Hydroxychloroquine-Derived Lipid Nanoparticles as a Dual-Function Delivery Platform for Personalized mRNA Neoantigen Cancer Vaccines: Integrating Endosomal Escape Optimization, Tumor Microenvironment Reprogramming, and Autophagy Inhibition in a Single Formulation

1. Introduction

The convergence of three technological revolutions—mRNA therapeutics, AI-driven neoantigen prediction, and protein structure modeling—has brought personalized cancer vaccines from theoretical concept to clinical reality. The KEYNOTE-942 Phase IIb trial demonstrated that intismeran autogene (mRNA-4157), a personalized neoantigen vaccine encoding up to 34 tumor-specific mutations, reduced melanoma recurrence by 49% at five years when combined with pembrolizumab [1]. Parallel developments in pancreatic cancer (BNT122, 100% recurrence-free survival in responders at 18 months) [3] and triple-negative breast cancer [2] have validated the premise that patient-specific mRNA vaccines can mobilize clinically meaningful anti-tumor immunity.

Yet a paradox haunts this field: the cancers that would benefit most from personalized vaccination are precisely those where current mRNA-LNP platforms work least effectively. Our computational analysis of 8,432 TCGA Pan-Cancer Atlas specimens using the Tumor Vaccine Responsiveness Score (TVRS) framework revealed a striking inverse correlation between clinical need and predicted vaccine responsiveness [9,10]. Melanoma scored TVRS 77.6, while pancreatic adenocarcinoma (TVRS 11.7) and glioblastoma (TVRS 15.3) scored lowest—not because they lack neoantigens entirely, but because their immunosuppressive, autophagy-dependent, “cold” tumor microenvironments prevent vaccine-induced T cells from functioning [9,14].

The delivery vehicle itself compounds this problem. Current LNPs face a well-characterized endosomal escape bottleneck: fewer than 2% of mRNA molecules reach the cytoplasm for translation [8]. In this hypothesis paper, we propose a unified solution by fundamentally reconceiving the role of hydroxychloroquine (HCQ) in cancer vaccine design. Rather than administering HCQ as a separate systemic drug, we propose incorporating HCQ-derived molecules as the ionizable lipid component of the LNP itself, creating a single nanoparticle that simultaneously optimizes mRNA delivery and reprograms the tumor immune environment (Figure 1).

2. The Three Convergent Pillars

2.1. Pillar I: the Endosomal Escape Crisis in mRNA Vaccine Delivery

The clinical efficacy of any mRNA vaccine is fundamentally constrained by the fraction of delivered mRNA that reaches the cytoplasm. The journey requires LNP internalization via endocytosis, transport through early endosomes (pH ~6.5) to late endosomes (pH ~5.5), and escape before lysosomal degradation (pH ~4.5) [8]. Quantitative studies estimate that fewer than 2% of mRNA molecules reach the cytoplasm [8]. The window between productive endosomal escape and destructive lysosomal degradation is narrow, and current ionizable lipids traverse it too slowly.

HCQ, as a lysosomotropic weak base, accumulates in acidic compartments and raises their pH, slowing the transition from late endosomes to lysosomes. In the context of an HCQ-derived ionizable lipid, this pH-modulating activity is not a systemic effect requiring oral dosing of 400–800 mg/day, but a local, nanoparticle-intrinsic property operating at the subcellular level within the compartments where mRNA escape occurs (Figure 2).

2.2. Pillar II: Tumor Microenvironment Reprogramming via M2→M1 Macrophage Repolarization

The failure of cancer vaccines in “cold” tumors is not primarily a failure of antigen presentation but of effector function. Even when neoantigen-specific T cells are generated, they encounter an immunosuppressive microenvironment dominated by M2-polarized tumor-associated macrophages (TAMs) [9]. A 2025 study demonstrated that LNPs formulated with HCQ-derived ionizable lipids achieve simultaneous spleen-tropic mRNA delivery and repolarization of M2 TAMs to M1 phenotype [5]. The mechanism involves HCQ-mediated elevation of lysosomal pH in macrophages, triggering calcium release via mucolipin-1 (TRPML1), activating p38 MAPK and NF-κB [5].

2.3. Pillar III: Autophagy Inhibition as a Vaccine Sensitizer

Autophagy is exploited by cancer cells as a survival mechanism under immune attack, particularly in KRAS-mutant tumors: pancreatic adenocarcinoma (KRAS mutated in 92%) and NSCLC (KRAS 33%) are among the most autophagy-dependent cancers [14]. Multiple clinical trials tested systemic HCQ with chemotherapy for pancreatic cancer with mixed results, likely due to insufficient intratumoral concentrations [26,27,28,29]. HCQ-LNPs solve this pharmacokinetic problem by delivering HCQ directly to the tumor microenvironment at concentrations orders of magnitude higher than achievable by oral dosing [5,6].

3. the Unified Hypothesis: HCQ-LNP Neoantigen Platform

We propose that a personalized mRNA neoantigen vaccine delivered via HCQ-derived ionizable lipid nanoparticles constitutes a quadruple-function therapeutic platform (Figure 1):

Function 1—Enhanced mRNA Delivery: HCQ-derived ionizable lipids modulate endosomal pH maturation, increasing mRNA escape to the cytoplasm and amplifying neoantigen presentation on MHC-I/II.

Function 2—TME Reprogramming: HCQ-LNPs reprogram M2 TAMs to M1 phenotype via lysosomal Ca²⁺/mucolipin-1/p38/NF-κB, converting immunosuppressive TME to immune-permissive.

Function 3—Autophagy Inhibition: Intracellular HCQ-lipid blocks autophagosome-lysosome fusion, disabling the cancer cell’s survival mechanism against immune attack.

Function 4—Anti-inflammatory Tolerance: HCQ-functionalized LNPs suppress proinflammatory cytokine storms, enabling multi-dose schedules (9 doses over 24 weeks) [6].

The unifying principle is that HCQ’s lysosomal pH-modulating activity, deployed at the nanoparticle level rather than systemically, simultaneously addresses delivery, microenvironment, and resistance barriers. This is a single molecular platform where every function emerges from lysosomal alkalinization at different biological scales.

4. Mechanistic Bridge from TB to Cancer: the pH Optimization Paradigm

This hypothesis represents a direct intellectual extension of our prior work on AI-guided phagolysosomal pH optimization for pyrazinamide-hydroxychloroquine synergy in tuberculosis (AIPH-TB framework) [13]. In that computational model, we demonstrated that HCQ’s ability to modulate phagolysosomal pH creates a tunable “pH window” optimizable for pH-dependent drug activation within macrophage phagolysosomes. The mechanistic parallel to cancer vaccine delivery is striking (Figure 3, Table 1).

5. AI-Driven Target Selection: Which Cold Tumors Benefit Most?

Our TVRS framework identifies tumors where the gap between conventional LNP vaccine responsiveness and HCQ-LNP potential is greatest. The pattern is clear: cancers with the lowest TVRS scores are characterized by high autophagy dependence, abundant M2 TAMs, and cold immune microenvironments—precisely the features HCQ-LNPs overcome (Figure 4, Table 2).

Pancreatic adenocarcinoma emerges as the ideal first target: its KRAS-driven autophagy addiction, dense M2-macrophage stroma, and 12% five-year survival create both the biological rationale and clinical urgency. This represents an inversion of conventional vaccine strategy—HCQ-LNP vaccines should target the tumors where the HCQ-derived lipid provides the greatest marginal benefit.

6. Proposed Experimental Validation Framework

6.1. Phase A: In Vitro Proof-of-Concept

Synthesize HCQ-derived ionizable lipids with varying alkyl chain lengths (C10, C12, C14). Formulate into LNPs (HCQ-lipid:DSPC:cholesterol:PEG-DMG at 50:10:38.5:1.5). Screen for transfection efficiency in human dendritic cells vs. SM-102 LNPs. Measure endosomal pH kinetics, M2→M1 repolarization, and autophagy inhibition (LC3-II/I ratio, p62 accumulation).

6.2. Phase B: In Vivo Preclinical

Using KPC pancreatic cancer mouse model, test: (1) PBS control; (2) anti-PD-1 alone; (3) neoantigen mRNA in SM-102 LNP + anti-PD-1; (4) neoantigen mRNA in HCQ-LNP + anti-PD-1; (5) empty HCQ-LNP + anti-PD-1; (6) oral HCQ + SM-102 vaccine + anti-PD-1. Primary endpoints: tumor volume, survival, intratumoral CD8+ T-cells. The critical comparison (arm 3 vs. 4) tests whether HCQ-LNP delivery demonstrates superior tumor control.

6.3. Phase C: AI-Optimized Neoantigen Selection

For clinical-grade vaccine: (1) tumor/normal WES with Mutect2/Strelka2; (2) NetMHCpan 4.1 + MHCflurry + DeepImmuno prediction; (3) Saudi HLA-optimized SNIP ensemble ranking; (4) AlphaFold 3 structural validation; (5) CodonBERT + LinearDesign mRNA optimization; (6) TVRS-based patient stratification [11,12,13].

7. Clinical Translation Pathway

Track 1 (Near-term): Phase I/II of personalized mRNA neoantigen vaccine in conventional SM-102 LNPs + pembrolizumab for advanced melanoma (Protocol FRHC-NEO-MEL-2026-001). Establishes clinical infrastructure and provides control data.

Track 2 (Medium-term): Following preclinical HCQ-LNP validation, Phase I trial specifically targeting autophagy-dependent cold tumors (pancreatic, glioblastoma) to test the central hypothesis that HCQ-LNP delivery overcomes immune resistance.

8. Potential Limitations and Safety Considerations

Several potential challenges must be acknowledged. First, the optimal HCQ-lipid concentration that balances endosomal escape enhancement against excessive lysosomal disruption is unknown and will require systematic titration. Excessive lysosomal damage could trigger galectin-mediated innate immune activation, paradoxically increasing inflammatory side effects. Second, the long-term safety of repeated intramuscular injection of HCQ-derived lipids is uncharacterized; while acute toxicity data from published studies are reassuring, chronic retinal and cardiac monitoring will be essential in clinical trials. Third, M2→M1 repolarization has been demonstrated only in mouse models; human TAMs may respond differently due to species-specific differences in macrophage biology. Fourth, autophagy inhibition by HCQ-LNPs may have unintended consequences on antigen cross-presentation by dendritic cells, which itself relies partially on autophagic pathways for MHC-II loading.

8.1. Critical Tumor-Type-Specific Consideration: The Zinc Ionophore Paradox

A particularly important safety consideration concerns HCQ’s well-characterized function as a zinc ionophore. HCQ directs zinc ions to lysosomes and increases intracellular free zinc concentrations. While zinc-mediated apoptosis amplification has been proposed as a therapeutic strategy in some cancers, and the combination of zinc with chloroquine enhances cytotoxicity in ovarian cancer models, recent evidence reveals a starkly different picture in triple-negative breast cancer (TNBC).

Specifically, brain-metastatic TNBC cells (MDA-MB-BrM2) contain double the zinc concentration of parental MDA-MB-231 cells. Elevated intracellular zinc in TNBC promotes: (a) upregulation of SerpinB2, which facilitates brain and lung metastasis by blocking niche defenses against cancer cell invasion; (b) enhanced tumorigenic capacity of breast cancer stem cells; and (c) activation of zinc-dependent matrix metalloproteinases (MMP-9, MMP-14) that drive epithelial-mesenchymal transition (EMT) and cancer cell motility. Furthermore, the zinc transporter ZIP6, transcriptionally activated by STAT3, promotes Snail-mediated E-cadherin repression, directly facilitating cell dissociation and metastatic spread. The zinc transporter LIV1 (SLC39A6) is overexpressed in TNBC and has been linked to EMT and early lymph node metastasis.

This tumor-type-specific duality of zinc biology represents a critical safety consideration for the HCQ-LNP platform. In KRAS-mutant pancreatic adenocarcinoma, where the risk-benefit profile favors autophagy inhibition and where zinc-mediated apoptosis may be beneficial, HCQ’s zinc ionophore activity is likely advantageous. However, in TNBC, the same zinc ionophore activity could paradoxically promote metastatic progression by fueling zinc-dependent pro-tumorigenic pathways.

We therefore propose that clinical application of HCQ-LNP neoantigen vaccines must be preceded by cancer-type-specific zinc profiling, including assessment of ZIP/ZnT transporter expression, baseline intracellular zinc levels, and zinc-dependent pathway activation status. For TNBC applications specifically, future formulations may require co-encapsulation of zinc chelators (e.g., TPEN) within the HCQ-LNP to neutralize the zinc ionophore effect while preserving the pH-modulating, autophagy-inhibiting, and M2→M1 reprogramming functions. This represents a tractable engineering challenge that does not invalidate the core hypothesis but rather refines its clinical application to be tumor-type-aware.

8.2. Future Directions: COX-2 Inhibition as an Adjunct Strategy

Beyond the HCQ-LNP platform itself, systemic COX-2 inhibition (e.g., celecoxib) may provide complementary benefit as an oral adjunct to HCQ-LNP vaccination. PGE2, the dominant immunosuppressive prostanoid in the tumor microenvironment of pancreatic and colorectal cancers, suppresses dendritic cell maturation and promotes M2 macrophage polarization through EP2/EP4 receptor signaling. Celecoxib-mediated COX-2 inhibition could synergize with HCQ-LNP’s M2→M1 reprogramming by simultaneously reducing de novo PGE2 production while the nanoparticle repolarizes existing M2 TAMs. This combination—HCQ-LNP neoantigen vaccine plus celecoxib plus pembrolizumab—represents a triple-blockade strategy against immunosuppression that warrants investigation in a separate study.

9. Conclusions

We present a unifying hypothesis that transforms hydroxychloroquine from a repurposed anticancer adjuvant into the structural foundation of a next-generation mRNA cancer vaccine delivery platform. By incorporating HCQ-derived ionizable lipids into the LNP formulation, a single nanoparticle achieves four therapeutic functions—enhanced mRNA delivery, TME reprogramming, autophagy inhibition, and anti-inflammatory tolerance—all from lysosomal pH modulation.

This hypothesis is rooted in the convergence of our TVRS framework, our prior AIPH-TB investigation, and recent independent demonstrations that HCQ-derived lipids achieve spleen-tropic mRNA delivery with immunomodulatory activity [5,6]. If validated, this platform could extend personalized mRNA cancer vaccines to cold tumors—pancreatic, glioblastoma, prostate—representing the greatest unmet need in oncology.