Self‐Cooperative RNA Vaccine Mitigates Dendritic Cell–Mediated Acquired Immune Resistance to Potentiate Cell Therapy for Solid Tumors

ABSTRACT

Conventional mRNA cancer vaccines are designed to maximize antigen potency but often overlook vaccination‐induced immune resistance. In this study, we identified a negative immune regulatory mechanism, whereby mRNA vaccination induces programmed death‐ligand 1 (PD‐L1) expression in dendritic cells (DCs) through type I interferon (IFN‐I) signaling. Elevated PD‐L1 expression impairs T‐cell priming in lymph nodes through engagement of programmed death receptor 1 (PD‐1) on T lymphocytes. To address this challenge, we developed a self‐cooperative RNA vaccine (SCORV) strategy by co‐delivering antigen‐encoding RNA and small interfering RNA against PD‐L1 (siPD‐L1) within a single lipid nanoparticle (LNP). Through iterative screening of >300 ionizable lipids, we optimized a DC‐targeted LNP formulation with high RNA delivery efficiency and minimal immunotoxicity. SCORV simultaneously suppresses PD‐L1–mediated immune resistance during antigen presentation and enhances T cell priming while alleviating T cell exhaustion. Importantly, SCORV potentiates the tumor reactivity of adoptively transferred tumor‐infiltrating lymphocytes and elicits robust antitumor immunity in murine melanoma and hepatocellular carcinoma models. This work highlights a rational design principle for mRNA vaccines that self‐correct vaccination‐induced immune resistance.

A dendritic cell‐targeting LNP platform enables a self‐cooperative RNA vaccine that co‐delivers antigen mRNA and siPD‐L1 to dendritic cells. By overcoming vaccination‐induced PD‐L1‐mediated immune resistance, this strategy enhances antigen presentation, T‐cell priming, and antitumor immunity, thereby improving therapeutic efficacy against solid tumors.

1. Introduction

mRNA‐based cancer vaccines have emerged as a transformative approach for cancer treatment, with recent clinical trials demonstrating their potential to elicit robust anti‐tumor immune responses [1, 2, 3, 4, 5]. Notably, a phase II trial of a neoantigen mRNA vaccine in pancreatic cancer reported significant improvements in progression‐free survival, underscoring the clinical viability of this modality [6]. Concurrent advances in lipid nanoparticle (LNP)‐based mRNA delivery systems, including novel ionizable lipids, optimized formulations, and organ‐targeting strategies, have further enhanced mRNA stability, cellular uptake, and therapeutic efficacy [7, 8, 9, 10]. The potency of mRNA vaccines relies on their ability to be internalized by dendritic cells (DCs), translated into tumor‐associated antigens, and presented via Major Histocompatibility Complex Class I (MHC‐I) to prime antigen‐specific cytotoxic T lymphocytes (CTLs) [11]. However, an effective mRNA vaccine requires a delicate balance between sufficient DCs activation to drive T‐cell immunity while avoiding excessive systemic inflammation [12]. Current vaccination strategies prioritize innate immune activation by incorporating lipid adjuvants (e.g., TLR or STING agonists) to stimulate innate immune pathways and enhance DCs maturation [13, 14, 15]. Yet, these approaches often lead to off‐target immune activation, limiting their therapeutic window [16]. Moreover, an underappreciated challenge is vaccine‐induced immune resistance, where innate immune stimulation paradoxically upregulates inhibitory checkpoints that dampen T‐cell responses [17, 18, 19].

Effective initiation of anti‐tumor immunity requires coordinated signaling between DCs and T cells [20]. The first signal, antigen recognition via the T‐cell receptor (TCR)–MHC‐I interaction, must be complemented by a second signal, wherein costimulatory molecules (e.g., CD80/CD86) engage CD28 on T cells to promote clonal expansion and effector differentiation [21]. However, DCs also express immune‐inhibitory ligands such as PD‐L1, which binds PD‐1 on T cells to mediate suppressive signals [22, 23]. Recent studies reported that PD‐L1/PD‐1 interactions of DCs and T cells in tumor‐draining lymph nodes (LNs) drive metabolic reprogramming and functional impairment of tumor‐specific T cells [24, 25, 26]. Thus, the net outcome of vaccination hinges on the balance between co‐stimulatory and inhibitory signals in the DC‐T‐cell synapse.

In this work, we first revealed that conventional mRNA vaccines paradoxically induce PD‐L1 upregulation on DCs through type I interferon signaling during antigen presentation. This vaccine‐induced PD‐L1 expression engages PD‐1 on T cells, creating an immunosuppressive checkpoint that compromises the activation and effector functions of tumor‐specific T cells. To address this challenge, we rationally developed a self‐cooperative vaccination strategy by integrating siPD‐L1 into the mRNA vaccine, which fosters robust antigen‐specific T cell responses through modulating activation‐inhibition signals in DC‐T interactions (Figure 1a). As crucial sites for DC‐mediated T cell priming, LNs are primary destinations for mRNA vaccines to induce cellular immunity. Developing effective LNPs for RNA delivery to DCs remains empirically challenging due to complex structure‐activity relationships [27, 28]. Through an integrated in silico–experimental strategy, we engineered LNPs for precise DC targeting in LNs. We then constructed an ionizable lipid library via Michael addition. 272 lipid variants were systematically evaluated through integrated computational physicochemical property prediction and experimental transfection efficiency assessment to define key structure–activity relationships for DC transfection (Figure 1b). In vivo validation yielded the lead formulation A2B19C2‐LNP (apparent pKa ∼6) with superior delivery efficiency to DCs in LNs, enabling efficient mRNA/circRNA and siRNA co‐delivery with minimal inflammatory response. The optimized‐formulation‐based self‐cooperative RNA vaccines (SCORV) significantly enhance the tumor‐reactivity of adoptive tumor‐infiltrating lymphocytes (TILs), demonstrating potent therapeutic efficacy in orthotropic melanoma and hepatocellular carcinoma (HCC) models (Figure 1c). Overall, we established a transformative RNA cancer vaccine by mitigating acquired immune resistance of DCs, which also potentiated TILs to overcome the antigen specificity limitations.

FIGURE 1.

Self‐cooperative RNA vaccine (SCORV) eliminates mRNA vaccine‐induced immune resistance of dendritic cells (DCs) to boost tumor‐infiltrating lymphocyte (TILs) therapy. (a), SCORV co‐delivers tumor antigen‐encoding mRNA and PD‐L1‐targeting siRNA (siPD‐L1) to DCs, modulating the activation/inhibition signals in DC‐T cell interactions to potentiate antigen‐specific T cell response. (b), Integrated in silico‐experimental strategy accelerates the discovery of potent ionizable lipids. (c), SCORV boosts adoptive TILs by enhancing tumor‐reactivity, demonstrating potent efficacy against melanoma and hepatocellular carcinoma (HCC).

2. Results

2.1. Type I Interferon‐Mediated PD‐L1 Upregulation in Antigen‐Presented DCs Limits mRNA Vaccine Efficacy

Emerging evidence underscores the importance of PD‐L1 on DCs in shaping the efficacy of immune checkpoint blockade (ICB) therapy [29]. Despite its clinical relevance, the cellular and molecular mechanisms governing PD‐L1 regulation in the context of mRNA vaccines remain incompletely understood. Specifically, we developed an ovalbumin (OVA)‐encoding mRNA vaccine using DLin‐MC3‐DMA (MC3)‐formulated lipid nanoparticles (LNPs) and identified a direct correlation between antigen presentation and PD‐L1 upregulation on DC surface. First, the SIINFEKL peptide was presented on the surface of murine bone marrow‐derived dendritic cells (BMDCs) in a mRNA vaccine dose‐dependent manner (Figure 2a and Figure S1). Notably, the PD‐L1 abundance on BMDCs dramatically increased in a dose‐dependent manner, which correlated positively with antigen presentation levels (Figure 2b,c and Figure S2). In contrast, the empty LNP (lipid nanoparticle) formulation showed minimal upregulation of PD‐L1 expression on BMDCs (Figure S3). In addition, a series of activators targeting innate immune pathways, including IFN‐α, LPS (TLR4 agonist), MSA‐2 (STING agonist), and R848 (TLR7/8 agonist) also all induced dose‐dependent varying degrees of PD‐L1 upregulation on BMDCs (Figure S4). These findings suggest that the acquired PD‐L1 upregulation on DCs could be driven by antigen presentation or immune agonists‐triggered innate immune pathway activation.

FIGURE 2.

Type I interferon‐mediated PD‐L1 upregulation in antigen‐presented DCs limits mRNA vaccine efficacy. (a), Correlation between DCs antigen presentation levels and mRNA vaccine doses (n = 5). (b), Correlation between PD‐L1 expression on DCs surface and mRNA vaccine doses (n = 5). (c), Positive correlation between the PD‐L1 abundance and antigen presentation levels of DCs. (d), At different doses of mRNA vaccines, PD‐L1 abundance on antigen‐presented DCs are higher than those on non‐antigen‐presented DCs (n = 3). (e, f), Transcriptome sequencing results of BMDCs treated with mRNA vaccines (n = 3), showing upregulation of H2‐T24 (MHC‐I), Cd274 (PD‐L1), and interferon‐stimulated genes (ISGs) (e); upregulated genes in BMDCs after vaccine treatment are enriched in the type I interferon response (f). (g), Increased secretion of IFNα by BMDCs after different treatments (n = 3). (h), Expression of related proteins in WT BMDCs and Ifnar1 ^−/− BMDCs after mRNA vaccine treatment. (i, j), PD‐L1 abundance on antigen‐presented Ifnar1 ^−/− BMDCs and non‐antigen‐ presented Ifnar1 ^−/− BMDCs (n = 3). (k), IFNAR1 knockout enhances DC priming efficiency for CD8⁺ T cells in LNs (n = 3). (l, m), Schematic diagram illustrating the mechanism by which mRNA vaccines (l) and SCORV (m) activate DCs and subsequently prime CD8⁺ T cells. (n), DCs treated with constant mOVA (2 µg/mL) and increasing siPD‐L1 concentrations show progressive PD‐L1 downregulation. (o, p), BMDCs treated with SCORV promote OT‐1 cell proliferation (o) and secretion of effector cytokines (p) (n = 3). Statistical significance was determined by one‐way ANOVA with Tukey's post hoc test (panels a, b, d, g, k, n–p) or linear regression (panel c) or unpaired t‐test (panel i). P values are indicated in figures.

To elucidate the mechanisms driving PD‐L1 upregulation of mRNA vaccine treated‐DCs, we further evaluated the PD‐L1 abundance on SIINFEKL/H‐2Kb⁺ BMDCs. Notably, antigen‐presented BMDCs exhibited higher PD‐L1 abundance (up to 2.3‐fold) in contrast to non‐antigen‐presented DCs across all mRNA vaccine doses (Figure 2d and Figure S5). Moreover, transcriptomic analysis of mRNA vaccine‐treated BMDCs revealed upregulation of genes associated with MHC‐I (H2‐M2, H2‐T24), PD‐L1 (Cd274), and interferon‐stimulated genes (ISGs) (Figure 2e and Figure S6). Gene set enrichment analysis further confirmed that upregulated genes were significantly enriched in type I interferon response pathways (Figure 2f). Consistent with these findings, we observed increased interferon‐α (IFN‐α) secretion by BMDCs following mRNA vaccine treatment whereas neither empty LNPs nor GFP‐mRNA‐LNP triggered detectable IFN‐α production (Figure 2g). Together, we hypothesized that type I interferon signaling mediates PD‐L1 upregulation in DCs during antigen presentation following mRNA vaccination.

We next employed Ifnar1 ^−/− BMDCs to definitively establish the requirement of type I interferon‐STAT1 signaling for PD‐L1 upregulation. Wild‐type BMDCs exhibited robust PD‐L1 upregulation following either mRNA vaccination or IFN‐α treatment, which was associated with STAT1 phosphorylation (p‐STAT1). In stark contrast, Ifnar1 ^−/− BMDCs showed complete abrogation of both p‐STAT1 activation and PD‐L1 induction (Figure 2h). These results definitively establish that IFNAR1‐mediated STAT1 signaling is essential for PD‐L1 upregulation during antigen presentation. Genetic ablation of IFNAR1 did not compromise antigen presentation capability, yet resulted in antigen‐presented BMDCs exhibiting PD‐L1 abundance slightly lower than non‐antigen‐presented DCs (Figure 2i,j and Figure S7). mRNA vaccine treatment led to enhanced priming of cytotoxic CD8⁺ T cells in IFNAR1‐deficient mice than wild‐type mice (Figure 2k and Figure S8). These findings demonstrate that antigen‐presented DCs upregulate PD‐L1 expression through type I interferon autocrine signaling. This negative immune regulation of antigen‐presented DCs can significantly attenuate mRNA vaccine‐induced CD8⁺ T cell activation through PD‐L1/PD‐1 interaction.

Given the dual role of type I interferon in both immune activation and resistance, we reasoned that direct PD‐L1 blockade would be preferable to avoid compromising immunostimulatory effects. Based on this rationale, we developed SCORV, an innovative mRNA vaccine platform that co‐delivers antigen‐encoding mRNA with siPD‐L1 to DCs. Unlike conventional mRNA vaccines (RV) that triggered PD‐L1 upregulation through IFN‐I signaling, our SCORV platform maintained potent antigen presentation while preventing PD‐L1 induction, thereby achieving superior T cell activation (Figure 2l,m). SCORV effectively prevented PD‐L1 expression in antigen‐presented BMDCs (Figure 2n and Figure S9). Notably, siPD‐L1 did not affect the antigen presentation levels of BMDCs (Figure S10). BMDCs treated with SCORV promoted the proliferation of OT‐1 CD8⁺ T cells and increased the secretion of effector cytokines (Figure 2o,p and Figure S11). Our findings demonstrated that SCORV downregulates PD‐L1 expression on antigen‐presented DCs to alleviate checkpoint‐mediated immunosuppression in DC‐T cell interactions, thereby promoting the activation and proliferation of antigen‐specific T cells.

2.2. Iterative Screening of Ionizable Lipids Platform for DC‐Targeted mRNA Delivery

The codelivery of siPD‐L1 and tumor antigen‐encoding mRNA to DCs in LNs is critical for SCORV's ability to potentiate antitumor immunity. However, conventional RNA‐LNP formulations (e.g., MC3‐LNP), when administered subcutaneously, primarily accumulate in the liver due to nonspecific tissue distribution, thereby limiting targeted RNA delivery to LNs [28, 30]. This off‐target accumulation underscores the need for advanced LNP designs to achieve efficient lymph node‐specific DC transfection. As the pivotal functional components of LNPs, the development of ionizable lipids is essentially hindered by the lack of predictive structure‐activity relationships. Consequently, the current approach predominantly depends on empirical optimization achieved via low‐throughput synthesis and screening methods [30].

To address the challenges, through a rational synthesis and iterative screen method, we designed a targeted RNA delivery system specifically for LNs. We initially employed a high‐throughput synthetic approach to generate a diverse library of ionizable lipids through a two‐step Michael addition reaction utilizing three molecular components. We mixed diacrylates and amines in precise ratios to create adjustable‐length ionizable linkers. These linkers were then terminated with alkyl acrylates (Figure 3a). By utilizing various monomers for each component, we efficiently generated a spacer‐length adjustable library of 272 ionizable lipids (Figure 3b). For initial LNP formulation screening, we employed a standard composition of ionizable lipid: DOPE: cholesterol: DMG‐PEG₂₀₀₀ at a molar ratio of 50: 10: 38.5: 1.5, with an ionizable lipid‐to‐mRNA mass ratio of 40:1. These LNPs were loaded with Fluc‐mRNA as a reporter system to evaluate transfection efficiency.

FIGURE 3.

Iterative screening of an ionizable lipids platform for DCs‐targeted mRNA delivery. (a), Schematic illustration of the ionizable lipid synthesis reaction and resulting products (spacer length = 1). (b), monomer library used for the synthesis of ionizable lipids in Round 2. (c), Transfection efficiency of LNPs formed by ionizable lipids synthesized in Round 2. (d‐f), Principal component analysis (PCA) of mRNA‐LNP transfection efficiency as a function of TPSA, MW, logD7.4, Flexibility, pKa, and LogS: variance explained by principal components (d), PCA score plot (e), and PCA loading plot (f). (g‐i), 3D surface plots showing the relationship between the physicochemical properties of ionizable lipids and LNP transfection efficiency: TPSA and MW (g), logD7.4 and flexibility (h), pKa and LogS (i).

We next employed an iterative screening approach to systematically optimize ionizable lipid structures for mRNA‐LNP formulations. Given the critical influence of spacer length on lipid architecture, our initial screening round (Round 1) focused on evaluating a select set of monomer combinations to identify optimal spacer lengths. Ionizable lipids with varying spacer lengths were formulated into mRNA‐LNPs and screened for transfection efficiency in DC2.4 cells, revealing spacer length 1 as optimal (Figure S12).

Building on these findings, in Round 2 screening, we comprehensively evaluated the ionizable lipids featuring the optimal spacer length, systematically assessing both mRNA transfection efficiency and cytotoxicity in vitro (Figure 3c and Figure S13). To establish quantitative structure‐activity relationships (QSAR), we computationally predicted key molecular descriptors for each lipid and correlated these parameters with biological performance metrics. Principal component analysis (PCA) of this dataset identified that lipid amphiphilicity, quantified through six physicochemical parameters determinants of transfection efficiency: topological polar surface area (TPSA), molecular weight (MW), distribution coefficient at pH 7.4 (logD7.4), molecular flexibility, acid dissociation constant (pKa), and aqueous solubility (LogS) (Figure 3d–f).

To further elucidate the structure‐activity relationships, we generated 3D response surface plots mapping the six key physicochemical parameters against transfection efficiency (Figure 3g–i). These visualizations revealed how pairwise combinations of (1) TPSA and MW, (2) logD7.4 and Flexibility, and (3) pKa and LogS collectively influenced LNP performance. Through this multidimensional analysis, we identified optimal parameter ranges that maximized delivery efficiency while maintaining acceptable cytotoxicity: TPSA (118‐160), MW (849‐1091), logD7.4 (3.3‐5.3; 5.7‐6.9), Flexibility (1.6‐5.2; 10.5‐15.9), pKa (5.8‐7.9), and LogS (‐6.8 to ‐3.7) (Figure S14). By coupling structure‐function analysis with performance‐based screening, we have streamlined the development of advanced ionizable lipids for SCORV.

2.3. A Novel LNP Platform Enables Targeted Co‐Delivery of siRNA and mRNA to DCs in LNs

Guided by the established parameter ranges, we designed and synthesized six ionizable lipids (A2B18C2 to A2B23C2) through systematic structural optimization (Figure 4a,b). We initiated our LNP optimization using A2B23C2 as the model lipid, employing design of experiments (DOE) methodology to systematically screen formulation parameters. This comprehensive approach identified an optimal lipid composition with a molar ratio of 25:15:57.5:2.5 (ionizable lipid/DOPE/cholesterol/DMG‐PEG₂₀₀₀) (Figure S15). Gel retardation assays demonstrated complete mRNA encapsulation at ionizable lipid‐to‐mRNA mass ratios ≥ 5:1 (Figure S16). Functional characterization revealed that a 20:1 ratio (applied to both mRNA and siRNA) achieved optimal transfection efficiency (Figure S17).

FIGURE 4.

The optimal LNP formulation enables targeted co‐delivery of siRNA and mRNA to DCs in LNs. (a), Structure of ionizable lipids candidates from screening. (b), Predicted physicochemical parameters of ionizable lipids candidates versus optimal ranges. (c), Apparent pKa characterization via TNS assay of the ionizable lipid‐based LNPs. (d, e), Cellular uptake (d) and lysosomal escape (e) of mRNA‐LNPs (n = 3). (f), Gene silencing efficiency of LNP‐si‐ZsGreen in DC2.4‐zsGreen cells, assessed by flow cytometry (n = 3). (g), Fluorescence image to show silencing efficiency of LNP‐si‐ZsGreen (A2B19B2), scale bar = 100 µm. (h–j), In vivo luciferase expression levels (h), lymph node expression (i), and lymph node‐to‐liver expression ratio (j) at 6 h post subcutaneous injection of LNP‐mFluc (n = 3). (k), Positive expression rates of LNP‐mGFP in different cell types within mouse LNs at 6 h post subcutaneous injection (n = 3). (l), In vivo and ex vivo organ distribution images of LNP‐Cy5‐siRNA at 6 h post subcutaneous injection. (m), Uptake of LNP‐Cy5‐siRNA by different cell types in LNs at 6 h post subcutaneous injection. (n), Co‐localization of mGFP expression and Cy5‐siRNA distribution in LNs at 6 h post‐subcutaneous injection of LNP‐mGFP and LNP‐Cy5‐siRNA, scale bar = 100 µm. Statistical significance was calculated via one‐way ANOVA with Tukey's test and p values were indicated (f, i‐k, m).

The above optimized parameters were then uniformly applied to the complete series of six ionizable lipids (A2B18C2‐A2B23C2) to enable direct comparative evaluation. All resulting LNPs exhibited uniform physicochemical properties, with a hydrodynamic diameter of ∼150 nm and a narrow size distribution (PDI ≈ 0.2) (Figure S18). Cryo‐electron microscopy revealed characteristic spherical nanoparticles with well‐defined concentric lipid bilayers (Figure S19). Remarkably, despite sharing identical formulation parameters, the lipid series maintained a tunable apparent pKa range spanning 6–9 (Figure 4c). This pKa tuning directly influenced surface charge characteristics. LNPs with pKa > 7 displayed a weakly positive surface charge of approximately +10 mV, whereas those with pKa < 7 were nearly neutral (Figure S20).

All LNPs demonstrated comparable or superior mRNA transfection efficiency relative to the clinical benchmark MC3‐LNP (Figure S21). Notably, higher apparent pKa formulations (8‐9; A2B22C2 and A2B23C2), exhibited significantly enhanced cellular internalization and improved lysosomal membrane destabilization compared to formulations with lower pKa values (Figure 4d,e and Figures S22–S24). This performance advantage originates from their enhanced protonation at physiological pH conditions. Increased positive charge density boosts cellular uptake via electrostatic membrane interactions while strengthening proton sponge effects for improved endosomal escape through membrane disruption.

We next assessed the gene silencing efficiency of LNP‐si‐ZsGreen in DC2.4‐ZsGreen cells. Flow cytometry and fluorescent microscope confirmed that LNPs with A2B19C2, A2B22C2, and A2B23C2 achieved over 80% silencing efficiency (Figure 4f,g and Figures S25 and S26). This robust knockdown efficiency demonstrates the platform's versatility for both mRNA delivery and siRNA‐based gene silencing applications.

We next evaluated LNP‐mediated mRNA delivery to DCs in LNs following subcutaneous administration. Kinetic analysis revealed peak m‐Fluc expression in inguinal LNs at 6 h post‐injection (Figure S27), which was subsequently used as the optimal timepoint for evaluation. Contrary to in vitro transfection results, the A2B19C2‐LNP formulation (apparent pKa ∼6) demonstrated superior transfection efficiency in LNs compared to other candidates (Figure 4h,i). Notably, A2B19C2‐LNP exhibited dramatically improved lymph node (LN) targeting compared with MC3‐LNP, achieving a 75‐fold increase in the LN‐to‐liver luciferase expression ratio (average radiance, p/sec/cm²/sr) (Figure 4j). We further compared A2B19C2‐LNP with the clinically benchmarked ALC‐0315‐LNP under subcutaneous (SC) administration and observed comparable LN transfection efficiency and mRNA expression (Figure S28). Given that intramuscular (IM) injection represents the conventional route for clinical mRNA vaccination, we next assessed the impact of administration route by comparing SC and IM delivery of A2B19C2‐LNP, and found that SC administration resulted in superior LN accumulation (Figure S29). Further analysis confirmed significantly enhanced mRNA expression in DCs for A2B19C2‐LNP relative to MC3‐LNP (Figure 4k and Figure S30), highlighting its potential for DC‐targeted vaccine applications. This unexpected preference for lower‐pKa formulations in vivo suggests distinct delivery requirements between in vitro and in vivo systems. Collectively, these results demonstrate that both lipid formulation and administration route critically determine in vivo LN delivery performance in DC‐targeted mRNA vaccine systems.

Our evaluation of siRNA delivery efficiency using Cy5‐labeled siRNA incorporated into LNPs demonstrated that A2B19C2‐LNPs achieved 5.3‐fold greater accumulation in LNs compared to MC3‐LNPs (Figure 4l and Figure S31). More significantly, these optimized nanoparticles showed particular efficiency in DCs‐targeted delivery in LNs, showing 1.5‐fold higher siRNA uptake in DCs relative to the MC3‐LNP benchmark (Figure 4m and Figure S32). Immunofluorescence staining analysis provided visual confirmation of the superior capability of A2B19C2‐LNPs for efficient co‐delivery of both mRNA and siRNA to DC population (Figure 4n).

The biocompatibility profile of our lead formulation showed dramatic improvements over first‐generation LNPs. While MC3‐LNPs induced substantial local inflammation with 9.5% CD45⁺ immune cell infiltration and accompanying tissue necrosis, the A2B19C2‐LNPs maintained excellent tissue compatibility with minimal immune cell recruitment (<1.2% CD45⁺) and complete absence of observable tissue damage (Figures S33 and S34). These findings demonstrate that the A2B19C2‐LNP platform successfully combines enhanced LN targeting, efficient DC transfection, and improved safety profiles—critical attributes for vaccine applications.

2.4. The SCORV Platform Demonstrates Potent Preventive and Therapeutic Efficacy Against Melanoma

Building upon the optimized LNP platform for targeted co‐delivery, we next evaluated SCORV, the formulated complex of antigen‐encoding mRNA and siPD‐L1 encapsulated in A2B19C2‐LNPs, for its immunotherapeutic potential. Single‐cell RNA sequencing demonstrated that SCORV substantially altered the DCs landscape compared to mRNA vaccines (RV+siNC). The platform selectively expanded cross‐presenting cDC1s to 50.02% of total DCs (+15.3% vs RV+siNC)—the crucial XCR1⁺CLEC9A⁺ subset for tumor antigen presentation to CD8⁺ T cells. Simultaneously, SCORV reduced immunosuppressive populations including tolerogenic InfDCs (−48.2% vs RV+siNC) and interferon‐producing pDCs (−22.2% vs RV+siNC), while moderately increasing inflammatory MoDCs (+16.6% vs RV+siNC) and cDC2s (+9.8% vs RV+siNC) (Figure 5a). This reprogramming established a cDC1‐dominated niche with coordinated upregulation of DC maturation markers (CD80/86), migratory receptors (CCR7), and activation of critical immune pathways such as JAK‐STAT and NF‐κB signaling (Figure 5b–e).

FIGURE 5.

SCORV platform demonstrated potent preventive and therapeutic efficacy against melanoma (a‐e), Single‐cell transcriptomic analysis of DC subsets in LNs following the indicated treatments after SCORV immunization: t‐SNE plot of DC subsets, Proportional distribution of DC subsets (a), and expression changes of representative genes involved in DC maturation (b), migration (c), the JAK‐STAT pathway (d), and the NF‐κB pathway (e). (f), Subpopulation analysis of immune cells (CD45⁺) in LNs following the indicated treatments after SCORV immunization. (g), PD‐L1 abundance on various immune cell subpopulations in LNs following the indicated treatments (n = 3). (h), Immunofluorescence analysis of PD‐L1 abundance in LNs following the indicated treatments after SCORV immunization, Scale bar = 100 µm. (i‐k), Proportion of antigen‐presented DCs (n = 4) (i), PD‐L1 abundance on antigen‐presented DCs (n = 4) (j), and proportion of antigen‐specific CD8⁺ T cells (n = 5) (k) within lymph node after following the indicated treatments. (l), C57BL/6 mice were immunized with the indicated formulations on days 0, 7, 14, and then were challenged with s.c. flank injection of 1 × 10⁶ B16‐OVA cells on days 21, and tumor growth was measured over time. (m), Changes in tumor size of treatment groups were monitored over time (n = 6). (n), C57BL/6 mice were challenged with s.c. flank injection of 1 × 10⁶ B16‐OVA cells on days 0, and then following the indicated treatments on days 4, 9, 14, and tumor growth was measured over time. (o, p), Tumor growth curves (n = 6, o), and animal survival curves were monitored over time (n = 6, p). (q), Proportion of antigen‐specific T cells in spleens (n = 5). (r), Proportion of IFNγ⁺ CD8⁺ T cells in tumors (n = 5). (s), Proportion of regulatory T cells (Tregs) in tumors (n = 5). (t), Confocal immunofluorescence analysis of IFN‐γ‐producing CD8⁺ T cells in tumors, Scale bar = 100 µm. Statistical significance was calculated via one‐way ANOVA with Tukey's test, and p values were indicated (i‐k, o, q‐s). Statistical significance was calculated via log rank (Mantel–Cox) test and the median survival time was calculated via Kaplan‐Meier Survival Analysis (p).

Flow cytometric analysis of LN immune populations revealed SCORV's distinctive ability to precisely modulate PD‐L1 expression in a cell‐type‐specific manner. In contrast to RV+siNC that upregulated PD‐L1 of DCs, SCORV precisely restored PD‐L1 expression on DCs to physiological baseline abundance (comparable to PBS‐treated samples) while maintaining normal PD‐L1 expression patterns in other immune cell types including T cells, B cells, and macrophages (Figure 5f,g). Confocal microscopy of lymph node sections revealed that SCORV significantly reversed the PD‐L1 upregulation observed in RV+siNC‐treated DCs (Figure 5h). This targeted immunomodulation was accompanied by enhanced functional capacity of DCs. SCORV‐treated antigen‐presented DCs showed both superior antigen presentation capacity (1.5‐fold increase in MHC I‐peptide complexes compared to RV+siNC) and selective PD‐L1 downregulation, while non‐antigen‐presented DCs maintained normal PD‐L1 expression (Figure 5i,j and Figures S35 and S36). The combination of enhanced antigen presentation with precise checkpoint modulation resulted in a 1.6‐fold increase of antigen‐specific CD8⁺ T cell activation compared to RV+siNC‐treated samples (Figure 5k and Figure S37).

Given SCORV's potent immunomodulatory effects, we next investigated its prophylactic potential in melanoma models. In preventive vaccination studies, SCORV provided complete protection against B16‐OVA tumor challenge, matching the efficacy of RV+siNC while significantly surpassing combinations of commercial adjuvants with tumor antigens plus anti‐PD‐L1 antibody (OVA+CpG+aPD‐L1) (Figure 5l,m). The evaluation of SCORV's therapeutic potential against established B16‐OVA melanoma tumors revealed significant improvements over current treatment approaches (Figure 5n). SCORV demonstrated remarkable tumor growth control, reducing progression kinetics by 51.6% compared to RV+siNC and by 82.9% relative to the OVA+CpG+aPD‐L1 combination therapy (Figure 5o and Figure S38). The survival outcomes SCORV‐treated mice showing a 37.0% extension in median survival time compared to RV+siNC (37 versus 27 days) and an impressive 94.7% increase over OVA+CpG+aPD‐L1‐treated samples (37 versus 19 days) (Figure 5p).

SCORV induced potent antitumor effects through coordinated systemic immune activation and tumor microenvironment (TME) reprogramming. Comprehensive immunophenotyping revealed that compared to conventional RV+siNC, SCORV induced enhanced systemic antigen‐specific immunity, as demonstrated by a 2.4‐fold increase in splenic antigen‐specific CD8⁺ T cell activation and a 2.0‐fold increase in tumor‐reactive CD8⁺ IFNγ⁺ T cell expansion (Figure 5q,r and Figures S39 and S40). Concurrently, SCORV treatment achieved a striking 77.1% reduction in immunosuppressive Treg populations within tumors (Figure 5s and Figure S41). Immunohistochemical and immunofluorescence analyses further validated these findings, showing significantly enhanced infiltration of CD8⁺ IFNγ⁺ T cells in tumor tissues (Figure 5t). In summary, SCORV mediated synergistic immune regulation in LNs by enhancing DCs antigen presentation while selectively downregulating PD‐L1. This dual‐action immunomodulation elicited potent antitumor immunity, significantly boosting both prophylactic and therapeutic efficacy against melanoma progression.

2.5. SCORV Potentiates Tumor‐Infiltrating Lymphocytes (TILs) Efficacy Through Immune Reprogramming in Melanoma

Tumor‐infiltrating lymphocyte (TIL) therapy, while offering intrinsic tumor‐homing capacity and low immunotoxicity, is constrained by the TME's enrichment of exhausted bystander T cells over rare tumor‐reactive clones and the unavoidable amplification of non‐specific T cells during expansion, yielding products with limited therapeutic specificity [31, 32, 33]. Compounding these issues, the labor‐intensive 4–6 weeks manufacturing process creates an unresolved efficacy‐efficiency trade‐off [34].

Capitalizing on SCORV's established capacity to activate endogenous tumor‐specific T cell responses, we propose that its combination with TILs therapy could overcome current limitations by simultaneously enhancing both the tumor‐reactive frequency and functional potency of adoptively transferred TILs. This integrated therapeutic strategy synergizes the unique advantages of each approach: TILs provide naturally tumor‐targeted lymphocytes, while SCORV activates TILs by modulating DC‐T interactions in LNs.

In a murine melanoma model, the combination of SCORV with CD8⁺ TILs demonstrated superior therapeutic efficacy compared to monotherapies (Figure 6a). The combined treatment significantly inhibited tumor progression, as evidenced by 93.8% reduction in tumor growth by day 21 (vs TILs alone) (Figure 6b–f). This enhanced antitumor effect translated to meaningful survival benefits, with 16.7% of treated animals (1/6) achieving complete tumor regression and a 2.2‐fold extension in median survival time compared to TILs alone (Figure 6g). Importantly, the combination regimen showed excellent safety profiles, with no observed toxicity as measured by stable body weights throughout treatment (Figure S42).

FIGURE 6.

SCORV potentiates CD8⁺ tumor infiltrating lymphocytes (TILs) efficacy through immune reprogramming in melanoma. (a), Schematic of the combined SCORV and CD8⁺ TILs therapy in tumor treatment. (b‐g), Tumor growth curves (b), individual tumor growth trajectories per group (c‐f), and survival analysis (g) (n = 6). (h), Schematic of CD45.1/CD45.2 immune profiling in the SCORV + TILs combined therapy. (i, j), SCORV increases antigen‐specific CD8⁺ TILs proportions in blood (i) and LNs (j) (n = 5). (k), SCORV elevates CD8⁺ TILs counts in tumors. (l), SCORV increases antigen‐specific CD8⁺ TILs proportions in tumors (n = 3). (m), SCORV modulates exhaustion markers (PD‐1, TIM‐3) on CD8⁺ TILs (n = 3). (n), SCORV enhances effector functions (Granzyme B⁺) of CD8⁺ TILs (n = 3). (o), SCORV promotes tumor infiltration by DiR‐labeled CD8⁺ TILs. (p), Time‐dependent infiltration kinetics of CD8⁺ TILs into tumors (n = 3). (q), Confocal microscopy of DiR‐labeled CD8⁺ TILs in tumors of indicated treatment (scale bar = 20 µm). (r‐s), SCORV boosts CD8⁺ TILs recruitment to LNs (n = 4). Statistical significance was calculated via one‐way ANOVA with Tukey's test (b, i‐n, s) or log rank (Mantel–Cox) test (g). P values were indicated. The median survival time was calculated via Kaplan‐Meier Survival Analysis.

To specifically evaluate SCORV's ability to reprogram adoptively transferred TILs, we utilized CD45.1⁺ CD8⁺ TILs isolated from B16‐F10 tumors to treat CD45.2⁺ mice bearing B16‐OVA tumors. This experimental design enabled precise tracking of tumor‐specific TIL reprogramming. (Figure 6h). Compared to RV+siNC, SCORV demonstrated superior systemic effects by maintaining elevated frequencies of antigen‐specific CD8⁺ TILs (up to 30.3%) in peripheral blood for 21 days post‐vaccination (Figure 6i; Figure S43). The treatment induced robust expansion of antigen‐specific CD8⁺ TILs across lymphoid and tumor compartments, achieving 2.2‐fold greater antigen‐specific CD8⁺ TILs accumulation in LNs (Figure 6j and Figure S44), along with a 2.5‐fold elevation in total CD8⁺ TIL density within tumors (Figure 6k), and a 5.1‐fold increased antigen‐specific CD8⁺ TILs infiltration in tumors (Figure 6l and Figure S45). Notably, SCORV preferentially enhanced functionally superior TIL subsets, particularly the PD‐1⁺ TIM‐3⁺ CD8⁺ population (3.1‐fold increase, Figure 6m and Figure S46) and Granzyme B⁺ effector cells (2.1‐fold increase, Figure 6n and Figure S47), demonstrating its ability to qualitatively improve tumor‐reactive TILs. These findings collectively reveal that SCORV enhances TIL therapy through sustaining systemic TIL persistence, amplifying lymphoid priming, and enriching functionally enhanced TIL subsets within the TME.

The spatiotemporal dynamics of TIL trafficking revealed SCORV's critical role in orchestrating lymphocyte homing. Following SCORV treatment, DiR‐labeled CD8⁺ TILs demonstrated enhanced tumor infiltration with distinct kinetic patterns (Figure 6o,p). Confocal microscopy visually confirmed successful intratumoral localization of adoptive CD8⁺ TILs (Figure 6q), while simultaneously revealing increased recruitment to lymphoid tissues (Figure 6r,s). These findings demonstrate SCORV's profound capacity to reprogram CD8⁺ TILs into tumor‐specific effector T cells by simultaneously optimizing their in vivo distribution and enhancing functional attributes to favor antitumor immunity. The combination of SCORV with TILs represents a promising strategy to overcome current limitations of TIL therapy and significantly improve antitumor efficacy.

2.6. CircRNA‐Encoded Neoantigen SCORV Potentiates TILs Therapy in Orthotopic Liver Cancer

Clinical‐stage circular RNA (circRNA) platforms leverage their natural resistance to exonucleases for durable antigen expression [35], while tumor neoantigens avoid central tolerance to elicit potent T cell responses [36]. Building on these advantages, we evaluated SCORV's compatibility with circRNA‐encoded neoantigens (PTPN2) of hepatocellular carcinoma (HCC) (Figure 7a). The TILs+SCORV combination demonstrated robust immunogenicity, driving marked expansion of neoantigen‐specific CD8⁺ T cell populations across multiple immune compartments. TILs+SCORV exhibited a 2.0‐fold increase in neoantigen‐specific CD8⁺ T cells in draining LNs compared to TILs+RV+siNC (Figure 7b and Figure S48). Peripheral blood immune analysis demonstrated SCORV's superior ability to sustain TIL expansion, with a 3.7‐fold greater persistence of antigen‐specific T cells compared to TILs+RV+siNC at day 21, maintaining comparable frequencies to day 7 levels (Figure 7c and Figure S49). This durable maintenance confirms SCORV's enhanced capacity to support TIL proliferation over RV+siNC.

FIGURE 7.

CircRNA‐encoded neoantigen SCORV potentiates CD8⁺ TILs therapy in orthotopic liver cancer. (a), Schematic of circRNA‐based SCORV encoding Hepa1‐6 neoantigens. (b,c), SCORV elevates neoantigen‐specific CD8⁺ TILs in blood (b) and LNs (c). (d), Bioluminescence imaging of orthotopic Hepa1‐6‐Luc tumors at indicated time points (n = 5). (e‐g), Average tumor bioluminescence intensity (n = 5). (h,i), Representative liver photographs (h) and weights (i) on day 25 (n = 3). (j), H&E‐stained liver sections (scale bar = 1 mm). (k,l), SCORV modulates exhaustion markers (PD‐1, TIM‐3), (k) and enhances Granzyme B⁺ CD8⁺ TILs (l) in Hepa1‐6‐Luc tumors (n = 3). (m), SCORV modulates regulatory T cell (Treg) frequencies (n = 3). (n), Immune cell composition within Hepa1‐6‐Luc tumors (mean values, n = 3). (o), Proposed mechanism of SCORV‐enhanced CD8⁺ TILs therapy. Statistical significance was calculated via one‐way ANOVA with Tukey's test (b, c, i, k‐m); p values were indicated.

The TILs+SCORV combination demonstrated significant therapeutic efficacy in established orthotopic Hepa1‐6‐Luc tumors. In vivo bioluminescence imaging revealed progressive tumor regression, with consistent 99.7% lower signal intensity versus TILs‐only controls across all mice (5/5) on day 25 (Figure 7d–g). Critically, SCORV‐treated mice exhibited near‐normal liver weights (2.1 ± 0.3 g) compared to the tumor‐induced hepatomegaly observed in controls (PBS: 3.4 ± 0.5 g; TILs‐only: 3.1 ± 0.4 g), confirming preservation of hepatic architecture (Figure 7h,i). Histopathological evaluation confirmed treatment efficacy, demonstrating minimal residual tumor foci (< 5% of sectional area) and intact lobular architecture in TILs+SCORV recipients versus extensive neoplastic infiltration (> 60% area) in PBS controls (Figure 7j).

Comprehensive immunophenotyping demonstrated TILs+SCORV's ability to fundamentally reshape the TME. The TILs+SCORV treatment induced a significant change of tumor‐infiltrating CD8⁺ T cell subsets, reducing PD‐1⁻ TIM‐3⁻ populations by 33.1% while including a 51.7% increase in intermediate‐effector PD‐1⁺ TIM‐3⁻ subsets compared to mice receiving TILs treatment (Figure 7k and Figure S50). This phenotypic shift correlated with enhanced cytotoxic function, as evidenced by increased frequencies of Granzyme B⁺ CD8⁺ T cells (2.2‐fold) and IFN‐γ⁺ CD8⁺ T cells (3.5‐fold) compared to TILs only (Figure 7l and Figures S51 and S52).

The treatment simultaneously attenuated immunosuppressive mechanisms, achieving a 67.0% reduction in Treg frequencies within tumors (Figure 7m and Figure S53). This dual modulation of enhancing effector T cell function while relieving immune suppression effectively reversed the hallmark immunosuppression of HCC, creating a microenvironment more permissive to antitumor immunity. Detailed immune cell profiling demonstrated significant microenvironmental remodeling across treatment groups. Compared to PBS controls and TILs alone, the TILs+SCORV treatment expanded DC, B cell, CD8⁺ T cell, and CD4⁺ T cell populations while reducing macrophage infiltration, which suggests a preferential decrease in immunosuppressive macrophage subsets (Figure 7n). This coordinated shift in immune cell composition indicated successful conversion to a more immunostimulatory TME. Our findings demonstrated that SCORV mediates profound functional improvements in tumor‐infiltrating lymphocytes by enhancing their intrinsic antitumor capabilities. The treatment boosts TIL recognition of tumor‐specific antigens while simultaneously improving their ability to penetrate and persist within tumor tissues. Furthermore, SCORV reinvigorates the cytotoxic effect of previously exhausted TIL populations, restoring their capacity to effectively eliminate tumor cells. This multifaceted reprogramming results in TILs with greater tumor specificity, enhanced tissue‐trafficking capacity, and amplified effector functions, which collectively addressed the major functional limitations that have constrained adoptive TIL therapy (Figure 7o).

3. Discussion

The emergence of mRNA vaccines has fundamentally transformed the landscape of cancer immunotherapy, offering unprecedented flexibility in encoding diverse tumor‐associated antigens and neoantigens. Conventional mRNA vaccine approaches typically rely on co‐administered adjuvants to stimulate innate immune responses and enhance antigen presentation. However, this strategy suffers from two major shortcomings that our work elucidates. First, the systemic administration of potent adjuvants frequently leads to dose‐limiting toxicities, including cytokine release syndrome and organ inflammation. More importantly, we discovered that the type I interferon response induced by conventional mRNA vaccines paradoxically upregulates PD‐L1 expression on dendritic cells precisely during antigen presentation. This vaccine‐induced immune checkpoint activation creates a fundamental limitation that suppresses T cell priming and ultimately compromises antitumor efficacy. Our findings provide the first direct evidence that endogenous interferon responses to mRNA vaccination can simultaneously enhance and constrain antitumor immunity through this dual‐edged mechanism. Furthermore, direct activation of innate immune pathways by the STING agonist MSA‐2, the TLR7/8 agonist R848, or exogenous IFN‐α was sufficient to induce PD‐L1 expression, indicating that multiple routes of IFN‐I activation can converge on this regulatory pathway. This observation raises the possibility that additional components of mRNA‐LNP formulations may contribute to IFN‐I signaling. Indeed, previous studies have implicated residual dsRNA contaminants, incomplete nucleotide modification, and the intrinsic immunostimulatory properties of certain ionizable lipids as potential drivers of innate immune activation in mRNA‐LNP systems [37, 38, 39]. Although these factors were not directly investigated in the present study, they may contribute to IFN‐I production and subsequent PD‐L1 regulation in specific formulation contexts.

SCORV is an integrated solution that coordinates antigen presentation and checkpoint blockade at the cellular level. The platform's innovative design employs low‐immunogenicity LNPs to co‐encapsulate both tumor antigen mRNA and PD‐L1‐targeting siRNA, enabling efficient delivery to dendritic cells within LNs. Compared to tumor vaccines combined with ICB antibodies, SCORV offers several potential advantages: First, it achieves precise spatial‐temporal coordination of antigen expression and PD‐L1 knockdown within the same dendritic cell populations. Second, it enhances T cell priming through selective checkpoint blockade specifically in antigen‐presenting cells, avoiding systemic immunosuppression. Third, it eliminates the need for extrinsic adjuvants while maintaining potent immunogenicity through optimized ionizable lipids.

The development of SCORV required overcoming significant challenges in LN‐targeted delivery. Through iterative screening of more than 300 LNP candidates, we provided a rational framework for LNP optimization and identified optimal formulations enabling efficient co‐delivery of mRNA and siRNA to lymph node DCs. Our studies revealed a crucial paradox: while high‐pKa LNPs (8‐9) show better in vitro performance, low‐pKa LNPs (∼6) demonstrate superior in vivo targeting and efficacy. This reflects a balance between cellular uptake/endosomal escape (favored by higher pKa) and in vivo stability/lymph node accumulation (favored by lower pKa).

The therapeutic potential of SCORV extends far beyond conventional vaccine applications. Our results demonstrate its unique ability to enhance adoptive T cell therapies, particularly tumor‐infiltrating lymphocyte (TIL) treatments. Current clinical protocols require high‐dose IL‐2 administration to support TIL persistence, often causing severe capillary leak syndrome and other toxicities. SCORV addresses this limitation by providing physiological cytokine support through vaccine‐mediated DC activation, while simultaneously reprogramming the TME to be more permissive to T cell infiltration and function. Remarkably, SCORV enhanced both adoptively transferred TILs and endogenous CTLs, achieving complete tumor regression in several tumor models while maintaining an excellent safety profile. The use of circRNA neoantigen‐encoding circRNA in SCORV provided additional benefits, including prolonged antigen expression and stronger T cell responses compared to linear mRNA. This aligns with growing evidence that circRNA structure enhances stability and immunogenicity, though direct comparisons in clinical trials will be necessary.

Our findings suggest SCORV orchestrates antitumor immunity through sequential immunomodulation. The treatment first optimizes dendritic cell‐T cell interactions in LNs by coordinately enhancing antigen presentation while preventing PD‐L1‐mediated immunosuppression through targeted knockdown in activated DCs. This dual action promotes more effective T cell priming by alleviating PD‐1‐dependent inhibition during initial activation. The enhanced T cell responses subsequently reshape the TME through several coordinated effects. The treatment boosts tumor antigen recognition through expanded antigen‐specific T cell clones while improving their infiltration capacity into tumor sites. Concurrent reduction in regulatory T cell frequencies after SCORV treatment further alleviates immunosuppressive pressure. Together, these changes foster a more favorable immunological milieu for antitumor activity while maintaining immune specificity.

For clinical translation, several key considerations emerge from our findings. First, injection site selection will be crucial to maximize delivery to tumor‐draining LNs, which serve as the primary inductive sites for antitumor immunity. Second, dose optimization studies should establish the ideal mRNA/siRNA ratio and administration schedule based on the kinetic relationship between DC antigen presentation and PD‐L1 upregulation. Third, treatment strategies may need tailoring to different tumor immune phenotypes—with particular attention to overcoming the suppressive environments of immune‐excluded and immune‐desert tumors. Additionally, the discovery of vaccine‐induced PD‐L1 upregulation suggests other immune checkpoints in LNs may similarly constrain vaccine efficacy.

In conclusion, our study overcomes immune resistance of DCs in LNs mediated by mRNA vaccine, targeting PD‐L1/PD‐1 pathway between DC‐T cell interaction, thus developing SCORV. Importantly, it is crucial to investigate the impact of other immune‐negative regulatory factors within TDLNs on antitumor immunity, providing a more comprehensive understanding of the immune landscape within TDLNs. By pinpointing key factors of ionizable lipids that affect LNP transfection in vitro or vivo, we gained valuable insights for optimizing LNP development. Moreover, SCORV enhanced endogenous CTLs responses and promoted adoptive TILs functionality, demonstrating far‐reaching implications for adoptive cell therapy.

Author Contributions

L.H. performed the study and wrote the manuscript. L.H., F.C., F.Z., B.H., X.L. and H.Y. designed the study. A.L., Z.C., Z.X. and W.W. conducted the experimental studies and analyzed the data. G.M. and W.L. assisted with the experimental studies. G.M., W.L., and M.F.L. provided critical suggestions on the experimental design and manuscript. S.L., J.G., Z.X., X.L. and H.Y. revised the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma73951‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.