mRNA-encoded mutant HPV16/18 vaccines promote specific T-cell responses and synergize with anti-PD-1 checkpoint blockade in mediating therapeutic tumor regression in mice

Abstract

Background

Persistent infection with high-risk human papillomavirus (HPV) 16 and 18 is a major driver of human cancer, including head and neck and cervical cancers. Although prophylactic vaccines prevent infection, effective therapies for established HPV-related cancers are needed. In this study, we developed a messenger RNA (mRNA)-based therapeutic vaccine encapsulated in lipid nanoparticles (LNP) encoding mutated E6/E7 antigens from HPV16/18 and an optimized co-stimulatory adjuvant (MTS107).

Methods

The mRNA backbone of the vaccine was engineered with mutated HPV16/18 E6/E7 at the N-terminus to prevent the degradation of p53 and pRb. A T2A self-cleaving peptide was incorporated to separate the antigenic components from the co-stimulatory signal genes. An optimal LNP formulation was identified based on its expression efficiency and safety profile both in vitro and in vivo. The efficacy and mechanism of action of the lead mRNA-LNP were subsequently evaluated in both TC-1 (HPV16⁺) and HPV18-transgenic MC38 syngeneic tumor models.

Results

The optimized mRNA antigen construct translated proteins at high levels in vitro without affecting p53 or pRb. In HPV16⁺ and HPV18⁺ syngeneic mouse tumor models, MTS107 effectively targeted dendritic cells and macrophages, inducing potent dose-dependent and time-dependent antitumor activity associated with the expansion of HPV-specific CD8⁺ T cells and enhanced intratumoral infiltration. Combination with an anti-programmed cell death protein-1 (PD-1) antibody (αPD-1) led to complete tumor remission.

Conclusions

These findings support the clinical evaluation of mRNA-based therapeutic vaccines like MTS107 for HPV-driven malignancies.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Persistent infection with high-risk human papillomavirus (HPV) 16 and 18 is a major cause of human cancers. Despite the availability of prophylactic vaccines, effective therapeutic strategies for established HPV-related cancers are needed.

WHAT THIS STUDY ADDS

• This study demonstrated that MLX0487-23, a lipid nanoparticle encapsulating a P016-2-an messenger RNA (mRNA) construct encoding four HPV antigens and the co-stimulatory factor granulocyte-macrophage colony-stimulating factor (GM-CSF) (referred to as MTS107), effectively expanded HPV-specific CD8⁺ T cells and elicited potent antitumor responses in preclinical models. These findings support its potential as a therapeutic mRNA vaccine candidate for the clinical evaluation of HPV-driven malignancies.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• MTS107 was established as a potent therapeutic vaccine that induces HPV-specific cytotoxic T lymphocyte responses through the expression of optimized HPV16 and HPV18 antigens enhanced by co-stimulatory adjuvants, such as GM-CSF and OX40L. This supports the advancement of mRNA-based therapeutic vaccines as a viable treatment option for HPV-related malignancies. These findings may guide future research, inform clinical trial design, and influence the development of therapeutic vaccines.

Background

Persistent infection with human papillomavirus (HPV) is a predominant cause of human cancer, including head and neck (HNC) and cervical (CC) cancers, leading to approximately 900,000 and 660,000 new cases annually, with 400,000 and 350,000 deaths, respectively.^{1 2} Globally, CC is the fourth leading cause of cancer-related mortality in women, and HNC ranks among the top 10, posing a considerable health burden. HPV infection is responsible for approximately 80% of these cases.^{3 4} Although prophylactic vaccines have greatly reduced the incidence of HPV-related infections, effective therapeutic treatments for established HPV-driven cancers are urgently needed.⁵

Messenger RNA (mRNA) vaccines that encode tumor-associated antigens, tumor-specific antigens, and associated cytokines represent a promising class of cancer immunotherapies.6,8 By using delivery systems, such as lipid nanoparticles (LNP), mRNA vaccines are protected from degradation and enhanced for targeted cellular uptake, thus improving druggability.⁹ The success of mRNA vaccines in combating the COVID-19 pandemic has demonstrated their safety and effectiveness, which has accelerated their development.^{10 11} mRNA-based vaccines present antigenic epitopes via the endogenous major histocompatibility complex class I, activating antigen-specific CD8⁺ T cells that can eventually mediate the specific killing of malignant tumor cells. Therefore, the activation of antigen-specific cytotoxic T lymphocytes (CTL) is a critical component of cancer immunotherapy.¹² Advances in artificial intelligence have facilitated the design of more efficient and stable mRNA sequences, allowing the inclusion of multiple antigens within a single construct and offering flexibility, precision, and rapid scalability in vaccine development.¹³

Among the 300 known HPV genotypes, HPV16 and HPV18 are the most strongly linked to tumor development. The early region (E) oncoproteins, E6 and E7, promote tumorigenesis by degrading p53 and pRb, respectively, leading to immune evasion and uncontrolled cell proliferation.¹⁴ Given their constitutive expression in HPV-associated cancers and their strong immunogenicity as foreign viral proteins, E6 and E7 are promising targets for antigen-specific therapeutic vaccines. As the activation of CTL requires co-stimulatory signals, incorporation of a second signal-associated co-stimulator into mRNA vaccines may enhance their efficacy.¹⁵

In this study, we aimed to develop an LNP-encapsulated mRNA vaccine encoding multiple antigenic signals, including mutated E6/E7 antigens from HPV16 and 18,^{16 17} along with an optimized co-stimulatory adjuvant to facilitate both antigen presentation and T-cell stimulation. The vaccine demonstrated strong antitumor activity in both HPV16⁺ and HPV18⁺ syngeneic mouse tumor models, inducing effective antigen-specific CTL activation after immunization. When combined with an anti-programmed cell death protein-1 (PD-1) antibody (αPD-1), it exhibited superior synergistic efficacy. Additionally, we explored the potential synergy between granulocyte-macrophage colony-stimulating factor (GM-CSF) and OX40L, which act at different stages of immune stimulation.^{18 19} We observed that the combination enhanced antitumor activity by increasing the activation and proliferation of antigen-specific CTL. These findings support the clinical evaluation of mRNA-based therapeutic vaccines against HPV-driven cancers.

Materials and methods

Cell culture and transfection

HPV16⁺ TC-1 cells and human embryonic kidney cell line HEK293T were purchased from Immocell Biotechnology (Xiamen, China) or the American Type Culture Collection. The HPV18⁺ MC38 (HPV18-E6E7) cell line was generated by Kyinno Biotechnology (Beijing, China) by transducing the HPV18 E6 and E7 genes into MC38 cells. TC-1 and MC38 (HPV18-E6E7) cells were cultured in Roswell Park Memorial Institute (RPMI)-1640 medium (Gibco), and HEK293T cells were cultured in Dulbecco’s Modified Eagle Medium (Gibco) supplemented with 10% fetal bovine serum in a 5% CO₂ incubator at 37°C. Different mRNAs were transfected into HEK293T cells using the transfection reagent MessengerMAX (Invitrogen), in accordance with the manufacturer’s instructions.

mRNA synthesis and LNP formulation

The mutated E6 and E7 proteins of HPV16 (E6^L57G/C113RE7^C24G/E26G) and HPV18 (E6^L52GE7^C27G/E29G) were arranged with signal peptide (SP) and a co-stimulator (figure 2A). The SP and four antigens were connected via a GS linker, whereas the co-stimulator and adjacent antigen sequence were linked using a T2A module. The mRNA was synthesized in vitro via T7 RNA polymerase-mediated transcription of a linearized DNA template containing N1-methylpseudouridine instead of uridine. The linearized DNA template incorporates 5′ and 3′ untranslated regions. Cap-modified mRNAs were synthesized through the co-transcriptional incorporation of CleanCap AG (TriLink) as a trinucleotide Cap 1 analog at a final concentration of 4 mM. The mRNA was polyadenylated using Escherichia coli poly(A) polymerase (Novoprotein) for 40 min at 37°C. Purification was performed through lithium chloride precipitation, and the mRNA was resuspended in ultrapure nuclease-free water. LNP formulations were prepared by mixing an ethanol-dissolved lipid mixture with 25 mM citrate buffer (pH 4.0) containing mRNA, using a microfluidic mixer. The generated formulations were concentrated via ultrafiltration (Millipore), filtered through a 0.22 µm membrane, and stored at 4°C until use. The particle size, polydispersity index, and RNA encapsulation efficiency were analyzed using Zetasizer Pro (Malvern), DynaPro NanoStar (Wyatt), and RiboGreen assays (Invitrogen), respectively.

Figure 2. P016-2 emerges as the lead mRNA candidate with potent antitumor activity. (A) Schematic of mRNA constructs: mutated HPV16-E6/E7 and HPV18-E6/E7 sequences were fused via GS linkers, whereas the co-stimulatory domains and adjacent antigens were linked by T2A modules. All constructs contained an N-terminal signal peptide (SP). (B) Western blotting confirms the expression of mRNA-encoded HPV18 E7 in transfected HEK293T cells. (C) ELISA detected secreted mGM-CSF in the supernatants from P016-1-transfected and P016-2-transfected cells. (D) Schematic representation of the mRNA screening efficacy model. (E) Body weight remained unaffected in mice treated with P016-1 or P016-2. (F) P016-2 induced dose-dependent tumor suppression, outperforming P016-1. (G) P016-2 exhibited superior dose-dependent survival benefits compared with P016-1. (H) Body weight was unaffected in P016-Neg and P016-2 treatment groups. (I) P016-2 exhibited superior dose-dependent antitumor efficacy compared with P016-Neg. (J) P016-2 exhibited superior dose-dependent survival benefits compared with P016-Neg. (K) P016-2 effectively promoted DC maturation, which was much stronger than that of P016-Neg. (L) Mutations in P016-2 prevented pRb and p53 degradation (mean±SD; n ≥3; *p<0.05, **p<0.01). CR, complete remission; DC, dendritic cell; HPV, human papillomavirus; mGM-CSF, mouse granulocyte-macrophage colony-stimulating factor; TGI, tumor growth inhibition.

Subcutaneous syngeneic tumor model

Female C57BL/6 mice (6–8 weeks old) were purchased from Charles River and housed under specific pathogen-free conditions. The HPV16⁺ syngeneic tumor model was established through subcutaneous implantation of a TC-1 tumor fragment (millet size) into the right flank. The HPV18⁺ model was generated by injecting 2×10⁶ MC38 (HPV18-E6E7) cells in 100 µL phosphate-buffered saline (PBS) into the right flank. mRNA vaccines at varying doses were administered intramuscularly into the rectus femoris of the right hind limb (50 µL per dose) every 5 or 7 days.

The experimental details are summarized as follows: the TC-1 xenograft model (established by implanting tumor tissue with an initial volume of approximately 6 mm) was used for LNP/mRNA screening and co-stimulatory factor synergy analysis. Vaccine dosing commenced at this volume (day 0), and vaccines were administered at 5-day intervals (0.02–0.5 µg/mouse in 50 µL; two doses). For the late-stage TC-1 tumor models, treatment was initiated at larger tumor volumes (TV) (approximately 100 or 500 mm³), with dosing at 7-day intervals (0.1–10 µg/mouse in 50 µL; two or three doses). In the MC38 (HPV18-E6E7) tumor model, treatment began at a TV of approximately 100 mm³ using: (1) mRNA vaccine (0.05–5 µg/mouse in 50 µL; three doses at 5-day intervals), and/or (2) anti-mouse PD-1 antibody (αPD-1, Bio X Cell; 200 µg in 100 µL PBS intraperitioneally injected (i.p) every 3 days; six doses). Mechanistic studies were conducted using TC-1 models at an initial TV of approximately 6 or 100 mm³. Mice received the mRNA vaccine (0.05–10 µg/mouse in 50 µL; two doses at 5-day intervals) and were sacrificed 7 days after the final dose for tissue collection and analysis. The groups comprised five to six mice each. Tumor size and body weight were measured two times per week. Spleens, tumors, and blood samples were collected at designated time points. Antitumor efficacy was evaluated based on tumor growth inhibition (TGI) and complete remission. Tumor-bearing animals were euthanized on reaching the following humane endpoints: >20% body weight loss, TV exceeding 2,000 mm³, or moribund status.^{20 21}

Western blotting

HEK293T cells were lysed 24 hours post-transfection using sodium dodecyl sulfonate (SDS) lysis buffer (Beyotime), and the culture media were collected. Proteins (30 µg per lane) were separated through SDS-polyacrylamide gel electrophoresis (PAGE) and transferred to polyvinylidene fluoride membranes (Millipore). Membranes were probed with primary antibodies against HPV18 E7 (Santa Cruz Biotechnology), pRb (CST), p53 (CST), or GAPDH (HUABIO), followed by horseradish peroxidase (HRP)-conjugated secondary antibodies (HUABIO). Signals were detected using an ECL kit (Millipore) and quantified using ImageJ software.²² Protein expression levels were normalized to those of GAPDH.

ELISA for GM-CSF analysis

Cell culture supernatants from transfected HEK293T cells were collected 24 hours post-transfection. GM-CSF levels were measured using a commercial ELISA kit (Abcam), according to the manufacturer’s protocol.

Quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR)

The construction of MC38 (HPV18-E6E7) cells was confirmed using qRT-PCR. Briefly, total RNA from MC38 (HPV18-E6E7) cells was extracted using TRIzol (QIAGEN) and reverse transcribed into complementary DNA (Vazyme). qRT-PCR was performed using SYBR Green (Vazyme) on an Applied Biosystems 7500 system (Thermo Fisher Scientific). Primers: HPV18-E6: F 5′-TAATAAGGTGCCTGCGGTGC-3′, R 5′-GTTGGAGTCGTTCCTGTCGT-3′. HPV18-E7: F 5′-ACATTTACCAGCCCGACGAG-3′, R 5′-GGTCGTCTGCTGAGCTTTCT-3′. GAPDH primers were used in the kit. Data were analyzed using the 2^−ΔΔCt method.

Tissue preparation

Tumors were dissociated using a tumor dissociation kit (Miltenyi) on a MACS Octo Dissociator. The cells were filtered through 70 µm strainers (Falcon). The spleens were homogenized through a 70 µm cell strainer and washed with RPMI-1640 medium (Gibco). The cell suspensions were centrifuged (300×g, 7 min) and resuspended in PBS or RPMI-1640 supplemented with 10% fetal bovine serum for further testing.

Immunohistochemistry staining

Tissues (muscle, liver, lungs, and spleen) from mice injected with LNP-encapsulated green fluorescent protein (GFP) mRNA (10 µg/50 µL/mouse) were fixed in formalin, embedded in paraffin, and sectioned. After deparaffinization and rehydration, endogenous peroxidases were blocked using 3% H₂O₂. Sections were stained with anti-GFP antibody (Abcam), followed by HRP-conjugated secondary antibody (Abways) and diaminobenzidine substrate. Counterstaining was performed using hematoxylin or PAS.

Flow cytometry

Flow cytometric analysis was performed on single-cell suspensions of blood, tumors, and spleens. Approximately 1×10⁶ cells per sample were stained with LIVE/DEAD-AmCyan (BioLegend) for 15 min at room temperature (RT, 15–25°C), followed by staining with H2-D^b HPV16 E7 tetramer-RAHYNIVTF-PE (MBL) for 30 min at RT. Cells were surface stained with anti-mTCR-FITC (BioLegend), anti-mCD4-APC-Cy7 (BioLegend), anti-mCD8-APC (BioLegend), anti-mCD19-BV605 (BioLegend), and anti-NK1.1-PE-Cy7 (BioLegend) for 20 min at RT for HPV-specific CTL and potential innate immune response detection. Dendritic cell maturation was assessed via surface staining with anti-mCD45-APC-7 (BioLegend), anti-mCD11c-BV421 (BioLegend), anti-I-A/I-E-PD594 (BioLegend), anti-mCD83-APC (BioLegend), or anti-mCD86-BV650 (BD Biosciences) for 20 min at RT.

To detect T-cell activation, after staining with tetramer, the cells were further surface stained with anti-mCD3-PerCP-Cy5.5 (BioLegend), anti-mCD4-APC-Cy7 (BioLegend), anti-mCD8-FITC (BioLegend), anti-mCD25-BV421 (BioLegend), anti-mCD69-BV650 (BioLegend), and Ki-67-APC (BioLegend) for 20 min at RT. To detect programmed death-ligand 1 (PD-L1) and PD-1 expressions, tool cells were first stained with anti-mPD-L1 (Kyinno), followed by the secondary antibody anti-human IgG-Fc-PE (Abcam). Single-cell suspensions from tumor tissues were surface stained with anti-mPD-1-BB700 (BD Biosciences). To evaluate LNP-targeting specificity, splenocytes were collected from mice administered MLX0487-23-encapsulated GFP or anchored variable heavy-chain domain of heavy-chain antibody (aVHH) mRNA (10 µg/mouse). Splenocytes from the aVHH mRNA group were surface stained with VHH-iFlour488 (AlpVHHs). Data was acquired on an Oncocyte Penteon flow cytometer (Agilent) and analyzed using FlowJo V.10.8.1.

Enzyme-linked immunosorbent spot (ELISpot) assay

The collected single splenocytes were resuspended in Opti-MEM (Thermo Fisher Scientific) at a density of 5×10⁵ cells/mL and then seeded at a density of 5×10⁴ cells/well in 96-well plates using an ELISpot assay kit (Dakewe). The splenocytes were stimulated with 100 µL of the indicated peptide pool (2 µg/mL/peptide), Opti-MEM medium only, or 250 ng/mL phorbol 12-myristate 13-acetate (PMA)+5 µg/mL ionomycin (Thermo) and incubated for 18 hours at 37°C with 5% CO₂. Interferon (IFN)-γ spots were quantified using an ELISpot kit (Dakewe).

Statistical analysis

All experiments were repeated at least in triplicate. All data from the study were analyzed using IBM SPSS Statistics V.25.0, and the results are presented as mean±SD. Means between the two groups of samples were compared using a t-test. One-way analysis of variance was used to compare means among multiple groups of samples. P values<0.05 were deemed significant.

Results

LNP MLX0487-23 demonstrates superior antitumor efficacy with spleen-targeting activity

Six lipids/LNPs were screened, with SM102-LNP serving as the benchmark LNP and BNT113-E7 mRNA as the active pharmaceutical ingredient.^{23 24} The METiS-LNPs had an average size of 70–130 nm, as measured by dynamic light scattering (figure 1A), a polydispersity index below 0.2 (figure 1B), and an encapsulation efficiency exceeding 95% (figure 1C), all of which were comparable to those of SM102-LNP. To evaluate the antitumor activity of the lead LNP, a TC-1 syngeneic tumor model was established in C57 mice, following the immunization timeline shown in figure 1D. Intramuscular administration was selected because of its superior efficacy compared with subcutaneous and intravenous injections (online supplemental figure S1). After two doses, all LNP-mRNAs exhibited dose-dependent antitumor activity (figure 1F) but had no significant effect on body weight (figure 1E).

Figure 1. METiS-LNP MLX0487-23 demonstrates superior antitumor efficacy with spleen-targeting activity. (A) Average size, (B) polydispersity index, and (C) encapsulation efficiency of SM102-LNP and six METiS-LNPs. “Empty” denotes LNPs without messenger RNA encapsulation. (D) Efficacy screening of LNPs was performed using a TC-1 tumor-bearing model. Two intramuscular vaccinations were administered starting at a tumor volume of~6 mm³ every 5 days. (E) Body weight remained unaffected across LNP-encoded BNT113-E7 (BNT) treatment groups. (F) MLX0487-23 exhibited optimal efficacy, achieving near-complete tumor regression at both low (0.05 µg) and high (5 µg) doses. (G) MLX0487-23 selectively targeted the spleen. (H) MLX0487-23 selectively targeted splenic dendritic cells and macrophages (mean±SD; n ≥3; *p<0.05, **p<0.01). aVHH, anchored variable heavy-chain domain of heavy-chain antibody; CR, complete remission; DC, dendritic cell; GFP, green fluorescent protein; LNP, lipid nanoparticles; TGI, tumor growth inhibition.

MLX0487-23 displayed the highest potency, achieving complete tumor regression at both low and high doses (figure 1F). To further confirm the efficacy of MLX0487-23, immunohistochemical staining and flow cytometry analyses were performed following intramuscular injection of GFP or aVHH mRNA encapsulated in these LNPs. The results revealed strong GFP expression in tissue-resident and infiltrating immune cells at the injection site, with much stronger and more abundant expression in spleen tissues (figure 1G). Flow cytometry confirmed intracellular GFP and surface aVHH expression in splenic macrophages and dendritic cells (figure 1H), highlighting their splenic-targeting capacity.

mRNA-encoded HPV16/18 therapeutic vaccine with co-stimulator GM-CSF demonstrates superior antitumor activity

The mRNA backbone of the vaccine was designed with mutated HPV16/18 E6/E7 to prevent p53 and pRb degradation, thereby reducing carcinogenic risk. GM-CSF was incorporated at the N-terminus or C-terminus of the antigens (figure 2A). To assess mRNA translation efficiency, an anti-HPV18 E7 antibody was used for western blotting, which showed that P016-Neg, P016-1, and P016-2 effectively expressed the HPV18 E7 protein in HEK293T cells, with P016-2 exhibiting the highest antigen expression (figure 2B). ELISA confirmed robust mGM-CSF secretion from both P016-1 and P016-2 cells, with P016-1 exhibiting the highest levels (figure 2C). Next, a TC-1 syngeneic tumor model was used to evaluate the activity of MLX0487-23-encapsulated mRNAs (figure 2D). Following treatment with the two doses, no significant changes in body weight were observed in any group (figure 2E,H). All three mRNAs exhibited dose-dependent antitumor efficacy, with P016-2 outperforming P016-1 and P016-Neg (figure 2F,G,I,J). Flow cytometric analysis of splenocytes revealed that P016-2 significantly enhanced dendritic cell (DC) maturation compared with P016-Neg, indicating a much stronger capacity to stimulate T cells (figure 2K). Furthermore, P016-2 transfection did not affect pRb or p53 levels in HEK293T cells, whereas P016-wt (the same format as P016-2 with wild-type HPV16/18 E6/E7) and 16-E6^wt+16-E7^wt (wild-type HPV16 E6/E7) significantly reduced pRb and/or p53 expression (figure 2L). These findings highlight P016-2 encapsulated by MLX0487-23 as a promising therapeutic mRNA vaccine candidate.

To investigate the therapeutic efficacy of the mRNA vaccine against HPV18⁺ tumors, an HPV18⁺ MC38 (HPV18-E6E7) cell line was constructed (figure 3A) and evaluated in a syngeneic mouse model (figure 3B). Immunization with the P016-2 vaccine had no significant effect on body weight (figure 3C). Tumor suppression data revealed a dose-dependent therapeutic effect, with both the medium (0.5 µg) and high (5 µg) doses achieving significant TGI rates of 48.83% and 71.79%, respectively (figure 3D). Survival analysis also revealed the dose-dependent efficacy of the P016-2 vaccine (figure 3E).

Figure 3. P016-2 vaccine exhibits dose-dependent antitumor efficacy in an HPV18⁺ MC38 (HPV18-E6E7) syngeneic mouse model. (A) qRT-PCR confirmed HPV18-E6E7 expression in engineered MC38 (HPV18-E6E7) cells. (B) MC38 (HPV18-E6E7) tumor-bearing mice received P016-2 vaccine at 0.05, 0.5, or 5 µg/dose. (C) Body weight remained stable across treatment groups. (D) P016-2 induced dose-dependent tumor suppression, with TGI values of 6.56%, 48.83%, and 71.79% at respective doses. (E) P016-2 vaccine exhibits dose-dependent survival benefits in an HPV18⁺ MC38 (HPV18-E6E7) syngeneic mouse model. (Mean±SD; n ≥3; *p<0.05, **p<0.01). HPV, human papillomavirus; qRT-PCR, quantitative real-time reverse transcription polymerase chain reaction; TGI, tumor growth inhibition.

Next, the therapeutic potential of the mRNA vaccine against late-stage tumors was assessed. Treatment was initiated when the TV reached either 100 mm³ or 500 mm³ on average, with three doses of 0.1, 1, and 10 µg (figure 4A). The P016-2 vaccine did not affect body weight in any of the dose groups (figure 4B,E). Tumor growth was robustly inhibited in a dose-dependent manner, with the highest dose (10 µg) achieving TGI rates of 58.11% and 62.73% in the two late-stage tumor models (figure 4C,F). The survival analysis results were consistent with the antitumor efficacy data, demonstrating that the P016-2 vaccine conferred dose-dependent survival benefits in late-stage tumor models (figure 4D,G).

Figure 4. P016-2 vaccine inhibits late-stage tumor progression. (A) TC-1 tumor-bearing mice received three P016-2 doses (0.1, 1, and 10 µg) starting at TV=100 mm³ or two doses at TV=500 mm³. (B) Body weight remained stable in mice that were treated starting at TV=100 mm³. (C) P016-2 vaccine provided a dose-dependent antitumor efficacy; high-dose P016-2 (10 µg) significantly suppressed tumors (TGI=58.11%) when administered at TV=100 mm³. (D) P016-2 vaccine exhibits dose-dependent survival benefits in late-stage TC-1 syngeneic mouse models (TV=100 mm³). (E) Body weight remained unaffected in mice that were treated starting at TV=500 mm³. (F) Two doses of P016-2 vaccine started at TV=500 mm³ exhibited dose-dependent antitumor efficacy; high-dose P016-2 (10 µg) significantly suppressed late-stage tumors (TGI=62.73%). (G) P016-2 vaccine exhibited dose-dependent survival benefits in late-stage TC-1 syngeneic mouse models (TV=500 mm³) (mean±SD; n=5; **p<0.01). TGI, tumor growth inhibition; TV, tumor volume.

P016-2 vaccine synergizes with immune checkpoint inhibitor

As the efficacy of a therapeutic vaccine depends on active T cells, T-cell exhaustion can limit its effectiveness. To address this, we explored the antitumor efficacy of the P016-2 vaccine combined with αPD-1 in the PD-L1-positive MC38 (HPV18-E6E7) tumor-bearing model (figure 5A). Flow cytometry analysis of tumor-infiltrating lymphocytes (TILs) from P016-2-treated tumors revealed high PD-1 expression, with a positivity rate exceeding 80% (figure 5B). MC38 (HPV18-E6E7) tumor-bearing mice were treated with either the P016-2 vaccine or αPD-1 alone or in combination, starting when the TV reached approximately 100 mm³ (figure 5C). Body weight monitoring revealed no significant changes among the treatment groups (figure 5D). Tumor growth analysis demonstrated that monotherapy with either P016-2 or αPD-1 significantly inhibited tumor growth but showed only moderate therapeutic efficacy. In contrast, the combination of P016-2 and αPD-1 led to complete tumor remission, which was much stronger than either monotherapy alone (figure 5E,F).

Figure 5. P016-2 vaccine synergizes with αPD-1 to enhance tumor regression. (A) Flow cytometry confirmed PD-L1 expression on MC38 (HPV18-E6E7) cells. (B) >80% of the tumor-infiltrating lymphocytes from P016-2 vaccine-treated MC38 (HPV18-E6E7) tumors were PD-1 positive. (C) MC38 (HPV18-E6E7) tumor-bearing mice received monotherapy (0.5 μg P016-2 vaccine or 10 mg/kg αPD-1) or combination therapy. (D) Body weight remained unaffected across groups. (E) Monotherapy with P016-2 (TGI=45.33%) or αPD-1 (TGI=50.97%) significantly reduced tumor growth, whereas combination therapy induced complete remission. (F) Monotherapy with P016-2 (survival rate=83.33%) or αPD-1 (survival rate=66.67%) significantly promoted the survival benefits in the HPV18⁺ MC38 (HPV18-E6E7) syngeneic mouse model, whereas combination therapy received a complete survival benefit (mean±SD; n≥3; **p<0.01). HPV, human papillomavirus; ICI, immune checkpoint inhibitor; i.m., intramuscular injection; i.p. intraperitoneal injection; mRNA, messenger RNA; PD-1, programmed cell death protein-1; PD-L1, programmed death-ligand 1; Q3D, every three days; Q5D, every five days; TGI, tumor growth inhibition.

P016-2 vaccine effectively stimulates HPV-specific T cells and enhances the immune cell infiltration

To understand the mechanism by which the P016-2 vaccine inhibits tumor growth, spleen and tumor tissues were collected to characterize the immune cells after treatment (figure 6A). Efficacy data on day 12 after two doses showed that the dose-dependent TGIs ranged from 72.99% to 88.37%, with no significant impact on body weight (figure 6B,C). To assess the activation of antigen-specific T cells, splenocytes were subjected to an IFN-γ ELISpot assay using HPV16-E7 peptide pools. A significant dose-dependent response was observed (figure 6D), with IFN-γ spot-forming units showing a strong positive correlation with tumor inhibition (figure 6E). Additionally, antigen-specific CD8⁺ T cells in the spleens and tumors were quantified using an H2-D^b HPV16 E7 tetramer throughflow cytometry. As shown in figure 6F, P016-2 significantly stimulated the production of HPV-specific CD8⁺ T cells, with 11.9% of the CD8⁺ T cells in the spleen being HPV16-E7 tetramer-positive. The proportion of HPV-specific CD8⁺ T cells in TILs increased in a dose-dependent manner, ranging from 11.0% to 31.4% (figure 6F). These findings further demonstrated a strong correlation between HPV-specific CD8⁺ T cells and tumor suppression (figure 6G). Moreover, the infiltration of CD3⁺ T cells significantly increased following P016-2 treatment (figure 6H). Most infiltrating CD3⁺ T cells were CD8⁺ T cells and not CD4⁺ T cells (figure 6I). Mechanistic validation was performed in an additional model, with vaccine administration initiated at a TV of approximately 100 mm³ and tissue harvested 7 days after the second vaccination. These results were consistent with those observed in the TV of the approximately 6 mm³ cohort (online supplemental figure S3), confirming the reproducibility across different tumor burdens. These results highlight the potent immunostimulatory effects of the P016-2 vaccine in generating a robust HPV-specific CD8⁺ T-cell response and promoting immune cell infiltration into tumors.

Figure 6. P016-2 vaccine drives HPV-specific CD8⁺ T-cell responses and T-cell infiltration. (A) TC-1 tumor model for mechanistic analysis. (B) Body weight remained unaffected pre-sampling. (C) P016-2 vaccine inhibited tumor growth in a dose-dependent manner pre-sampling. (D) IFN-γ ELISpot revealed elevated HPV-specific T-cell frequencies. (E) Inverse correlation between HPV-specific IFN-γ⁺ T cells and tumor volume. (F) Flow cytometry with H2-D^b HPV16 E7 tetramer confirmed HPV-specific CD8⁺ T cells induced by P016-2 vaccine. (G) Inverse correlation between HPV-specific CD8⁺ T cells and tumor volume. (H) Increased T-cell infiltration in tumors induced by P016-2 vaccine. (I) Infiltrating CD3⁺ T cells were predominantly CD8⁺, with minimal CD4⁺ populations (mean±SD; n≥3; *p<0.05, **p<0.01). DMSO, dimethyl sulfoxide; ELISpot, enzyme-linked immunosorbent spot; HPV, human papillomavirus; IFN, interferon; mTCR, anti-mouse T cell receptor antibody; PMA, phorbol 12-myristate 13-acetate; SFU, spot forming units; TGI, tumor growth inhibition.

GM-CSF synergizes with OX40L to enhance immune microenvironmental remodeling and antitumor activity

To further optimize the immunostimulatory effects of the vaccine, we evaluated the combination of GM-CSF and OX40L, two co-stimulators that enhance DC and T-cell activation, respectively. OX40L is encoded by P024-4, which shares an mRNA backbone similar to that of P016-2 (figure 7A). TC-1 tumor-bearing mice were treated with P016-2 or P024-4 alone or in combination (figure 7B) and did not show any change in body weight (figure 7C). Although both monotherapies significantly inhibited tumor growth, the combination therapy exhibited superior efficacy, achieving a TGI of 87.14% compared with 78.05% with P016-2 alone and 58.84% with P024-4 monotherapy (figure 7D). Flow cytometric analysis revealed that the addition of P024-4 promoted the expansion of HPV-specific CTLs (figure 7E) and their proliferation (figure 7F). These findings highlight the potential of co-stimulatory molecule synergy to improve the therapeutic efficacy of mRNA-based cancer vaccines.

Figure 7. GM-CSF synergizes with OX40L to augment tumor regression. (A) Structure of P024-4 mRNA. (B) TC-1 tumor-bearing mice received monotherapy (0.15 μg P016-2 or P024-4) or combination therapy (0.075 µg each). (C) Body weight remained unaffected across groups. (D) Combination therapy achieved stronger tumor suppression (TGI=87.14%) versus monotherapy (P016-2: 78.05%; P024-4: 58.84%). (E) Combination therapy further expanded HPV-specific CTLs compared with monotherapy. (F) Combination therapy further expanded total CTLs compared with monotherapy. (Mean±SD; n≥3; *p<0.05, **p<0.01). CTLs, cytotoxic T lymphocytes; GM-CSF, granulocyte-macrophage colony-stimulating factor; HPV, human papillomavirus; SP, signal peptide; TGI, tumor growth inhibition.

Discussion

In this study, we demonstrated the efficacy of a P016-2 mRNA-based therapeutic vaccine encapsulated in MLX0487-23 LNP for treating HPV16/18-positive tumors. Through the induction of HPV-specific CTLs and the inclusion of the co-stimulatory factor GM-CSF, the vaccine exhibited strong antitumor activity. Although several therapeutic DNA vaccines targeting HPV have been developed with notable success, they require electroporation for delivery and carry the potential risks of genomic integration, local swelling, and genotoxicity.⁵ These challenges highlight the need for alternative platforms, such as mRNA-based vaccines. Preclinical evaluations of LNP-mRNA vaccines have demonstrated favorable safety and efficacy profiles. Although current candidates employ either HPV16 monovalent or HPV16/18 bivalent designs, further optimization of vaccine constructs remains necessary, and additional clinical validation is needed to understand their therapeutic benefits.25,28 For our mRNA vaccine, we used a bivalent strategy targeting both HPV16 and HPV18 with parallel co-expression of a co-stimulatory factor to potentially enhance immunogenicity.

Our vaccine leverages the spleen-targeting capabilities of MLX0487-23 LNP, which offers considerable advantages. On intramuscular administration, the vaccine recruits DCs and macrophages to the local injection site and initiates an immune response through the rapid accumulation of mRNA-translated tumor antigens. Moreover, intramuscular injections provide better patient compliance than subcutaneous and intravenous routes, making them a favorable option for clinical applications. The spatial organization of antigens and co-stimulatory factors within the mRNA construct, separated by a T2A element,²⁹ plays a crucial role in vaccine efficacy. Prioritizing antigen expression while maintaining low levels of GM-CSF ensures optimal immune activation. Our results confirmed that this configuration leads to robust antitumor activity.

HPV-specific CTLs have been identified as the primary effector cells responsible for tumor elimination.²⁸ Although CD4⁺ T cells are known to support CTL activation,^{30 31} we did not observe a significant increase in CD4⁺ T cells in splenocytes or TILs. Despite this, CD4⁺ T cells in splenocytes comprised 54% of the total CD3⁺ T cells, suggesting that sufficient helper activity was present to support CTL function. The strong correlation between HPV-specific CTL levels in both the spleen and TILs and tumor suppression further underscores the central role of CTLs in mediating antitumor immunity.

The P016-2 vaccine exhibited potent dose-dependent tumor regression, even in late-stage tumors, indicating its potential efficacy in advanced HPV-related cancers. Moreover, the vaccine elicited robust memory T-cell responses, with stable levels of CD8⁺ central memory T cell (T_cm) and a dose-dependent increase in CD8⁺ effector memory T cell (T_em). The high percentage of T_em cells in TILs (>90%) suggested an active immune state capable of sustained tumor clearance (online supplemental figure S2). Given that therapeutic vaccines rely on active T cells, overcoming T-cell exhaustion is essential to maximize efficacy.³² Immune checkpoint inhibitors, such as αPD-1, have been widely used to reverse T-cell exhaustion in patients with HPV-associated HNC and CC.³³ Our study demonstrated that the combination of P016-2 and αPD-1 significantly enhanced antitumor activity in the MC38 (HPV18-E6E7) syngeneic mouse model, in which tumors express PD-L1.

To further enhance the efficacy of our mRNA vaccine, we explored various co-stimulatory factors, including Flt3L, HMGB1, HSP70p, CCL21, GM-CSF, and OX40L. Among these, only GM-CSF and OX40L demonstrated immune-enhancing activity, with GM-CSF showing the strongest stimulation. Given their complementary roles, with GM-CSF promoting DC maturation and OX40L enhancing T-cell activation,^{19 34} we hypothesized that combining them would lead to the synergistic promotion of immunity. Although no significant increase in DC maturation was detected, CTL proliferation was markedly enhanced. As a single DC can activate thousands of T cells, even a small number of DCs can drive potent T-cell responses, which explains the strong antitumor effect observed despite minimal DC expansion.^{35 36} Furthermore, mRNA itself is a ligand for toll-like receptors 3, 7, and 8, which can elicit the activation of innate immunity.³⁷

In conclusion, our study demonstrated that the MLX0487-23 (LNP)-encapsulated P016-2 (mRNA) vaccine effectively induced an HPV-specific CTL response through the expression of optimized antigens from HPV16 and HPV18 combined with potent co-stimulatory adjuvants such as GM-CSF and OX40L. The vaccine exhibited robust antitumor activity against HPV-positive tumors and showed promise for clinical application to treat HPV-associated cancers. Combinatorial strategies using immune checkpoint inhibitors can further enhance therapeutic efficacy. These findings support the continued development of mRNA-based therapeutic vaccines as viable treatment options for HPV-associated malignancies.

Data availability statement

Data are available upon reasonable request.