Structurally-optimised HPV16 E7/E6 mRNA–LPP mediates dose-sparing efficacy via tumour microenvironment reprogramming
Shucai Sun, Yao Deng, Jiao Ren, Xiaotian Han, Jialuo Bing, Tangqi Wang, Zhanyihao Hao, Houwen Tian, Liang Zhang, Wenjie Tan
TL;DR
A new mRNA vaccine delivery system effectively activates immune responses against HPV-related cancers at low doses by reprogramming the tumor environment.
Contribution
The study introduces a codon- and untranslated region-optimized self-amplifying mRNA vaccine with dose-sparing efficacy via localized immune modulation.
Findings
Both nr-mRNA and sa-mRNA vaccines activated systemic antitumor immune responses and increased CD8⁺ T cells and NK cells in tumors.
sa-mRNA–LPP achieved comparable efficacy at one-fifth the dose of nr-mRNA–LPP.
The immune response from sa-mRNA–LPP was localized to the tumor site without activating peripheral lymphoid organs.
Abstract
The development of therapeutic vaccines against human papillomavirus (HPV)–associated malignancies remains challenging due to the immunosuppressive tumour microenvironment and the limited efficacy of existing delivery platforms. In this study, we designed and systematically compared two mRNA vaccine strategies based on a core–shell structured lipopolyplex (LPP) delivery system: a codon- and untranslated region-optimised non-replicating mRNA (nr-mRNA) and a self-amplifying mRNA (sa-mRNA). In the TC-1 murine model of HPV-driven cancer, both vaccine formulations effectively activated systemic antitumour immune responses, significantly enhancing the infiltration of functional CD8⁺ T cells and natural killer cells into tumours and promoting the repolarisation of tumour-associated macrophages towards an M1 phenotype. Notably, the sa-mRNA–LPP platform achieved comparable therapeutic efficacy…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7- —National key research and development plan of China
- —Medical Science Research Project of Hebei
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsRNA Interference and Gene Delivery · Immunotherapy and Immune Responses · Immune cells in cancer
Introduction
The World Health Organization (WHO) has launched a global initiative to eliminate cervical cancer, with targets for 2030 including: vaccinating 90% of girls by age 15 years, screening 70% of women with high-performance tests at least twice in their lifetime, and providing treatment to 90% of women with cervical cancer or precancerous lesions [1, 2]. Recommendations within China’s national approach include developing novel technologies to treat human papillomavirus (HPV) infection and cervical precancerous lesions, such as therapeutic HPV vaccines [3]. Currently, therapeutic HPV vaccines targeting HPV16/18 E6/E7 antigens are in the clinical research phase [4, 5]. Therapeutic mRNA cancer vaccines are among the most promising immunotherapeutic approaches, designed to activate T-cell responses against specific tumour-associated antigens and tumour-specific targets [6]. In preclinical studies, mRNA vaccines have demonstrated therapeutic efficacy against HPV-associated tumours [7–10]. However, the optimal mRNA vaccine optimisation strategy or platform remains to be determined.
The immunogenicity of non-replicating mRNA (nr-mRNA) vaccines is determined by their primary sequence-encoded information and critically regulated by their higher-order structural conformations [11]. Optimisation of nr-mRNA coding sequences requires a multidimensional strategy, including the adjustment of GC content [12], optimisation of secondary structure [11], enhancement of the codon adaptation index (CAI) [13], and modulation of minimum free energy (MFE). CAI and MFE are key regulatory elements that collectively determine nr-mRNA stability and translational efficiency [14]. Therefore, in developing therapeutic nr-mRNA vaccines against HPV-positive cervical precancerous lesions, utilising CAI and MFE as core optimisation parameters is expected to significantly improve clinical outcomes.
Beyond coding sequence design, the 5′ and 3′ untranslated regions (UTRs) play a pivotal role in nr-mRNA vaccine efficacy, as their regulatory sequences substantially influence mRNA stability [15]. Although natural UTRs derived from human α- and β-globin and rabbit β-globin have been widely adopted in vaccine design [16, 17], researchers are actively developing synthetic UTR sequences with enhanced regulatory functions [18, 19]. To systematically identify UTRs that improve protein expression, multiple screening technologies have been established, including systematic evolution of ligands by exponential enrichment [19] and computational design algorithms [20]. Notably, one fundamental distinction between Pfizer/BioNTech and Moderna vaccines lies in their unique UTR configurations [21]. However, no comprehensive head-to-head comparison has been conducted to evaluate the UTRs from these two approved vaccines. Consequently, developing novel mRNA vaccines with synchronously optimised UTRs holds promise for addressing current limitations in UTR screening systems for efficacy prediction.
Self-amplifying mRNA (sa-mRNA) is an innovative vaccine platform based on the alphavirus replication system, featuring two open reading frames that encode RNA-dependent RNA polymerase (replicase) and structural proteins [22]. Preserving intact viral replication machinery [23], the platform sustains self-amplification for up to 2 months in vivo, significantly enhancing the level and duration of antigen expression [24]. This unique mechanism enables sa-mRNA vaccines to induce stronger and more durable immune responses at lower doses [25]. However, although HPV16 sa-mRNA vaccines have shown efficacy in tumour models, their dose advantage over nr-mRNA remains unconfirmed [10].
In this study, core–shell structured lipopolyplex (LPP) nanoparticles, which have been established and characterised in previous studies [26–28], were used to encapsulate two RNA vaccine platforms: synthetic nr-mRNA incorporating synonymous codon optimisation with optimised 5′ and 3′ UTRs, and sa-mRNA. In the TC-1 tumour model, both platforms demonstrated potent antitumour efficacy by enhancing CD8⁺ T-lymphocyte infiltration and promoting M1 macrophage polarisation, thereby reprogramming the immunosuppressive tumour microenvironment (TME). Notably, 1 µg of E7E6 sa-mRNA–LPP exhibited therapeutic efficacy comparable to that of 5 µg of E7E6 nr-mRNA–LPP. These results confirm that both optimised mRNA vaccine platforms provide therapeutic benefits against HPV-associated malignancies through synergistic effects between enhanced antigen-specific cellular immunity and TME reprogramming. The findings of this study provide robust support for advancing the clinical development of rationally designed nr/sa-mRNA–LPP vaccines.
Results
Design and in vitro screening of nr-mRNA sequences with optimised structural stability and codon usage
To develop an effective nr-mRNA vaccine against HPV16, we first designed and screened nr-mRNA sequences encoding a mutated, non-oncogenic E7E6 fusion antigen (M-E7E6) with optimal expression properties. We engineered the M-E7E6 antigen by mutating key oncogenic domains of the native HPV16 E6 and E7 proteins (Fig. 1a and Table S1) [29], and then designed a series of nr-mRNA sequences (M0, M1, M2, M3) with varying degrees of codon optimisation and predicted secondary structures [11, 14]. Their characteristics were evaluated based on MFE (a key indicator of structural stability) and the CAI (a measure of codon optimality) (Fig. 1b and Tables S2–4).Fig. 1. Enhancing E7E6 protein expression and tumour inhibition using codon-optimised HPV16 E7E6 nr-mRNA. a Schematic representation of HPV16 E7E6 nr-mRNA vaccine design. b 2D scatter plot of the minimum free energy (MFE) and codon adaptation index (CAI) (M0, M1, M2, and M3) for the HPV16 E7E6 sequence. c Transfection of naked nr-mRNAs into HEK293T cells and subsequent analysis of the antigen expression by western blotting. d Quantitative analysis of the antigen expression levels derived from panel (c). e Schematic of the TC-1 tumour model establishment and treatment regimen. f Tumour growth kinetics post-vaccination with nr-mRNA-LPP formulations. Data represent mean ± SEM (n = 4–7 mice per group). Significance was determined by a two-sided Student’s t-test (panels d, f); *P < 0.05, **P < 0.01
A detailed analysis of the predicted metrics and in vitro expression revealed key distinctions. While the M1 and M2 sequences exhibited comparably favourable (low) MFE values (Fig. 1b and Table S4), their differential in vitro expression levels (Fig. 1c, d) suggested that nuanced differences in codon optimisation were a determining factor. In contrast, the M3 sequence, despite having the most stable predicted structure (lowest MFE), showed suboptimal expression, likely attributable to its lower CAI [13]. Notably, the fully optimised M0 sequence (CAI = 1) did not yield the highest expression, consistent with previous reports that over-optimisation can be detrimental to protein yield [11, 14, 30]. Furthermore, the expression of all nr-mRNAs was dose-dependent and transient over time (Fig. 1c, d). In summary, based on the comprehensive in vitro profiling, the M1 and M2 sequences were selected for subsequent in vivo evaluation due to their superior balance of structural stability and codon usage.
In vivo antitumour efficacy of lead nr-mRNA candidates
We next evaluated the antitumour efficacy of M1 and M2, formulated in lipid nanoparticles (LPP). TC-1 tumour-bearing mice received a single 5 or 30 µg dose of M1-LPP or M2-LPP four days post-tumour engraftment, using enhanced green fluorescent protein (eGFP)–LPP as a control (Fig. 1e). Both candidates elicited a strong, dose-dependent antitumour response compared to the control (Fig. 1f).
Critically, no significant difference in efficacy was observed between M1 and M2 at either dose. Therefore, we selected the M1 sequence, which demonstrated equivalent potency in vivo alongside superior in vitro expression, as our lead candidate for further investigation.
M1-LPP vaccine remodels the TME
To define the immunomodulatory mechanism of the M1-LPP vaccine, we analysed its impact on the tumour immune microenvironment. In the TC-1 tumour model, early leukocyte infiltration was observed post-immunisation (Fig. 2a), accompanied by significant expansion of natural killer (NK) cells (Fig. 2b), CD8⁺ T (Fig. 2c) and functional CD8⁺ T cell populations expressing interferon gamma (IFN-γ) or tumour necrosis factor alpha (TNF-α) (Fig. 2d). At a dose of 30 µg, this response was notably more pronounced than that in mice treated with 5 µg M1-LPP or controls. Conversely, the M1-LPP vaccine did not alter the frequency or proliferative capacity of regulatory T cells (Fig. 2e). Spatial analysis of tumour sections demonstrated a substantial increase in CD8^+^ T cell counts in tumours treated with 30 µg M1-LPP compared to those treated with 5 µg M1-LPP (Fig. 2f).Fig. 2. Influence of HPV16 E7E6 nr-mRNA-LPP vaccination on effector immune cell recruitment. a–c, e Percentage distributions of tumour-infiltrating CD45^+^ leukocytes, NK cells, CD8.^+^ T cells, and Treg cells depicted graphically. d Frequencies of IFN-γ⁺ and TNF-α⁺ among CD8⁺ T cells within TILs following E7₄₉-₅₇ peptide restimulation. Representative dot plots are displayed alongside the summary graph. f Immunofluorescence staining of CD8⁺ T cells (green) in TC-1 tumour sections. Nuclei are counterstained with DAPI (blue). Scale bar, 100 µm. g–j, Modulation of the myeloid compartment by nr-mRNA–LPP monotherapy. Shown are frequencies among CD45⁺ cells for CD11b⁺ cells (g), MDSCs (h), TAM (i), and M1‑like, M2‑like subsets and the M1/M2 TAM ratio (j). For panels a–e and g–j, data points represent individual mice (n = 4–7 per group). Data are presented as mean ± SEM. P values were calculated using an unpaired, two-tailed Student's t-test (a–e, g–j), with significance denoted by *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001
Concurrently, M1-LPP reshaped the immunosuppressive myeloid compartment. Flow cytometric analysis of CD11b⁺ myeloid populations revealed that the 30 µg dose significantly reduced total myeloid cells and myeloid-derived suppressor cells (MDSCs), an effect not observed at the 5 µg dose (Fig. 2g, h). Although the total tumour-associated macrophage (TAM) abundance remained unchanged (Fig. 2i), the 30 µg M1-LPP vaccine selectively expanded pro-inflammatory M1-like TAMs while depleting M2-like subsets, resulting in a 3.4-fold increase in the M1/M2 ratio (Fig. 2j). This reprogramming was not observed with the lower, less efficacious dose.
In summary, the M1-LPP vaccine elicits potent antitumour immunity by enhancing cytotoxic lymphocyte function and reprogramming the myeloid compartment towards an antitumour state, effectively reversing TME immunosuppression.
UTR optimisation enhances nr-mRNA translation efficiency
To investigate the impact of UTR sequences on protein expression, we systematically engineered nr-mRNA constructs by replacing their original UTRs with those derived from BNT162b2 (Fig. 3a). The resulting constructs, designated B-E7E6, exhibited significantly enhanced expression levels compared to their predecessors (Fig. 3c, d and Fig. S2). From this series, we selected two lead candidates—B1 and B2—which retain the identical coding sequences of M1 and M2, respectively, but are distinguished by the optimised UTRs (Tables S2–4). The B1 construct demonstrated the highest expression level, a property potentially attributable to its approximately 10% reduction in MFE, suggesting improved structural stability (Fig. 3b). Together, these results underscore the critical importance of UTR selection in enhancing gene expression and support the use of B1 as a lead candidate for further development.Fig. 3. Construction of HPV16 E7E6 nr-mRNA with various UTR elements, validation of its expression, and suppression of established tumours. a Schematic representation of the compositions of the two nr-mRNA-LPP vaccines (5′-mRNA1273-E7E6-mRNA1273-3′ and 5′-BNT162b2-E7E6-BNT162b2-3′). b 2D scatter plot of the minimum free energy (MFE) and codon adaptation index (CAI) (M1, M2, B1, and B2) for the HPV16 E7E6 sequence. c Western blot showing the antigen expression in HEK293 cell lysates at 6, 12, and 24 h after nr-mRNA transfection. d Quantification of the antigen expression calculated from (b). e Tumour growth kinetics in mice vaccinated with the indicated formulations (eGFP-LPP, M1/B1-LPP at 5 or 30 µg). Data represent mean ± SEM (n = 4–5). A two-sided Student’s t-test (panels d-e) was used to assess significance versus the control group (*P < 0.05, **P < 0.01)
UTR optimisation enables potent antitumour efficacy with reduced dosage
To determine whether the enhanced in vitro expression of B1 translated to improved in vivo performance, we prepared B1-LPP nanoparticles and evaluated their efficacy using the established therapeutic vaccination protocol (Fig. 1e). Dose–response analysis demonstrated that a single 5 µg dose of B1-LPP induced significantly stronger antitumour activity than the control vaccine. Notably, increasing the dose to 30 µg provided no additional therapeutic benefit, indicating that maximal efficacy was already achieved at the lower dose. Importantly, the antitumour protection conferred by 5 µg of B1-LPP matched that of 30 µg of the original M1-LPP vaccine (Fig. 3e).
These findings demonstrate that UTR optimisation enhances nr-mRNA stability and translational efficiency in vivo, enabling effective antitumour protection at a substantially reduced dose.
Low-dose B1-LPP vaccine elicits localised antitumour immunity within the TME
To determine whether the superior antitumour efficacy of the UTR-optimised B1-LPP vaccine at a lower dose was associated with enhanced TME immunity, we characterised the local and systemic immune responses following vaccination. Using the TC-1 tumour model, we performed comprehensive immunophenotyping of tumour-infiltrating leukocytes using flow cytometry and assessed antigen-specific T cell responses in the spleen using an ELISPOT assay.
Our analysis revealed that while the 5 µg dose of B1-LPP did not increase overall immune cell infiltration (as measured by CD45⁺ cell frequency; Fig. 4a), it specifically enhanced the recruitment and activation of key cytotoxic effectors within tumours. This included significant expansion of NK cells (Fig. 4b), CD8⁺ T cells (Fig. 4c), and functional CD8⁺ T cells producing IFN-γ (Fig. 4d) compared to both control and 5 µg M1-LPP vaccinated mice. Notably, these localised responses, though significantly stronger than those induced by the original M1-LPP at the same dose, were generally less pronounced than those achieved with 30 µg doses of either vaccine.Fig. 4. Enhanced intratumoural immune infiltration by low-dose B-E7E6 vaccine. Cellular composition of excised tumours was assessed using flow cytometry. a–c Percentages of tumour-infiltrating CD45⁺ leukocytes (a), NK cells (b), and CD8⁺ T cells (c). d CD8⁺ tumour-infiltrating lymphocytes were restimulated ex vivo with E7₄₉–₅₇ peptide; representative dot plots and frequencies of IFN-γ⁺ and TNF-α⁺ CD8⁺ T cells are shown. e IFN-γ ELISpot analysis of E7₄₉–₅₇‑specific T cell responses in splenocytes from each experimental group. Data are shown as mean ± SEM (n = 4–5), with significance assessed by two-sided Student’s t-test (a–d) or one-way ANOVA (e); *P < 0.05, **P < 0.01, ***P < 0.001
In contrast to the robust TME responses, systemic immunity remained largely unaltered. ELISPOT analysis showed no significant E7-specific T cell responses in the spleen across all treatment groups (Fig. 4e). Consistently, splenic flow cytometry revealed only minimal changes in immune cell activation, with modest increases in CD4⁺ T, IFN-γ⁺CD4⁺, and IFN-γ⁺CD8⁺ T cell frequencies in B1-LPP-treated mice, and no differences in NK, CD3⁺, or total CD8⁺ T cell populations compared to controls (Fig. S3).
In summary, the UTR-optimised B1-LPP vaccine at a 5 µg dose elicits potent and selective immune activation within the TME, characterised by enhanced recruitment and functional activation of cytotoxic lymphocytes, without inducing substantial systemic immune responses, highlighting its efficiency in generating localised antitumour immunity.
Efficient and sustained expression of HPV16 E7E6 from sa-mRNA constructs
To develop an alternative mRNA platform with potential dose-sparing advantages, we engineered a self-amplifying mRNA (SAM) system for HPV16 E7E6 antigen expression based on the Venezuelan equine encephalitis virus (VEEV) replicon. The SAM construct encodes the VEEV RNA-dependent RNA polymerase (nonstructural proteins nsP1-4) and places the E7E6 antigen under the control of a subgenomic promoter, enabling intracellular RNA amplification and sustained antigen production (Fig. 5a).Fig. 5. Design and validation of SAM vaccines. a Structured diagram of HPV16 E7E6 integration into a Venezuelan equine encephalitis virus-based SAM construct. Western blot analysis confirming the presence of E7E6 in lysates following SAM (b) and SAM–LPP (c) transfection. d Immunofluorescence images showing the antigen expression (green) in transfected HEK293 cells. Nuclei are counterstained with DAPI (blue). Scale bar, 100 µm. e Mean fluorescence intensity quantified from the immunofluorescence images in (d). Data are shown as mean ± SEM (n = 3), with significance assessed by one-way ANOVA with Tukey’s multiple comparison test, *P < 0.05, **P < 0.01
We first evaluated the kinetics and magnitude of antigen expression in vitro using both naked SAM and SAM encapsulated in lipid nanoparticles (SAM–LPP). Quantitative analysis showed that E7E6 protein levels increased rapidly at 6 h post-transfection and were maintained at high levels throughout the monitoring period across all tested SAM concentrations, with no significant differences observed between different doses (Fig. 5b). The SAM–LPP formulation exhibited similar kinetics and dose-independent expression patterns, with robust protein expression beginning at 6 h and accumulating to substantial quantities between 12 and 24 h post-transfection (Fig. 5c). Immunofluorescence imaging confirmed comparable expression patterns for both SAM and SAM–LPP, demonstrating persistent high-level protein production that was maintained throughout the 72-h observation period under all conditions and across all dose levels tested (Fig. 5d, e).
In summary, the VEEV-based SAM platform enables sustained expression of the HPV16 E7E6 antigen, independent of dosage within the tested range. This consistent expression profile, achieved with both naked and LPP-formulated SAM, supports the exploration of lower dosing regimens in subsequent immunisation studies.
Subcutaneous SAM–LPP vaccine induces more significant antitumour effects
To determine the optimal delivery route for the sa-mRNA vaccine, we compared the antitumour efficacy of SAM–LPP administered subcutaneously versus intramuscularly. The results showed that mice receiving subcutaneous immunisation exhibited significantly superior tumour protection compared to both the intramuscular group and phosphate-buffered saline (PBS) control. Notably, the complete response rate was 40% in the subcutaneous group, whereas no complete responses were observed in either the intramuscular group or the PBS control group (Fig. S3). These results indicate that subcutaneous administration is a more effective delivery approach for the SAM–LPP vaccine.
Comparison of the antitumour efficacy between SAM–LPP and B1-LPP vaccines
To evaluate the dose effects of the two mRNA platforms, we compared the antitumour efficacy of SAM–LPP and B1-LPP. Both 1 and 5 µg of SAM–LPP significantly inhibited tumour growth. Direct comparison revealed that 1 µg of SAM–LPP exhibited antitumour effects comparable to 5 µg of B1-LPP (Fig. 6a). These results indicate that SAM–LPP requires a lower dose to achieve similar therapeutic efficacy, demonstrating its dose-sparing advantage.Fig. 6. Examination of antitumour responses following the SAM vaccine administration. Tumour growth and tumour-infiltrating immune cell dynamics after SAM–LPP vaccination. a Analysis of tumour growth kinetics post-immunisation with various vaccine formulations (n = 4–5). b–e Percentage distributions of tumour-infiltrating CD45^+^ leukocytes, CD4^+^ T cells, Treg cells, and CD8^+^ T cells, depicted graphically (n = 3–4). f Ratio of CD8^+^ T cells to Tregs (CD8.^+^ T/Treg) among TILs (n = 3–4). g Analysis of CD8⁺ T cell responses in TILs. Frequencies of IFN-γ⁺ and TNF-α⁺ cells (upon E7₄₉–₅₇ peptide restimulation) and E7 multimer-specific cells are summarised, with corresponding representative dot plots (n = 3–4). h–j Modulation of the myeloid compartment by SAM–LPP monotherapy. Shown are frequencies among CD45⁺ cells for CD11b⁺ cells (h), MDSCs (i), and TAM-including M1-like, M2-like subsets and the M1/M2 ratio (j) (n = 3–4). a-J An unpaired, two-tailed Student's t-test was used for statistical comparisons. Results are shown as mean ± SEM, with asterisks indicating significance levels (*P < 0.05, **P < 0.01, ***P < 0.001)
SAM–LPP vaccine elicits localised antitumour immunity and reprograms the myeloid compartment
To delineate the immunological mechanisms underlying the antitumour efficacy of the SAM–LPP vaccine, we characterised its impact on immune cell populations in the TME and in systemic immune organs. To this end, we performed a comprehensive flow cytometric analysis of immune cells in TC-1 tumour tissues, spleen, and draining lymph nodes following vaccination with either 1 or 5 µg of SAM–LPP.
Within tumour tissues, SAM–LPP vaccination induced robust immune activation characterised by significantly enhanced CD45⁺ leukocyte infiltration (Fig. 6b). This was accompanied by substantial expansion of total CD8⁺ T cells (Fig. 6e), E7-multimer-specific CD8⁺ T cells (Fig. 6g), and functional CD8⁺ T cells producing IFN-γ or TNF-α compared to PBS controls. Critically, only the 5 µg dose significantly reduced regulatory T cell (Treg) frequency (Fig. 6d) and markedly increased the CD8⁺ T/Treg ratio (Fig. 6f), while the vaccine did not alter total CD4⁺ T cell numbers (Fig. 6c).
Analysis of the myeloid compartment revealed that 1 and 5 µg SAM–LPP significantly reduced total CD11b⁺ myeloid cell infiltration compared to PBS controls (Fig. 6h), although neither dose significantly altered MDSC or total TAM frequencies (Fig. 6i, j). Importantly, only the 5 µg dose specifically enhanced the M1/M2 TAM ratio by simultaneously expanding M1-polarised TAMs and reducing M2-polarised TAM infiltration (Fig. 6j). This dual effect of reducing immunosuppressive Tregs and promoting pro-inflammatory M1 macrophage polarisation provides a key mechanistic explanation for the superior antitumour efficacy observed at the 5 µg dose level.
In contrast, systemic immune responses were minimal. Splenic analysis revealed only a modest increase in TNF-α⁺CD8⁺ T cells, with no changes in total CD8⁺ T cells or CD45⁺ leukocyte populations (Fig. S4a). Lymph node immunophenotyping showed no vaccine-induced alterations in immune cell composition (Fig. S4b).
In summary, the SAM–LPP vaccine activates potent antitumour immunity primarily within the TME, with the 5 µg dose uniquely reducing immunosuppressive Treg populations and promoting beneficial M1 macrophage polarisation, accounting for its superior therapeutic efficacy.
SAM–LPP vaccine activates the PD-1/PD-L1 axis in the TME
To determine whether the SAM–LPP vaccine induces a microenvironment susceptible to immune checkpoint inhibition, we measured PD-1 and PD-L1 expression in TC-1 tumours. Flow cytometry analysis revealed that vaccination significantly upregulated PD-1 across all immune populations examined: CD45⁺ leukocytes, bulk CD8⁺ T cells, and E7-specific CD8⁺ T cells (Fig. 7a). Concurrently, PD-L1 levels were markedly elevated on tumour cells and MDSCs in vaccinated mice compared to controls, while TAMs showed no significant change (Fig. 7b). These findings demonstrate that the SAM–LPP vaccine co-ordinately enhances both PD-1 and PD-L1 expression in the TME. Given the established role of this axis in T cell exhaustion [7, 31], our results support combining the vaccine with PD-1/PD-L1 blockade to potentiate antitumour immunity.Fig. 7SAM–LPP vaccine induces PD-1/PD-L1 overexpression in the TME. Analysis of PD-L1 expression in TC-1 cells (a), TAMs (b), and MDSCs (c) isolated from tumours in vivo, and PD-1 expression in TC-1 tumour-infiltrating CD45^+^ (d), CD8^+^ T (e), and HPV16 E7-specific CD8.^+^ T(f) cells. Data are shown as mean ± SEM (n = 3–4), with significance assessed by one-way ANOVA with Tukey’s multiple comparison test, *P < 0.05, **P < 0.01
Discussion
The choice of synonymous codons can significantly influence the translation rate of nr-mRNA and the resultant protein yield [32]. Recent studies have elucidated the mechanisms by which codon use modulates translation, emphasising the importance of jointly optimising CAI and MFE [14]. This study comprehensively analysed nr-mRNA sequences designed based on structural stability and codon optimality and identified M1 and M2 as promising vaccine candidates. These results indicate that an optimised coding-region configuration enhances protein production and stimulates robust antitumour immune responses in nr-mRNA vaccines. Beyond coding sequence optimisation, UTR selection critically regulates nr-mRNA stability and translational efficiency [33]. The B-E7E6 construct employing BNT-162b2 UTRs demonstrated markedly higher protein expression than M-E7E6. Collectively, our data demonstrate that the combined optimisation of coding regions and UTR pairs synergistically enhances translational output, thereby stimulating robust antitumour immune responses in nr-mRNA vaccines. Importantly, this systematic optimisation of mRNA structural elements enabled a significant reduction in the required vaccine dose while maintaining high immunogenicity.
Consistent with the therapeutic efficacy observed for nr-mRNA vaccines, sa-mRNA vaccines elicited potent antitumour activity at both doses. The fundamental distinction between nr-mRNA and sa-mRNA vaccines lies in sa-mRNA's autonomous replication capacity post-delivery, enabling sustained antigen production with dose-sparing properties in murine models [25, 34]. Preclinical evaluations demonstrate sa-mRNA's enhanced immunogenicity: GRT-R910 elicited neutralising antibody titres comparable to those observed with conventional mRNA vaccines at doses of 10–100 µg in non-human primates [35]. Similarly, 1.25 µg sa-mRNA provided influenza protection equivalent to 80 µg of standard mRNA, representing a 64-fold dose reduction [36]. Notably, our findings reveal that a 1 µg dose of SAM–LPP elicited therapeutic efficacy comparable to levels observed with 5 µg B1-LPP, further underscoring the dose-saving advantage of our optimised platform.
This mRNA delivery system is a powerful, versatile platform with great potential for anticancer therapy. In this study, we developed core–shell structured LPP nanoparticles that encapsulated two types of RNA vaccines. Unlike conventional mRNA delivery platforms, the LPP platform enters dendritic cells via phagocytosis, and nucleic acid vaccine-treated dendritic cells exhibit an enhanced antigen presentation capacity [37]. This mechanism effectively prevents organ-specific side effects induced by vaccine particles, such as the observed hepatic accumulation associated with the lipid nanoparticle mRNA formulation [38, 39]. The biosafety of LPP has been verified by SW-BIC-213, an LPP-delivered mRNA vaccine evaluated in clinical trials and granted EUA in Laos [40, 41]. The SW-BIC-213 mRNA vaccine manifests a favourable safety profile. All adverse reactions observed during the study were tolerable, transient, and resolved spontaneously. While the complex microenvironment of human tumours (such as dense stroma) poses challenges for nanomedicine delivery, LPP technology itself has clinically demonstrated its effectiveness in barrier penetration and immune activation through other mRNA vaccines [28]. This provides a crucial theoretical and practical foundation for advancing our vaccine to clinical applications.
Notably, in the investigation of administration routes for SAM–LPP, subcutaneous injection demonstrated a key advantage: it elicited significantly superior antitumour efficacy compared to intramuscular injection, achieving a 40% complete response rate, an outcome absent in the intramuscular group. Furthermore, although the vaccine failed to induce significant CD8⁺ T-cell responses in peripheral lymphoid organs or the spleen, it triggered potent immune activation specifically within the TME. We hypothesise that this distinct response pattern and the enhanced therapeutic effect may stem from robust antigen-specific T-cell responses initiated by optimised antigen presentation through skin dendritic cells [28]. Future studies are warranted to further elucidate the precise mechanisms by which vaccination routes modulate T-cell immunity.
To systematically delineate the antitumour immune mechanisms, we performed comprehensive flow cytometric profiling. Our analysis revealed a coordinated shift in the immune landscape following vaccination: significant expansion of cytotoxic CD8⁺ T cells and NK cells, along with polarisation of TAMs towards the M1 phenotype [42–44]. This proinflammatory reprogramming was accompanied by a reduction in key immunosuppressive populations [45]—specifically, decreased MDSCs with the 30 µg M1-LPP vaccine and reduced Treg infiltration with the 5 µg SAM–LPP vaccine. Furthermore, the 5 µg SAM–LPP vaccine induced upregulation of PD-1 on T cells and PD-L1 on tumour cells, indicative of adaptive immune resistance. The striking contrast between the robust TME response and the minimal activation detected in spleen or lymph nodes confirms that the vaccine's critical antitumour immunity is primarily localised within the tumour milieu.
Our study revealed a nonlinear dose–response relationship for B1-LPP, with the 5 µg dose achieving antitumour efficacy comparable to that of the 30 µg dose. This plateau effect suggests that the therapeutic response may have reached a saturation state. We propose that this phenomenon could be associated with multiple immunosuppressive mechanisms within the tumour microenvironment (TME). Established research has clearly demonstrated that under hypoxia-induced metabolic stress, CD8⁺ T cells develop mitochondrial dysfunction and impaired metabolic adaptability [46]; disrupted proteostasis in T cells directly leads to exhaustion of effector functions [47]; upregulation of immune checkpoint molecules such as PD-1/PD-L1 further suppresses T cell activation and cytotoxic capacity [48]; and polarisation of regulatory T cells (Tregs) towards a TH1-Treg phenotype exacerbates the overall immunosuppressive state [49]. These interconnected mechanisms collectively create the paradoxical situation of "increased T-cell quantity but compromised functionality."
Moreover, although the CD8⁺ T-cell immune responses induced by 1 and 5 µg of SAM–LPP were comparable in magnitude, the latter led to more substantial tumour regression. This seemingly paradoxical observation stems from the unique immunomodulatory capacity of the 5 µg SAM–LPP dose: it significantly reduced Treg cell infiltration and the Treg/CD8⁺ T-cell ratio in the TME and reprogrammed TAMs, polarising them towards the antitumour M1 phenotype. In multiple solid tumours, increased Treg infiltration is associated with reduced patient survival and higher risk of recurrence [50, 51], while a high Treg/CD8⁺ T-cell ratio often indicates poor prognosis and resistance to immunotherapy [52]. Furthermore, M1-type TAMs can synergistically suppress tumour progression through multiple mechanisms, including direct phagocytosis of tumour cells, cytokine secretion to activate T cells, and enhancement of antibody-dependent cellular cytotoxicity [53–56]. Our findings further demonstrate that the reduction in Treg infiltration, together with M1-type TAM polarisation, constitutes a key antitumour pathway independent of conventional CD8⁺ T-cell immunity and plays an essential role in achieving effective tumour suppression. Based on these results, we propose that this mechanism underlies the superior antitumour efficacy of the 5 µg SAM–LPP dose.
"Adaptive immune resistance," a key mechanism of tumour immune escape, occurs when activated T cells upregulate PD-1 and secrete IFN-γ, leading to concomitant PD-L1 upregulation on tumour cells and subsequent inhibition of T-cell function [57, 58]. Our study demonstrates that although the SAM–LPP vaccine therapy potently activates antitumour immunity, it precisely triggers the synchronous upregulation of PD-1 and PD-L1. These findings provide a compelling rationale for combining the SAM–LPP vaccine with PD-1/PD-L1 blockade, a strategy designed to overcome this limitation and achieve synergistic therapeutic efficacy.
While this study advances understanding of HPV16-targeted mRNA vaccines, two critical limitations emerge. First, the antitumour effect is primarily associated with enhanced infiltration of CD8⁺ T and NK cells, reduced Treg cell infiltration, and M1-type macrophage polarisation. Current studies have clearly demonstrated the critical roles of CD8⁺ T [10, 58, 59], NK [58, 60], and Treg [61] cells in antitumour immune responses. However, the specific molecular mechanisms by which vaccines induce M1-type macrophage polarisation remain unclear.
Second, the HPV antigen design is restricted to the E6/E7 oncoproteins, neglecting immunogenic non-oncogenic antigens such as E2 and E5. The incorporation of E2/E5 epitopes may enhance the broad-spectrum efficacy of vaccines by mitigating immune evasion mechanisms [62, 63] and augmenting tumour-infiltrating antigen-specific CD4^+^ T and CD8^+^ T cell responses within the TME [64]. Future research should focus on developing polyvalent mRNA vaccines that incorporate E2/E5 together with E6/E7, combined with single-cell profiling technologies to systematically analyse the dynamic interactions between CD8⁺ T and NK cells in preclinical models. Furthermore, although the TC-1 model was chosen as the field standard for initial HPV vaccine evaluation, additional in vivo and ex vivo models, such as a patient-derived xenograft model, are required to strengthen the translational relevance of this study.
In summary, we developed and systematically optimised nr/sa-mRNA-LPP platforms that achieve significant dose reduction while eliciting potent antitumour immunity. The sa-mRNA-LPP platform in particular demonstrated superior immunogenicity and therapeutic efficacy. Crucially, our comprehensive immune profiling reveals that these vaccines orchestrate a multifaceted antitumour response within the TME, characterised by coordinated activation of cytotoxic effector cells and reduction of immunosuppressive populations. These findings strongly support the clinical translation of this promising vaccine platform for HPV-associated malignancies.
Materials and methods
Cell lines
HEK 293 T cells (ATCC CRL-3216) were maintained in high-glucose Dulbecco’s modified Eagle medium (Corning) supplemented with 10% heat-inactivated foetal bovine serum (FBS; Gibco) and 1% penicillin–streptomycin (Sigma-Aldrich). The TC-1 murine tumour line (HPV16 E6/E7-transformed C57BL/6 pulmonary epithelial cells [65]) was authenticated through short tandem repeat profiling and mycoplasma testing prior to establishing master cell stocks. Experiments utilised cells between passages 5–9 post-revival.
Design and development of mRNA vaccines encoding the HPV16 E7E6 fusion protein
The HPV16 nr/sa-mRNA vaccine construct used in this study adopts an E7-E6 gene order. To achieve fusion expression of the two genes, the stop codon of E7 was removed along with the start codon of E6. No additional linker peptide sequence was introduced, resulting in direct fusion. Regarding key inactivation mutations, in E7, we introduced C24G and E26G mutations to abrogate its binding capacity to pRB and its transforming function [66, 67]; furthermore, in E6, we introduced an L57G mutation, which effectively disrupts its oncogenic activity of mediating p53 degradation [68, 69]. E7E6 was optimised for structural stability and codon optimality [11, 14]. The E7E6 coding sequences were cloned into the optimised mRNA production plasmids to generate nr-mRNA or sa-mRNA. Two mRNA variants were produced: 1) nr-mRNA containing N1-methylpseudouridine substitutions and 2) unmodified sa-mRNA. Both formulations were encapsulated in LPP nanoparticles using microfluidic mixing technology as previously characterised [70]. Control constructs expressed eGFP under identical formulation conditions.
Engineering of an mRNA-LPP nanoplatform
This study utilised a structurally stabilised and codon-optimised HPV16 E7E6 fusion gene to construct both N1-methylpseudouridine-modified nr-mRNA and unmodified sa-mRNA. All mRNA sequences were encapsulated into lipid-protamine-polyplex nanoparticles using microfluidic technology. The mRNA-LPP nanoparticles were prepared according to an established method [26–28], entailing a two-step procedure: First, protamine sulphate was dissolved in 25 mM sodium acetate buffer (pH 5.2), then mixed with the mRNA solution in 10 mM citrate buffer (pH 4.0) at a 5:1 volume ratio, and incubated at room temperature for 30 min to form polyplexes. Subsequently, the lipid phase containing an ionisable lipid, DOPE (Avanti Polar Lipids, Birmingham, USA), cholesterol (A.V.T Pharmaceutical, China), and mPEG-DMG (Avanti Polar Lipids, Birmingham, USA) at specified molar ratios was prepared in ethanol. This organic phase was combined with the aqueous polyplex solution at a 1:3 volume ratio (ethanol:aqueous) using a microfluidic mixer to drive nanoparticle self-assembly. The assembled formulations were purified by dialysis against PBS (pH 7.4), concentrated using ultrafiltration, and sterilised through 0.22 µm filters. The final products were characterised by dynamic light scattering to determine particle size and polydispersity index, by cryo-transmission electron microscopy to assess morphology, and by a fluorescence-based quantification method to evaluate mRNA encapsulation efficiency, as previously described [26–28]. mRNA nanoparticles encoding eGFP were prepared using identical procedures to those for the controls.
Western blotting
HEK 293 T cells (1 × 10⁶) were plated in complete medium and allowed to adhere for 12 h prior to transfection. Cells were transfected with varying concentrations of E7E6 mRNA (0.5, 1.0, and 2.0 µg) using Lipofectamine MessengerMAX™-based transfection reagent. Cellular lysates were prepared at specified time intervals (6, 12, and 24 h post-transfection) through mechanical disruption in radioimmunoprecipitation assay buffer containing protease inhibitors. Protein quantification was performed using a bicinchoninic acid protein assay kit (Beyotime Biotechnology) according to standardised protocols.
For immunoblotting, equivalent protein quantities were separated through an automated capillary-based western blot system (ProteinSimple). Target proteins were detected using specific primary antibodies: mouse monoclonal anti-HPV16 E7 (Santa Cruz Biotechnology, SC-65711; 1:200 dilution) and anti-β-actin (ZSGB-BIO, TA-09; 1:1000 dilution) as loading control. Immunocomplexes were quantified through chemiluminescent detection using the Compass software (ProteinSimple), with three biological replicates performed for each experimental condition. Signal intensities were normalised to β-actin expression levels for comparative analysis.
Animal models and therapeutic protocols
All experiments utilised female C57BL/6 wild-type mice (8–10 weeks old) sourced from Charles River Laboratories and maintained under specific pathogen-free conditions. Animal procedures were conducted in accordance with AAALAC International guidelines and were approved by the Institutional Animal Ethics Committee of the National Institute for Viral Control and Prevention, China CDC (Ethical Approval #20,230,504,032). To establish tumour models, mice received subcutaneous injection of 2 × 10^4^ TC-1 cells suspended in 50 µL PBS into the right flank, designated as Day 0. We randomised all mice into treatment cohorts using a computer-generated random number table, without consideration of physiological characteristics, and confirmed comparable initial mean body weight across groups. On Day 4 post-tumour implantation, these mice were administered a single subcutaneous dose of the respective mRNA-LPP formulations. Tumour growth was tracked twice per week at intervals of 3–4 days (e.g., every Wednesday and Sunday) by blinded operators using digital callipers, with volume calculated as (Width^2^ × Length)/2. The study concluded on Day 28, when all mice were euthanised and tumour tissues collected for immune profiling.
Tissue processing and single-cell suspension preparation
Tumour specimens underwent enzymatic and mechanical disaggregation using a Miltenyi Tumour Dissociation Kit (cat. 130–096–730) in conjunction with a gentleMACS™ dissociator following the manufacturer’s optimised protocol to generate single-cell suspensions. Concurrently, splenic and lymph node tissues were mechanically homogenised using sterile plungers and then filtered through a 70 µm nylon mesh into RPMI 1640 medium (Gibco). All cellular suspensions underwent identical centrifugation conditions (1000 × g, 10 min, 4 °C) in a refrigerated centrifuge, after which the pelleted cells were reconstituted in RPMI 1640 supplemented with 10% heat-inactivated FBS (Gibco) for downstream immunological assays.
Flow cytometry
Single-cell suspensions from tumour, spleen, and lymph node tissues were prepared by mechanical dissociation and enzymatic digestion, followed by red blood cell lysis using an ammonium chloride solution. Cell viability was assessed using a viability dye, such as Zombie NIR™ Fixable Viability Kit (BioLegend, 423,106) and subsequently incubated with purified anti-mouse CD16/32 antibody (Fc block, BD Pharmingen) at room temperature for 5 min. The cells were then incubated with a cocktail containing PE-conjugated H-2Dᵇ/E7₄₉–₅₇ (RAHYNIVTF) tetramer and fluorophore-conjugated monoclonal antibodies against the following cell surface markers: CD45, CD8α, CD4, NK1.1, CD11b, F4/80, MHC II, Gr-1, and CD25 (Table S5) for 30 min at 4 °C in the dark, followed by washing with staining buffer (PBS containing 2% FBS and 0.1% sodium azide). For intracellular antigen staining, specialised kits were used according to the target antigen: the Cytofix/Cytoperm kit (BD Biosciences, 554,714) was used for cytokines (IFN-γ and TNF-α) and CD206, while the Transcription Factor Buffer Set (BD Biosciences, 562,574) was used for the transcription factor Foxp3 (Table S5). Cells were incubated with specific intracellular antibodies for 30 min at 4 °C in the dark and washed with permeabilisation/wash buffer. For nuclear antigen staining, cells were permeabilised with a nuclear permeabilisation buffer and incubated with nuclear antibodies for 50 min at room temperature in the dark, followed by washing with nuclear wash buffer. For cytokine analysis, T cells require peptide stimulation prior to staining. CD4^+^ T/CD8^+^ T cells were stimulated with 10 µg/mL E7_49–57_ (RAHYNIVTF, SciLight Biotechnology, LLC) peptide with an equal volume of DMSO as the control in the presence of 10 µg/mL Brefeldin A (Sigma) for 5 h at 37 °C. Samples were stored at 4 °C in the dark until flow cytometric analysis. Data were acquired using a FACSCanto II flow cytometer (BD Biosciences) and analysed using the FlowJo software version 10.8 (TreeStar), with compensation set using single-stain controls and cells gated sequentially to identify populations of interest. Gating strategies are shown in Table S6 and Fig. 5.
Immunofluorescence microscopy
Paraffin Sects. (8 µm) of the tumours were deparaffinised and rehydrated through a series of xylene and graded ethanol solutions. Antigen retrieval was performed in EDTA Antigen Retrieval Buffer (pH 9.0) for 30 min using a microwave. The sections were then blocked for 1 h at room temperature in Dulbecco’s PBS containing 1% bovine serum albumin (BSA), 5% goat serum, and 0.2% Triton X-100. For immunostaining, the sections were incubated overnight at 4 °C with a primary antibody against CD8^+^ T cells (Abcam, ab217344, 1:200 dilution). The following day, the sections were incubated for 1 h at room temperature with the secondary antibody, Goat anti-rabbit IgG Alexa Fluor 488 (Thermo Fisher Scientific, A11008, 1:1000 dilution). Nuclei were counterstained with 4',6-diamidino-2-phenylindole (DAPI; Solarbio). Immunofluorescence images were captured using the tile-scanning function of a confocal laser-scanning microscope (Leica Microsystems).
Statistical analyses and data presentation
Data are reported as the mean ± standard error of the mean (SEM). Statistical analyses were conducted using GraphPad PRISM 9. Comparisons between two groups were made using an unpaired two-tailed Student's t-test, while comparisons across multiple groups were conducted by one-way analysis of variance (ANOVA) with a Tukey’s multiple comparison test.
Supplementary Information
Supplementary Material 1.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Gote V, Bolla PK, Kommineni N, Butreddy A, Nukala PK, Palakurthi SS, et al. A Comprehensive Review of m RNA Vaccines. Int J Mol Sci. 2023;24(3). 10.3390/ijms 24032700.10.3390/ijms 24032700 PMC 991716236769023 · doi ↗ · pubmed ↗
- 2Xia X. Detailed Dissection and Critical Evaluation of the Pfizer/Bio N Tech and Moderna m RNA Vaccines. Vaccines (Basel). 2021;9(7). 10.3390/vaccines 9070734.10.3390/vaccines 9070734 PMC 831018634358150 · doi ↗ · pubmed ↗
- 3Blakney A, Ip S, Geall A. An Update on Self-Amplifying m RNA Vaccine Development. Vaccines. 2021;9(2). 10.3390/vaccines 9020097.10.3390/vaccines 9020097 PMC 791154233525396 · doi ↗ · pubmed ↗
- 4Minnar CM, Chariou PL, Horn LA, Hicks KC, Palena C, Schlom J, et al. Tumor-targeted interleukin-12 synergizes with entinostat to overcome PD-1/PD-L 1 blockade-resistant tumors harboring MHC-I and APM deficiencies. J Immunother Cancer. 2022;10(6). 10.1136/jitc-2022-004561.10.1136/jitc-2022-004561 PMC 924093835764364 · doi ↗ · pubmed ↗