mRNA-LNP vaccination orchestrates systemic immunity to control human papillomavirus-positive head and neck squamous cell carcinoma

Abstract

Messenger RNA (mRNA) vaccines have demonstrated significant potential in cancer immunotherapy by activating both innate and adaptive immunity. However, the detailed cellular and molecular dynamics underpinning these systemic immune responses remain incompletely understood. In this study, we characterized the systemic immune landscape following human papillomavirus (HPV)-targeted mRNA-lipid nanoparticle (LNP) vaccination using single-cell RNA sequencing (scRNA-seq) in a murine model of HPV-positive head and neck squamous cell carcinoma (HNSCC). Our study revealed a coordinated remodeling of the systemic immune landscape, involving the tumor microenvironment (TME), tumor-draining lymph nodes (TDLNs), spleen, and blood. Notably, we pioneered a distinct interferon-stimulated gene (ISG) signature across multiple lymphoid subsets in TDLNs, driven by the LNP component, which contributed to rapid, non-antigen-specific immune activation. Additionally, HPV mRNA-LNP vaccination induced an antigen-specific cycling burst of immune cells that mediated tumor control through a systemic coordination of multi-directional differentiation into anti-tumor cell compositions. These findings enhance our understanding of how mRNA-LNP vaccination orchestrates systemic anti-tumor responses and highlight the therapeutic potential of targeting ISG-expressing and cycling immune cells to improve vaccine efficacy, paving the way for future clinical applications in HPV-related cancers.

Graphical abstract

mRNA vaccines reshape the entire body’s immune landscape to fight cancer. Ren and colleagues discovered a rapid, innate immune program in lymph nodes and a coordinated burst of antigen-specific cell cycling. These findings reveal how mRNA vaccines orchestrate a systemic attack on tumors, advancing immunotherapy for HPV-related cancers.

Introduction

Messenger RNA (mRNA) vaccines have been remarkably successful in combating infectious disease pandemics, generating widespread interest in their application for tumor therapy across a wide range of malignancies.¹ mRNA sequences can be easily tailored to encode any target antigen, thus enabling large-scale and rapid development of mRNA vaccines against various cancers including solid tumors and hematological malignancies.^2,3,4 This versatility makes mRNA vaccines a powerful platform with broad applicability beyond infectious diseases, opening new avenues for cancer immunotherapy. The wildly used mRNA delivery carrier, lipid nanoparticle (LNP), ensures efficient mRNA expression in vivo and endows the vaccine with inherent adjuvant characteristics.^5,6 As a consequence, mRNA-LNP vaccination has demonstrated remarkable potential in orchestrating systemic anti-tumor responses, not only within the local tumor environment but also across distal immune sites.⁷

mRNA-LNP vaccines are known to have a unique capacity to activate both the innate and adaptive immune system.^8,9 Upon delivery, LNP can facilitate the uptake of mRNA into antigen-presenting cells (APCs). Once inside the cell, the mRNA is translated into the antigen of interest, which can then be processed and presented on the cell surface to prime CD8⁺ cytotoxic T cells and CD4⁺ helper T cells.^10,11,12 CD4⁺ helper T cells then support multiple aspects of the adaptive response, including enhancing the cytotoxic activity of CD8⁺ T cells, and aiding B cells in the production of antigen-specific antibodies.^13,14 Moreover, the mRNA component itself acts as a potent activator of innate immune pathways.¹⁵ The presence of foreign mRNA can trigger pattern recognition receptors (PRRs), such as toll-like receptors (TLRs), which recognize viral RNA and initiate a rapid immune response. Activation of these PRRs leads to the production of pro-inflammatory cytokines and type I interferons, which serve as signaling molecules that recruit additional immune cells to the site of vaccine administration and amplify the immune response.^16,17 These studies indicate that the coordination of immune responses across various compartments is crucial for the anti-tumor effects of mRNA vaccines. However, current research is insufficient to elucidate how mRNA cancer vaccines regulate the activation of local and distant immune sites, and the detailed cellular and molecular dynamics of systemic immune responses remain to be fully characterized.

We previously developed a mRNA-LNP vaccine encoding human papillomavirus (HPV) 16 E7 antigen, which can mediate the regression of HPV-positive head and neck squamous cell carcinoma (HNSCC) by activating HPV-specific CD8⁺ T cells and regulating their functional commitment.¹⁸ In this study, we aimed to depict the single-cell transcriptional landscape of systemic immunity elicited by HPV mRNA-LNP vaccination and characterize key immune subpopulations influencing vaccination efficacy. This study might progress our understanding of how mRNA-LNP vaccination orchestrates systemic anti-tumor immunity and provide insights into its optimization.

Results

Remodeling of the systemic immune landscape induced by mRNA-LNP vaccination

As we previously reported, intravenous administration of the HPV mRNA-LNP vaccine generated a burst of HPV-specific CD8⁺ T cells that were effective in mediating the regression of established tumors.¹⁸ Mice were implanted with mEERL tumors on the subcutaneous flank and treated on day 9 (prime), day 14 (first booster), and day 19 (second booster) with HPV mRNA-LNP vaccination. Consistent with our prior data, mice that received HPV mRNA-LNP vaccination successfully mediated tumor control (Figures S1A–S1D) without inducing detectable pathological damage and lethal cytokine storm (Figures S1E and S1F). To comprehensively characterize the vaccination-induced alterations in the systemic immune landscape, we utilized flow cytometry to sort CD45⁺ immune cells from tumors, blood, tumor-draining lymph nodes (TDLNs) and spleens of tumor-bearing (TB) mice on day 25, 6 days after the third dose of HPV mRNA-LNP vaccination, with mice mock-vaccinated with phosphate-buffered saline (PBS) being used as controls (Figure 1A). This methodology allowed for a detailed analysis of immune cell populations and provided insights into the cellular dynamics post-vaccination. To further establish baseline immune cell populations, we also sorted CD45⁺ immune cells from the spleens of no tumor (NT) mice, mock-vaccinated with PBS or vaccinated with HPV mRNA-LNPs, on day 16. This high-resolution analysis enabled the identification and visualization of seven CD8⁺ T cell subsets, eight B cell subsets, 16 CD4⁺ T cell subsets, and 18 myeloid subsets, based on canonical markers (Figures 1B; S2–S6; Table S2. Classical markers of four major cell types in scRNA-seq samples (related to Figure 1), Table S3. Differentially expressed genes of different CD8+ cell clusters applied in scRNA-seq (related to Figure S3), Table S4. Differentially expressed genes of different CD4+ cell clusters applied in scRNA-seq (related to Figure S4), Table S5. Differentially expressed genes of different B cell clusters applied in scRNA-seq (related to Figure S5), Table S6. Differentially expressed genes of different myeloid cell clusters applied in scRNA-seq (related to Figure S6)). In addition, leveraging the power of single-cell RNA sequencing (scRNA-seq), we were able to further dissect the transcriptional landscape of these immune subsets, revealing nuanced shifts in immune cell states post-vaccination.

Figure 1.

Remodeling of the systemic immune landscape induced by mRNA-LNP vaccination

(A) Schematic overview of the experimental workflow for this study. (B) t-SNE plot showing 49 clusters identified from NT and TB mice, with different colors representing CD8⁺ T cells (pink), CD4⁺ T cells (brown), B cells (yellow), and myeloid cells (green). (C) t-SNE plots showing the tissue distribution (peripheral blood, spleen, TDLN, tumor, left) and treatment distribution (vaccine vs. control, right) and mice distribution (NT vs. TB, lower) of single cells. NT: no tumor; TB: tumor bearing. TDLN: tumor-draining lymph node. (D) Tissue (spleen, blood, TDLN and tumor) distribution and treatment (vaccine vs. control) distribution of six main immune cell clusters. (E) Tissue distribution (spleen, blood, TDLN, and tumor) and treatment (vaccine vs. control) distribution of 49 immune cell subclusters.

Figure 1 illustrates the dynamic changes in immune cell composition following HPV mRNA-LNP vaccination, highlighting how these changes correspond to the immune response observed in different tissues, including tumors, TDLNs, and spleens (Figures 1C and 1D). This visualization not only emphasizes the shifts in immune cell populations but also underscores the complex interplay between the tumor microenvironment (TME) and peripheral immune system. Additionally, we observed significant alterations in the frequency and activation status of specific immune cell subsets, reinforcing the importance of a robust systemic immune response in mediating anti-tumor efficacy.

Interestingly, we found that vaccination or tumor implantation alone resulted in similar changes of the immune microenvironment in the spleens from NT mice, which were characterized by the activation and migration of naive B cells as well as the rapid accumulation of antigen-presenting cells (APCs) including dendritic cells (DCs) and macrophages. Meanwhile, across treatment and tissue groups, we identified a sophisticated immune network corresponding to vaccination-mediated anti-tumor efficacy in TB mice (Figure 1E). Overall, TME, secondary lymphoid organs and peripheral blood exhibited divergent changing patterns following HPV mRNA-LNP vaccination. Critically, as the body’s largest reservoir of immune cells, the spleens following the vaccination were characterized by a remarkable decrease of activated T cells and B cells, accompanied by the rapid accumulation of APCs, which was consistent with the spleen-centered sequential activation of adaptive immunity.^8,19 Besides, HPV mRNA-LNP vaccination induced a burst of naive and memory B cells in the blood, indicating the activation and subsequent migration of B cells from secondary lymphoid organs to distal activation sites via the systemic circulation.²⁰ Meanwhile, a rapid increase of T cells, especially those highly expressing interferon (IFN)-stimulated genes (ISGs), were identified in TDLNs following the vaccination, suggesting of early T cell priming by APCs which contributed to a quick immune response.^21,22 Moreover, the infiltration of effector memory and exhausted CD8⁺ T cells in TME experienced a remarkable enhancement following the vaccination, implicating the successful establishment of anti-tumor immunity.^23,24

Collectively, our study revealed evidence for a sophisticated but coordinated remodeling pattern of systemic immune landscape corresponding to vaccination-mediated anti-tumor efficacy.

ISG signature elicited by LNP compositions contributes to non-antigen-specific quick immune responses

Type I interferons (IFN-I) mediate numerous immune interactions during tumors and viral infections, from the establishment of a defending state to the regulation of innate and adaptive immunity.^25,26 In this study, we unexpectedly identified that ISG signature was not only confined to previously reported CD8⁺ T cells and CD4⁺ T cells,^22,27 but also expressed in many other lymphoid subsets, thus might represent a unique but extensive functional state induced by HPV mRNA-LNP vaccination. Therefore, we comprehensively characterized all the immune cells with ISG signature and identified a total of 12 subclusters including CD8⁺ T cells, CD4⁺ T cells, γδ T cells, and B cells (Figures 2A; S7A and S7B; Table S7). Interestingly, proportions of all these ISG⁺ subclusters experienced a dramatic increase in TDLNs following the vaccination, indicating that the generation of ISG signature was solely induced in TDLNs, which were the primary sites at which anti-tumor lymphocytes are activated.²⁸ To investigate whether these subclusters underwent distinct differentiation trajectories following HPV mRNA-LNP vaccination, we traced the transcriptional alterations corresponding to their functional commitment. We found that all of these subclusters underwent similar differentiation trajectories, which originated from the naive clusters and directly progressed to ISG⁺ clusters, indicating that the differentiation program of ISG⁺ clusters was quick, unique but extensive for immune cells following HPV mRNA-LNP vaccination (Figures 2C–2E; S7C). To further understand that ISG signature was elicited by which part of HPV mRNA-LNPs, we examined the expression levels of ISGs in these immune subsets elicited by different treatment (Figures 2F–2H; S7D). The results showed that empty LNP vector alone could generate similar potent expression of ISGs in all the immune subsets as HPV mRNA-LNPs, indicating that ISG signature was associated with a non-antigen-specific immune response elicited by LNP compositions. As expected, Gene Ontology (GO) analysis of ISG⁺ clusters revealed functional enrichments in anti-viral immunity, response to type I interferon signaling and positive regulation of innate immunity (Figure 2I).

Figure 2.

ISG signature elicited by LNP compositions contributes to non-antigen-specific quick immune responses

(A) The t-SNE projection of ISG⁺ cells, showing the formation of 12 subclusters. Each dot represents an individual cell, colored according to cell cluster number. (B) Line graph showing changes in the proportions of total ISG⁺ cells, ISG⁺ B cells, ISG⁺CD4⁺ T cells and ISG⁺CD8⁺ T cells across different tissues in vaccine-treated or untreated TB mice. (C) Differentiation trajectories for ISG⁺ B cells inferred using pseudotime and projected onto t-SNE maps. (D) Differentiation trajectories for ISG⁺CD8⁺ cells inferred using pseudotime and projected onto t-SNE maps. (E) Differentiation trajectories for ISG⁺CD4⁺ cells inferred using pseudotime and projected onto t-SNE maps. (F) Gene expression levels of Isg15 and Ifi208 in B cells isolated from TDLN of TB mice across the PBS, vector, and vaccine groups (n = 7–9/group). Statistics were assessed using Tukey’s multiple comparison tests. Error bar = mean ± SEM. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. (G) Gene expression levels of Isg15 and Ifi208 in CD8⁺ cells isolated from TDLN of TB mice across the PBS, vector, and vaccine groups (n = 7–9/group). Statistics were assessed using Tukey’s multiple comparison tests. Error bar = mean ± SEM. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. (H) Gene expression levels of Isg15 and Ifi208 in CD4⁺ cells isolated from TDLN of TB mice across the PBS, vector, and vaccine groups (n = 7–9/group). Statistics were assessed using Tukey’s multiple comparison tests. Error bar = mean ± SEM. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. (I) Enrichment heatmap of GO pathways using specific dropout genes from each ISG⁺ cell subclusters.

Collectively, these findings suggest that LNPs were equipped with inherent adjuvant properties, which could contribute to non-antigen-specific quick immune responses by eliciting a unique but extensive ISG signature in lymphoid subsets.

Cycling signature elicited by HPV mRNA-LNP vaccination correlated with potent antigen-specific anti-tumor immunity

Previous studies have demonstrated that T/B cell proliferation is a surrogate of cell function because they need to be able to replicate to mount an effective immune response, especially in the setting of vaccination.^29,30 Nevertheless, our study revealed evidence that HPV mRNA-LNP vaccination induced the expression of a potent proliferation/cycling gene signature on almost all of the immune subsets (Figures 3A; S8A). A total of seven cycling subclusters were identified, including T cells, neutrophils, macrophages, DCs, B cells, and plasma cells (Figures 3B; S8B–S8D; Table S8). Notably, proportions of different cycling subclusters exhibited distinguished changing patterns corresponding to HPV mRNA-LNP vaccination (Figure 3C). Specifically, HPV mRNA-LNP vaccination induced a proliferative burst of T cells in the blood and TME, whereas the proliferation of myeloid cells was mainly observed in the spleens. Nevertheless, following HPV mRNA-LNP vaccination, the proportion of cycling B cells seemed to experience a rapid decrease in the spleens, which might be attributed to the differentiation and migration of B cells from secondary lymphoid organs to distal activation sites upon antigen exposure.²⁰

Figure 3.

Cycling signature elicited by HPV mRNA-LNP vaccination correlated with potent antigen-specific anti-tumor immunity

(A) Cell cycle phase analysis of cycling cells. The approximate cell cycle phase was determined by scoring individual cells based on their expression levels of genes associated with the S-phase, G1 phase, and G2M transition. (B) The t-SNE projection of cycling cell, showing the formation of eight subclusters. Each dot represents an individual cell, colored according to cell cluster number. (C) Line graphs showing changes in the proportions of total cycling cells, cycling B cells, cycling T cells, and cycling myeloid across different tissues in vaccine-treated or untreated TB mice. (D) Immunofluorescence analysis of representative spleen and tumor tissues showing CD45 and Ki67 protein expression in different treatment groups, with the percentage of Ki67⁺ cells among CD45⁺ cells (n = 3/group). Statistics were assessed using Tukey’s multiple comparison tests. Error bar = mean ± SEM. (E) Bubble plot showing GO functional enrichment for each cycling sub-cluster. (F) Flow diagram illustrating the process of adoptive transfer therapy. (G) Tumor growth curves of different recipient groups post-adoptive transfer therapy (n = 7/group).

To further understand that cycling signature was elicited by which part of HPV mRNA-LNPs, we examined the expression of cycling markers in immune cells elicited by different treatment using paraffin-embedded samples (Figure 3D). For quantifying cell infiltration, we performed a full slide scan and calculated the percentage of cycling cells in all immune cells. Multiplex immunohistochemistry (mIHC) image analysis showed that immune cells upon HPV mRNA-LNP vaccination stimulation generally exhibited a stronger proliferation profile, which were consistent with the results of our scRNA-seq analysis, indicating that this cycling signature might reflect the establishment of antigen-specific immune responses. And as expected, GO analysis of cycling clusters revealed functional enrichments in cell cycle phase transition, DNA replication and nuclear division (Figure 3E). We then sorted cycling immune cells from the vaccinated spleens and adoptively transferred them into tumor-bearing mice 2 days following tumor inoculation, which demonstrated potent anti-tumor immunity at the endpoint (Figures 3F and 3G).

Collectively, our study revealed evidence that HPV mRNA-LNP vaccination induced a characteristic proliferative burst of immune cells, which was closely associated with the establishment of anti-tumor immunity.

Cycling immune cells mediate tumor control through multi-directional differentiation into anti-tumor cell compositions

To further investigate the underlying mechanisms of the anti-tumor immunity mediated by cycling immune cells, we first projected cycling cells onto their non-proliferating reference by performing proliferation mapping (Figure S8E) to postulate their potential origins. The results showed that cycling B cells were more likely to belong to naive B cells, whereas most of cycling T cells were projected onto effector memory and exhausted CD8⁺ T cells. Meanwhile, cycling myeloid cells were projected onto a more heterogeneous group containing neutrophils, macrophages, and DCs. These findings suggest that cycling immune cells might be a heterogeneous population derived from varied subtypes of immune cells. Next, to trace the in vivo differentiation trajectories of these cycling immune cells, we sorted cycling immune cells in the spleens from vaccinated-CD45.2 TB mice and adoptively transferred them into CD45.1 TB mice 2 days following the tumor inoculation (Figure S9A). The spleens, blood, TDLNs, and tumors were harvested at 25 days following the tumor inoculation, of which CD45.2⁺ cells were sorted for subsequent scRNA-seq (Figure 4A). In general, cycling immune cells underwent a multi-directional differentiation into functional cell compositions (Figures 4B–4D; Table S9. Differentially expressed genes of post-ACT B cell clusters applied in scRNA-seq (related to Figure 4), Table S10. Differentially expressed genes of post-ACT myeloid cell clusters applied in scRNA-seq (related to Figure 4), Table S11. Differentially expressed genes of post-ACT T cell clusters applied in scRNA-seq (related to Figure 4)).

Figure 4.

Cycling immune cells mediate tumor control through multi-directional differentiation into anti-tumor cell compositions

(A) The workflow showing the change of pre-ACT immune cells and post-ACT immune cells. ACT: adoptive transfer therapy. (B) The t-SNE projection of post-ACT myeloid cell, showing the formation of 12 subclusters. Each dot represents an individual cell, colored according to cell cluster number. (C) The t-SNE projection of post-ACT B cell, showing the formation of four subclusters. Each dot represents an individual cell, colored according to cell cluster number. (D)The t-SNE projection of post-ACT NK/T cell, showing the formation of four subclusters. Each dot represents an individual cell, colored according to cell cluster number. (E) Inferred developmental trajectory of post-ACT myeloid cell by RNA velocity. Cells are colored according to their cluster origins. (F) Inferred developmental trajectory of post-ACT B cell by RNA velocity. Cells are colored according to their cluster origins. (G) Inferred developmental trajectory of post-ACT NK/T cell by RNA velocity. Cells are colored according to their cluster origins. (H) Histogram depicting the changes in the percentage of myeloid cells among CD45⁺ cells before and after ACT. (I) Histogram depicting the changes in the percentage of B cells among CD45⁺ cells before and after ACT. (J) Histogram depicting the changes in the percentage of NK/T cells among CD45⁺ cells before and after ACT. (K) Galaxy diagram showing the differentially regulated genes in pre-ACT and post-ACT myeloid cells across different tissues, categorized per known or predicted function(s), literature, and sequence similarity. Circle size is proportional to the number of differentially expressed genes (significance <0.05). (L) Galaxy diagram showing the differentially regulated genes in pre-ACT and post-ACT B cells across different tissues, categorized per known or predicted function(s), literature, and sequence similarity. Circle size is proportional to the number of differentially expressed genes (significance <0.05). (M) Galaxy diagram showing the differentially regulated genes in pre-ACT and post-ACT NK/T cells across different tissues, categorized per known or predicted function(s), literature, and sequence similarity. Circle size is proportional to the number of differentially expressed genes (significance <0.05).

As shown in Figure 4E, the majority of pre-adoptive cell transfer (ACT) cycling myeloid cells progressed through Ly6c2⁺ monocytes, early activated macrophages (Fabp4⁺ macrophages and C1qb⁺ macrophages), and ended in terminally differentiated macrophages (Spp1⁺ macrophages and Foxp1⁺ macrophages) and DCs (Fscn1⁺ DCs and Clec9a⁺ DCs). These differentiation trajectories have shown distinct tissue preference. After ACT, cycling myeloid cells rapidly entered into the blood and the secondary lymphoid organs, as well as acquired a naive monocyte signature and an early activated macrophage signature, whereas they became more terminally differentiated after infiltrating into the TME (Figure 4H). Meanwhile, functional analysis showed that pre-ACT cycling myeloid cells were barely characterized by the enrichment in pathways related to cell proliferation, including cell cycle and DNA repair, while post-ACT myeloid cells from the TDLNs were mainly enriched in pathways related to antigen processing and presentation. And as expected, post-ACT myeloid cells from the TME have shown versatile functional enrichments with anti-tumor potentials, including the activation of TNF signaling pathways, NF-κB signaling pathways and chemokine signaling pathways. Moreover, once entering the circulation, pre-ACT cycling B cells rapidly migrated into the spleens and TDLNs and generated the germinal center (GC) response (Figures 4F and I), suggesting of the formation of a specialized microenvironment within the B cell follicles of secondary lymphoid tissues upon immunization.¹³ Consistently, post-ACT B cells in secondary lymphoid organs were enriched for functional pathways related to antigen processing and presentation and B cell receptor signaling pathways. Besides, FcγR-mediated phagocytosis process was activated in the TME, indicating that antigen-specific humoral immunity might play an important part in the vaccination-mediated anti-tumor effects (Figure 4L). Similarly, pre-ACT cycling NK/T cells underwent a proliferative burst as well as effector-biased differentiation once entering into the secondary lymphoid organs through the circulation (Figures 4G and J). Although we observed a relatively low degree of NK/T cell infiltration in the TME, which might be attributed to the fierce counterattack of tumor cells, continuous supply of fresh troops from peripheral immune system could make up for the sustained anti-tumor immunity. Consistently, post-ACT NK/T cells in the TME were enriched for functional pathways related to NK cell-mediated cytotoxicity (Figure 4M).

Collectively, our study revealed evidence that cycling immune cells mediate tumor control through a systemic coordination of multi-directional differentiation into anti-tumor cell compositions.

Discussion

Our study provides key insights into the systemic immune responses orchestrated by HPV mRNA-LNP vaccination, revealing a complex yet coordinated remodeling of the immune landscape that drives effective anti-tumor immunity. Importantly, these findings not only extend our understanding of mRNA vaccines in the context of HPV-positive head and neck squamous cell carcinoma but also establish a potential framework for developing mRNA-LNP vaccines for other cancer types, particularly those related to viral oncogenesis, such as cervical or head and neck squamous cell carcinomas.³¹ Given the success of mRNA vaccines in recent human clinical trials for infectious diseases, there is substantial potential to translate these findings into clinical settings for cancer patients, where immune modulation by mRNA vaccines could be explored as an adjunct or alternative to existing therapies. Future clinical trials will be necessary to assess how these vaccines perform in diverse patient populations, particularly in cases of advanced or recurrent cancers where immune evasion mechanisms are prominent.

The extensive ISG signature observed across multiple immune subsets, including CD8⁺ T cells, CD4⁺ T cells, γδ T cells, and B cells, is particularly striking and parallels findings from recent human mRNA vaccine studies, where robust ISG expression was linked to potent immune activation. This broad ISG expression underscores the inherent adjuvant properties of LNPs, suggesting that mRNA-LNP platforms may induce dual immune responses: antigen-specific immunity coupled with non-antigen-specific immune potentiation.^25,26 These dual mechanisms provide the foundation for utilizing mRNA vaccines not only for tumor immunotherapy but also for other diseases requiring enhanced immune responses, such as chronic infections or autoimmune diseases. Moreover, the observation that ISG expression was induced by the empty LNP vector alone highlights the adjuvant properties of LNPs, consistent with recent studies demonstrating that LNP components can activate PRRs, such as TLRs, independent of the mRNA antigen.^32,33 Notably, in our study, this ISG signature was mainly detected in TDLNs, which was consistent with previous reports that subcapsular sinus macrophages from TDLNs have a major role in the secretion of IFN-I and restriction of viral spread as well as tumor progression.^21,25,26,34 This finding reinforced the notion that LNPs can contribute to non-antigen-specific immune activation, enhancing the overall efficacy of the vaccine. Nevertheless, detailed immunological effects underlying different ISG⁺ immune subsets remained to be further investigated.

Another key finding of our study is the proliferative burst of immune cells, particularly cycling T cells, B cells, and myeloid cells, following HPV mRNA-LNP vaccination. The importance of immune cell proliferation in generating effective anti-tumor responses has been well documented in both preclinical and clinical studies of nucleic acid vaccines.^{1,12,35,36,37,38,39,40} However, most of these studies focused on the clonal expansion of direct effector cells like T cells and B cells, with few studies demonstrating the roles of other cycling clusters.^{41,42,43,44,45,46,47} Our study revealed evidence that not only T/B cells could undergo a proliferative burst following the vaccination, but also myeloid cells formed a substantial part of cycling cells, which were equipped with the capacity of multi-directional differentiation once entering into the circulation and fulfill their anti-tumor potential at systemic level. This observation highlights that cycling signature in immune cells is essential for sustained immune surveillance and tumor control. Meanwhile, we also identified distinct tissue-specific differentiation trajectories for cycling myeloid cells, with the activation of antigen-processing pathways in myeloid cells within secondary lymphoid organs and the activation of TNF signaling pathways, NF-κB signaling pathways, and chemokine signaling pathways within the TME. These findings underscore the role of these cycling myeloid cells in linking innate and adaptive immune responses.

In addition to the myeloid compartment, B cells were also shown to play a critical role in the vaccination-mediated anti-tumor response. Our data revealed that B cells rapidly migrated to secondary lymphoid organs and subsequently contributed to antigen presentation in the TME, a finding that is consistent with studies highlighting the role of B cell-mediated humoral immunity in tumor regression.^48,49 Meanwhile, the activation of FcγR-mediated phagocytosis and B cell receptor signaling pathways in the TME further supported the involvement of B cells in orchestrating both humoral and cellular immunity following mRNA-LNP vaccination.^50,51

Notably, while our study employed MC3-based LNPs, a formulation validated for preclinical mRNA delivery, the mechanistic insights into systemic immune activation likely extend to clinically approved LNP platforms such as SM-102 (Moderna) and ALC-0315 (Pfizer-BioNTech). These LNPs share critical design principles with MC3, including ionizable lipid architectures that enable endosomal escape and TLR4/STING-mediated innate immune activation.^52,53,54 The ISG⁺ lymphoid subsets we identified in TDLNs mirror the IFN-I-driven immune priming observed in human trials of SM-102/ALC-0315-based COVID-19 vaccines,^55,56 underscoring the translational relevance of our findings. Furthermore, the proliferative burst of myeloid and lymphoid cells aligns with clinical reports of LNP-enhanced GC reactions and antigen-presenting cell trafficking.^53,57 Future studies could explore integrating our HPV antigen design with clinical-grade LNPs to optimize tumor-specific immunity while leveraging their established safety profiles.

Collectively, our results provide strong evidence that HPV mRNA-LNP vaccination orchestrates a systemic immune response characterized by ISG expression, immune cell proliferation, and tissue-specific differentiation. These findings expand the mechanistic understanding of mRNA-LNP vaccines and highlight the potential of optimizing LNP compositions to enhance both innate and adaptive immune activation. Future studies should explore the therapeutic potential of cycling immune cells, particularly in the context of ACT, to further harness their ability to generate potent and durable anti-tumor immunity. Our study also underscores the need for continued investigation into how different immune compartments contribute to long-term tumor control, with the goal of improving the efficacy of mRNA-LNP vaccines in cancer immunotherapy.

Conclusion

In summary, our findings highlight the complex immune dynamics elicited by HPV mRNA-LNP vaccination and its potential applications in cancer immunotherapy. These results underscore the importance of targeting specific immune cell subsets to enhance vaccine efficacy and suggest avenues for future research, particularly in exploring the therapeutic potential of mRNA vaccines in other cancer types and settings.

Materials and methods

Sex as a biological variable

This study used male C57BL/6 mice to minimize variability from the estrous cycle, maintain consistency with established HPV⁺ HNSCC models, and reflect the higher incidence of HPV⁺ HNSCC in men.

Animals

Male C57Bl/6J mice (6–8 weeks old) were purchased from Byrness Weil biotech Ltd (Chongqing, China), while CD45.1⁺ congenic mice (B6.SJL-PtprcaPepcb/BoyJ) were obtained from Shanghai Model Organisms Center, Inc. Mice were housed under specific pathogen-free conditions with ad libitum food and water. At designated time points, retro-orbital blood sampling was performed under isoflurane anesthesia (4% induction, 2% maintenance) with continuous oxygen flow (500 mL/min). Following blood collection, the animals were humanely euthanized via cervical dislocation while still under deep isoflurane anesthesia to minimize distress and ensure a painless procedure. Subsequently, spleens, tumors, and lymph nodes were harvested for further analysis. All animal experiments followed ethical guidelines set by the Animal Ethics Committee of West China Hospital (approval number: 20220511001).

Cell line and tumor implantation

The mEERL cell line, derived from the oropharyngeal epithelium of a C57BL/6J mouse and expressing HPV-16 E6, E7, and H-Ras, was cultured in Prigrow IV medium (cat. no. TM004) at 37°C with 5% CO₂ in a humidified incubator as previously reported. Mycoplasma contamination was routinely tested and found negative. For tumor implantation, 2 × 10⁶ mEERL cells were injected subcutaneously into the right flank of mice. Tumor size was monitored every 2–3 days using vernier calipers, and tumor volume was calculated using the formula: volume = π/6 × length × width.² Mice were euthanized when tumors exceeded 1,000 mm³ or ulcerated.

Vaccination

The HPV mRNA-LNP vaccine was generated and validated as previously described.¹⁸ On days 9, 14, and 19 post implantation, tumor-bearing C57Bl/6J mice (tumor size: 50–100 mm³) were randomized into treatment and control groups and intravenously administered PBS, LNP vector, or HPV mRNA-LNP (10 μg in 100 μL).

Cytokine measurement

Cytokine levels were measured using the Bio-Plex Pro Mouse Cytokine Group I Panel 23-plex kit (Bio-Rad) as per the manufacturer’s protocol by Wayen Biotechnologies (Shanghai, China). Samples were analyzed on the Bio-Plex MagPix System, and data were processed with Bio-Plex Manager software (v6.1, Luminex). The measured cytokines included Eotaxin, G-CSF, GM-CSF, IFN-γ, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12(p40), IL-12(p70), IL-13, IL-17, KC, MCP1, MIP-α, MIP-1β, RANTES, and TNF-α.

Tissue collection and single-cell suspension preparation

Fresh samples from whole blood, spleen, lymph node, and tumor were collected from tumor-bearing and non-tumor-bearing mice on day 25. In addition to blood samples, each tissue was fragmented into 1-mm cubic pieces using surgical scissors, followed by enzymatic digestion using dissociation media, which contained 0.5 mg/mL collagenase type I (Gibco, 17100017), collagenase type II (Gibco, 17104019), collagenase type IV (Gibco, 17101015), and 0.1 mg/mL DNase I (Sigma-Aldrich, 11284932001). Then tissues were incubated on a rocker (160 RPM) for 15–40 min at 37°C. The digested tissues were then passed through a 70-μm cell strainer to remove residual cell aggregates, centrifuged at 500 × g for 5 min. The flow-through was resuspended in red blood cell lysis buffer for 5–10 min, and then centrifuged and resuspended in PBS prior to surface staining. Immune cells were stained at 1 × 10⁶ cells per mL with antibodies (CD45, clone 30-F11, BioLegend, 103130; CD45.2, clone 104, BioLegend, 109805) for 30 min at 4°C, and then washed and suspended in 0.04% bovine serum albumin (BSA). 7-AAD viability staining solution (BioLegend, 420404) was added 5 min before flow cytometry sorting for dead cell discrimination. CD45⁺ cells were sorted with FACS Aria II Cell Sorter (BD Biosciences); post-sort purity was routinely >95% for the sorted populations. All subsequent steps were performed on ice or at 4°C unless specified otherwise.

Peripheral blood mononuclear cells (PBMCs) were isolated from whole blood by density gradient centrifugation in Ficoll-PaqueTM PLUS (Cytiva, 17144002) for 20 min at 800 × g without breaks. PBMCs from the interface between the plasma and the Ficoll-Paque gradient were collected and washed in ice-cold PBS, followed by centrifugation at 400 × g for 5 min. The pellet was resuspended in 3–5 mL of red blood cell lysis buffer for 5 min. The method to sort CD45⁺ immune cells from PBMCs is the same as that described above for isolating from tissues.

Adoptive transfer experiments

Male C57BL/6J (CD45.2⁺) mice, aged 6 to 8 weeks, were subcutaneously implanted with 2 × 10⁶ mEERL cells on day 0. Immune cycling cells (CD45.2⁺UBE2C⁺) were isolated from spleens on day 25. After sorting, cells were centrifuged and resuspended in PBS. A total of 1 × 10⁵ cells were transferred via retroorbital intravenous injection into 6- to 8-week-old male C57BL/6 (CD45.2⁺) or CD45.1⁺ congenic mice, which had been subcutaneously implanted with 2 × 10⁶ mEERL cells 2 days prior. Tumor size was monitored every 2–3 days post-transfer. CD45.2⁺ cells from blood, spleens, lymph nodes, and tumors were isolated on day 25 for further analysis.

Flow cytometry

Cells were stained with fluorophore-conjugated antibodies at 4°C for 30 min, excluding dead cells with Zombie NIR Fixable Viability Dye (1:1000, BioLegend, 423105). After washing and blocking with anti-CD16/32 (BD Pharmingen, 553142), cells were resuspended in PBS and analyzed with a BD FACSAria SORP Flow Cytometer. The following antibodies used were included: CD45 (BUV395, clone 30-F11, BD Biosciences; PerCP, clone 30-F11, BioLegend), CD3 (PE, clone 17A2, BioLegend), CD19 (APC, clone 6D5, BioLegend), CD4 (APC/Fire 750, clone GK1.5, BioLegend), CD8a (BV510, clone 53-6.7, BioLegend), CD45.2 (FITC, clone 104, BioLegend), and UBE2C (AF647, Santa Cruz Biotechnology). Data were acquired using the BD FACSAria SORP system and the Cytek Aurora CS, then analyzed with FlowJo v10.8.1 software. Compensation and gating were performed using unstained, single-stained, and fluorescence-minus-one (FMO) controls for accurate analysis.

Quantitative polymerase chain reaction for expression analysis

Total RNA was extracted from CD4⁺ T cells, CD8⁺ T cells, and CD19⁺ B cells using TRIzol reagent (Life Sciences, 15596026) according to the manufacturer’s protocol. Cells were lysed in 1 mL TRIzol, incubated for 5 min, and RNA was isolated by chloroform extraction and isopropanol precipitation. RNA was washed with 75% ethanol, air-dried, and resuspended in RNase-free water. RNA purity and concentration were assessed by OD260/OD280, with ratios between 1.7 and 2.0. Reverse transcription was performed with the HiScript III RT SuperMix kit (Vazyme Biotech, R323-01), followed by quantitative PCR (qPCR) using primers for Ifi208, Isg15, and Gapdh (reference gene). Reactions were performed in a QuantStudio3 PCR system (ThermoFisher) with 5 μL master mix, 0.3 μL of each primer, 1 μL cDNA, and 3.4 μL RNase-free water in a 10 μL final volume. Cycling conditions were 95°C for 30 s, followed by 39 cycles of 95°C for 10 s and 60°C for 30 s. Relative gene expression was calculated using the delta-delta Ct method, with Gapdh as the internal control.

Histological examination by hematoxylin and eosin staining

Mice tissues collected from the animal studies were fixed in 10% neutral formalin at room temperature 1 h or longer, then embedded in paraffin and then the samples were cut into 6-μm thickness sections and stained with hematoxylin and eosin. Stained slides were viewed and acquired using the Olympus VS200 (Olympus).

Multiplex immunofluorescence staining

Eighteen FFPE tissue samples from spleens and tumors were sectioned at 4 μm for multiplex immunohistochemistry using the OPAL Polaris system (Akoya Biosciences). After deparaffinization and hydration, sections were stained with CD45 (1:200, CST) and Ki67 (1:200, CST) antibodies, with nuclei counterstained using spectral DAPI. Primary antibody detection was performed with Opal Polymer HRP MsRo detection reagent. The Opal 7-Color IHC kit, using Opal 520 (Ki67), Opal 690 (CD45), and DAPI, was applied per the manufacturer’s protocol. Slides were scanned with the Vectra PolarisTM system and analyzed using QuPath software (version 0.5.1).

scRNA-seq library preparation and sequencing

scRNA-seq libraries were prepared according to 10X Genomics manufacturer’s recommendations (Next GEM Single Cell 5′ GEM Kit v2, 1000244). Then scRNA-seq libraries were sequenced using NovaSeq 6000 and each cell was sequenced to an average of 30,000 reads.

scRNA-seq library pre-processing, quality-control, and cell type annotation

Raw scRNA-seq data were processed with Cell Ranger (v7.0.1) and aligned to the refdata-gex-mm10-2020-A transcriptome. Sequencing quality was assessed by evaluating valid cell barcodes, mapping rates, and Q30 base frequencies. Data were combined into a gene/barcode matrix and pre-processed with Seurat (v4.3.1), excluding clusters with low UMIs or high mitochondrial transcripts. Doublet cells were removed using scDblFinder (v1.16.0). To integrate cells from different samples, batch effect correction was performed with the Harmony algorithm (R package, v0.1.0). Clustering and dimensionality reduction were conducted with Seurat’s FindCluster and RunTSNE functions. Differentially expressed genes were identified using FindAllMarkers (min.pct = 0.25, logfc.threshold = 0.25), and immune cell types were annotated based on these DE genes and known markers.

Trajectory analysis

Differentiation trajectories among immune cell subtypes were analyzed using Monocle3 (v1.3.7) and Slingshot (v2.2.1), with the naïve-like population as the starting point. For cycling immune cell subtypes, RNA velocity was calculated using Velocyto (v0.17.16) and scvelo (v0.3.1) to derive loom files and calculate RNA velocity for each gene, which was then embedded into t-SNE space.

Functional annotation analyses

Kyoto Encyclopedia of Genes and Genomes (KEGG) and GO enrichment analyses were carried out for DEGs between two groups or target genes of cell-to-cell communication by the R package clusterProfiler (v4.10.0). Then the ClusterGVis R package (v0.1.1) was used to cluster and visualize the gene expression matrix via default parameters.

Scoring of gene expression signatures in cycling immune cells

Gene expression signatures were computed using Seurat’s AddModuleScore function. For cycling myeloid cells, the antigen presentation signature was calculated by the antigen presentation associated genes (HLA-DMA, HLA-DMB, HLA-DOA, HLA-DOB, HLA-DPA1, HLA-DPB1, HLA-DQA1, HLA-DQA2, HLA-DQB1, HLA-DQB2, HLA-DRA, HLA-DRB1, HLA-DRB5), the complement activation signature was calculated by the complement associated genes (C1QC, C1QB, C1QA, APOE, APOC1), and the angiogenesis signature was calculated by the following genes: CCND2, CCNE1, CD44, CXCR4, E2F3, EDN1, EZH2, FGF18, FGFR1, FYN, HEY1, ITGAV, JAG1, JAG2, MMP9, NOTCH1, PDGFA, PTK2, SPP1, STC1, TNFAIP6, TYMP, VAV2, VCAN, VEGFA. The phagocytosis signature was calculated by the related genes (MRC1, CD163, MERTK, C1QB), the M1 and M2 signatures were calculated by the M1 type associated genes (CCL5, CCR7, CD40, CD86, CXCL9, CXCL10, CXCL11, IDO1, IL1A, IL1B, IL6, IRF1, IRF5, KYNU) and M2 type associated genes (CCL4, CCL13, CCL18, CCL20, CCL22, CD276, CLEC7A, CTSA, CTSB, CTSC, CTSD, FN1, IL4R, IRF4, LYVE1, MMP9, MMP14, MMP19, MSR1, TGFB1, TGFB2, TGFB3, TNFSF8, TNFSF12, VEGFA, VEGFB, VEGFC), the maturation signature was calculated by the genes (CD40, CD80, CD86, RELB, CD83), the regulatory signature was calculated by the genes (CD274, PDCD1LG2, CD200, FAS, ALDH1A2, SOCS1, SOCS2), and the migration signature was calculated by the genes (CCR7, MYO1G, CXCL16, ADAM8, ICAM1, FSCN1, MARCKS, MARCKSL1).

For NK/T cell signatures, the properTy R package (v1.1.0) was performed to calculate and visualize the immune property score, including activation, cell cycle, co-stimulatory, cytokines, cytolytic, cytotoxic, exhausted, inhibitory, and naive/memory function score in each subtype.

Statistical analysis

Statistical analyses and graph generation were performed using GraphPad Prism 10. A two-way ANOVA followed by Tukey’s multiple comparisons test was used for statistical evaluations. Error bars indicated the standard error of the mean (SEM). Statistical significance was defined as p < 0.05, with significance levels indicated in the figures as follows: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Study approval

All animal experiments were performed in accordance with the guidelines of the Animal Ethics Committee of West China Hospital (approval number: 20220511001) and adhered to the ethical standards set forth by the Guide for the Care and Use of Laboratory Animals, as outlined by the China Association for Laboratory Animal Care.

Data availability

Sequence data that support the findings of this study are available via NCBI Sequence Read Archive (SRA) under the BioProject accession code: BioProject PRJNA1271238. The code and scripts that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgments

The authors thank Li Li, Fei Chen, Chunjuan Bao, and Yang Deng (Institute of Clinical Pathology, West China Hospital, Sichuan University) for their assistance with histological staining. We also express our sincere appreciation to Linqiao Tang from the Core Facilities of West China Hospital, Sichuan University, for their invaluable support and guidance. Additionally, we are grateful to Yi Zhang and Yue Li from the Research Core Facility of West China Hospital, Sichuan University, for their help with spectral imaging. The computations in this paper were supported by the High Performance Computing platform at West China Biomedical Big Data Center, West China Hospital, Sichuan University. The graphic abstract was created in BioRender, Qiu, K. (2024) BioRender.com/p66q842.

This work was supported by the National Natural Science Foundation of China (J.J.R., grant #82272777, #82471153; K.Q., grant#82501393; Y.F.R., grant#82403589; D.N.C, #82501390), the Science and Technology Department of Sichuan Province (K.Q., grant#2024NSFSC1510; D.N.C., grant#2024NSFSC1513; J.J.R., grant#2024YFHZ0335; Y.S., grant#2025ZNSFSC1530; Y.F.R., grant#2025ZNSFSC1528), Chengdu Science and Technology Bureau (J.J.R., grant#2024-YF05-00908-SN), the Postdoctoral Research Fund of West China Hospital, Sichuan University (K.Q., grant#2024HXBH111; Y.F.R., grant#2023HXBH121; Y.S., #2025HXBH041; D.N.C., #2025HXBH004), the National Key Research and Development Program of China (X.R.S., grant#2023YFC3403200); and Sichuan University Interdisciplinary Innovation Fund (K.Q., D.N.C.).

Author contributions

J.R., X.S., and K.Q. conceived the study. K.Q, X.D., M.M., Y.S., L.F., Y.L., and C.J. designed and performed all scRNA-seq and animal experiments with help from J.J. and S. Lang F.C., X.P, Y.Z., H.W., and J.L. collected and processed the tissue samples. Y.R., D.C., and X.S. analyzed all scRNA-seq with help from W.X., G.L., and S. Li. K.Q., M.M., Y.S., H.H., Y.W., H.L., M.Z, L.L., and S.W. designed and performed cell experiments. K.Q, M.M, Y.S., and Y.R. performed the biological analysis and interpretation. K.Q., M.M., Y.S., and Y.R. wrote the manuscript. J.R., X.S., Y.Z., W.X., G.L., S. Li, J.J., and S. Lang revised the manuscript with input from all authors. The authorship order was established by mutual agreement, in accordance with the relative contributions of each author to the conduct of the study and the preparation of the manuscript.

Declaration of interests

The authors have declared that no conflict of interest exists.