome (27). A single intravenous infusion of clodrosome specifically depleted CD11b⁺F4/80⁺ macrophages, but not CD11c⁺MHCII⁺ DCs, in the spleens of animals 1 d.p.i. (Supplementary Fig. S2H–S2J). Depletion of infiltrating macrophages in the muscles after vaccination required increased doses of clodrosome on −3, 0, and 3 d.p.i. (with respect to DNA vaccination), which significantly reduced macrophage infiltration into the muscles by approximately 4-fold (Fig. 2G). This depletion scheme significantly attenuated induced CD8⁺ T-cell responses following one vaccination of DLnano_LS_GT8 14 d.p.i. in C57BL/6 mice by approximately 3-fold (Fig. 2H).

It has previously been reported that monocytic populations can transfer sequestered peptides through gap junctions to conventional (c)DCs, which can more efficiently prime CD8⁺ T cells (28). We investigated the role of CD11c⁺CD8α⁺ cDCs in cross-presentation of DLnano-immunogens using BATF3-KO transgenic mice, which lack splenic development of CD8α⁺ cDCs (29). Single immunization of DLnano_LS_GT8 induced similar humoral responses to GT8 in BATF3-KO and wild-type C57BL/6 mice (Supplementary Fig. S2K), even though induced CD8⁺ T-cell responses were significantly attenuated in BATF3-KO mice (Fig. 2I). Taken together, the data highlight the importance of both phagocytic macrophages and CD8α⁺ cDCs in priming CD8⁺ T cells.

To demonstrate the functional relevance of CTL priming by DLnano-vaccines in the treatment of cancer, we designed nano-vaccines displaying immunodominant CD8⁺ T-cell epitopes. The self-assembling nanoparticles were engineered using a structure-guided process as described in the Materials and Methods section. We designed nanoparticles presenting 60 copies of Trp2₁₈₈ peptide (DLnano_LS_Trp2₁₈₈) and Gp100₂₅ peptide (DLnano_LS_Gp100₂₅). Homogeneous in vitro assemblies of DLnano_LS_Trp2₁₈₈ and DLnano_LS_Gp100₂₅ were observed by SEC (Fig. 3A; Supplementary Fig. S3A) and negative-stain electron microscopy (nsEM; Fig. 3B; Supplementary Fig. S3B), respectively. We also demonstrated the multivalent nature of the designed DLnano_LS_Trp2₁₈₈ and DLnano_LS_Gp100₂₅ nanoparticles, which bound to VRC01 with affinity similar to that of eOD-GT8-60mer, and higher than that of GT8 monomer (Supplementary Fig. S3C).

Groups of C57BL/6 mice were immunized with designed DNA-launched nanoparticle Trp2₁₈₈ or Gp100₂₅ vaccines individually and were observed to induce significantly improved CD8⁺ T-cell responses to the Trp2₁₈₈ and Gp100₂₅ peptides compared with DNA vaccines encoding monomeric GT8-scaffolded Trp2₁₈₈ and Gp100₂₅, respectively (DLmono_Trp2₁₈₈ and DLmono_Gp100₂₅; Fig. 3C and D). When DLnano_LS_Trp2₁₈₈ and DLnano_LS_Gp100₂₅ were administered in separate sites in a single animal, the coadministration resulted in improved elicitation of CD8⁺ T-cell responses to both Trp2 and Gp100 peptides compared with DLmono_Trp2₁₈₈ and DLmono_Gp100₂₅ (Supplementary Fig. S3D). We evaluated the potency of the designed DLmono_Trp2₁₈₈, DLmono_Gp100₂₅, DLnano_LS_Trp2₁₈₈, and DLnano_LS_Gp100₂₅ in a therapeutic B16-F10 melanoma model (Supplementary Fig. S3E). We observed DLnano_LS_Trp2₁₈₈ and DLnano_LS_Gp100₂₅ treatments, in the absence of anti–PD-1, significantly extended median survival compared with pVAX treatment. Potency of DLnano_LS_Trp2₁₈₈ and DLnano_LS_Gp100₂₅ treatments could be further enhanced when anti–PD-1 was coadministered. DLnano_LS_Trp2₁₈₈ and DLnano_LS_Gp100₂₅ combined with anti–PD-1 significantly improved survival compared with DLmono_Trp2₁₈₈ and DLmono_Gp100₂₅ alone, extending the median survival by 9 days.

We next compared CTL responses induced by DLnano-vaccines to CpG-adjuvanted peptide vaccines. We observed that both DLnano_LS_Trp2₁₈₈ and CpG-adjuvanted Trp2₁₈₈ peptide vaccination induced peptide-specific CD8⁺ T-cell responses, although responses induced by DLnano_LS_Trp2₁₈₈ vaccination were significantly higher (Supplementary Fig. S3F). For the Gp100₂₅ peptide, prior studies suggest that Gp100₂₅ peptide vaccination induces weaker and less consistent CD8⁺ T-cell responses, unless combined with lymph node–targeting agents or TLR7 agonists (30, 31). Our study indicated that vaccination with only DLnano_LS_Gp100₂₅, but not CpG-adjuvanted Gp100₂₅ peptide, induced robust peptide-specific CD8⁺ T-cell responses (Supplementary Fig. S3G–S3I).

We next compared induction of CD8⁺ T-cell responses to Trp2₁₈₈ and Gp100₂₅ epitopes in B16-F10 melanoma–bearing mice, which received treatments as shown in the scheme (Fig. 3E; ref. 32). We observed CD8⁺ T-cell responses to Trp2 only in mice treated with DNA vaccinations, but not those receiving anti–PD-1 treatment alone or anti–PD-1 plus protein vaccination (Fig. 3F). Induced Trp2-specific CD8⁺ T cells exhibited effector phenotypes (IFNγ⁺CD107a⁺; Fig. 3G) and were also polyfunctional (IFNγ⁺TNFα⁺IL2⁺; Fig. 3H). Similar observations were also made for Gp100-specific CD8⁺ T cells (Supplementary Fig. S3J–S3L). In this therapeutic model, DNA, but not protein, vaccination of Trp2₁₈₈ and Gp100₂₅-60mer was observed to suppress B16-F10 tumor growth, and significantly prolong median survival of mice by 11 days (Fig. 3I–J). The effect was even more pronounced when mice were vaccinated prior to tumor inoculation. Two vaccinations of DLnano_LS_Trp2₁₈₈ and DLnano_LS_Gp100₂₅ prior to B16-F10-Luc challenge completely prevented tumor growth in 80% (4/5) mice, whereas tumor growth was observed in all pVAX control mice or mice treated with protein Trp2₁₈₈ and Gp100₂₅-60mer (Fig. 3K). Although all mice in the pVAX and protein groups died, 80% of mice in the DNA group had tumor-free survival (Fig. 3L). Finally, protection by DLnano-vaccination was observed to be CTL mediated, as indicated by CD8⁺ T-cell depletion (Fig. 3M; Supplementary Fig. S3M). Taken together, the data suggest DLnano-vaccination can elicit CD8⁺ T-cell responses to confer protection against melanoma both in prophylactic and therapeutic settings.

CD8⁺ T cells can play an extremely important role in surveillance against intracellular pathogens and tumors. In cancer, presence of tumor-infiltrating CD8⁺ T cells correlates with improved prognosis in patients (33, 34). Such observations have led to the use of T-cell–based therapies, such as DC vaccines or in vitro expansion and adoptive transfer of tumor-infiltrating lymphocytes, in patients with cancer, achieving varying degrees of success (35, 36). Alternatively, several reports also describe the antitumor activity of antigen-specific CD4⁺ T cells. T_H1 cells, for instance, are responsible for immune responses against tumors by either enhancing CD8⁺ T-cell response or activating macrophages to phagocytose cancerous cells. CD4⁺ CTLs are another subset of CD4⁺ T cells that have acquired cytolytic activity and demonstrate clear antitumor activity (37, 38). Our work here mainly focuses on designing DLnano-vaccines scaffolding CD8⁺ T-cell epitopes, but nano-vaccines scaffolding CD4⁺ T-cell epitopes should also be closely examined in future work.

Induction of CTL responses by vaccination can be challenging and not readily achieved by vaccination with protein or inactivated virus (39). DNA vaccines have been shown to elicit CD8⁺ T-cell responses both in preclinical animal models and in clinical trials (9, 16). Here, we observed that electroporation resulted in transient tissue apoptosis and macrophage infiltration, and significantly enhanced CTL responses by DLnano-vaccines. Bolus protein nanoparticle vaccinations and electroporation, however, did not result in CTL induction, possibly suggesting that direct in vivo production of nano-vaccines in host's cells would be required. The work also demonstrated the importance of both electroporation and macrophages in CTL induction by DLnano-vaccines. However, the study did not directly demonstrate that electroporation-induced muscle-infiltrating macrophages were critical because the clodrosome treatments used here systemically depleted macrophages, including, but not limited to, the electroporation-induced muscle-infiltrating macrophages. Future study to characterize the biology of these muscle-infiltrating macrophages. We also demonstrated the importance of CD8⁺ T-cell priming by DLnano-vaccines in a melanoma model. We designed a generalizable nanoparticle platform for displaying CD8⁺ T-cell epitopes to various tumor antigens. As a proof of principle, we engineered two nanoparticle vaccines scaffolding 60 copies of CD8⁺ epitopes from Trp2 and Gp100. DNA cassettes encoding antimelanoma LS-nanoparticles scaffolding Trp2 and Gp100 epitopes could elicit significantly improved CTL responses to both targets compared with conventional monomeric DNA vaccines, CpG-adjuvanted peptide vaccines, and protein nano-vaccines, and conferred 80% protection when the vaccine was prophylactically administered before tumor challenge. Finally, DLnano-vaccines could induce CTL responses to whole antigens, such as GT8, influenza hemagglutinin, or to selected peptides. Whole antigen- or neoantigen peptide–oriented DLnano-vaccines may be designed for various cancer targets and presented as a viable and attractive strategy in the exciting era of cancer immunotherapy.

Z. Xu reports a patent for US.62784318 pending. M.C. Wise reports personal fees from Inovio Pharmaceuticals during the conduct of the study. P.D. Fisher reports employment with and owns shares and stock options in Inovio Pharmaceuticals. K. Schultheis reports personal fees and other from Inovio Pharmaceuticals (salary and benefits including ownership of stock and stock options) outside the submitted work. K.E. Broderick reports salary and stock options from Inovio Pharmaceuticals during the conduct of the study and outside the submitted work, as well as a patent for DNA delivery pending. L. Humeau reports salary and stock options from Inovio Pharmaceuticals during the conduct of the study. D.W. Kulp reports a patent for nanoparticle vaccine pending. D.B. Weiner reports grants, personal fees, and other from Inovio Pharmaceuticals (stock) during the conduct of the study; personal fees from Sanofi (advisory board), BBI (consultant), GeneOne (consultant), and AstraZeneca (speaker) outside the submitted work; as well as patents for DNA vaccines licensed to Inovio Pharmaceuticals. No potential conflicts of interest were disclosed by the other authors.

Z. Xu: Conceptualization, resources, data curation, software, formal analysis, validation, investigation, methodology, writing–original draft, writing–review and editing. N. Chokkalingam: Resources, data curation, writing–review and editing. E. Tello-Ruiz: Data curation. M.C. Wise: Data curation, supervision. M.A. Bah: Data curation. S. Walker: Data curation. N.J. Tursi: Data curation. P.D. Fisher: Resources, data curation. K. Schultheis: Resources, data curation. K.E. Broderick: Resources, funding acquisition. L. Humeau: Resources, funding acquisition. D.W. Kulp: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, validation, writing–original draft, writing–review and editing. D.B. Weiner: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, writing–original draft, writing–review and editing.

The authors thank Animal Facility staff at the Wistar Institute for providing care to the animals. They thank the Imaging Facility Core at the Wistar Institute for assistance with in vivo imaging experiment. The authors thank the Histotechnology Core at the Wistar Institute for assistance with sectioning/preparation of sample specimens. They also thank the Flow Core at the Wistar Institute for assistance on the flow experiments. The following reagent was obtained through the AIDS Reagent Program, Division of AIDS, NIAID, NIH: VRC01 antibody from Xueling Wu, Zhi-Yong Yang, Yuxing Li, Gary Nabel, and John Mascola. This research was supported by NIH IPCAVD grant U19 Al109646-04, NIH/NIAID Collaborative Influenza Vaccine Innovation Centers contract 75N93019C00051, Inovio Pharmaceuticals Virus grant 5181101374 (awarded to D.B. Weiner), the W.W. Smith Charitable Trust 68112-01-383 (awarded to D.W. Kulp), and the Wistar Monica H.M. Shander Memorial Fellowship (awarded to Z. Xu).

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