Lyophilizable Thermostable Nano‐Aluminum Adjuvant and Combinations Induced Robust Antigen‐Specific Humoral and Cellular Immunity in Mice

ABSTRACT

Aluminum adjuvants have been used in licensed human vaccines for almost a century. However, they can neither induce an efficient cellular immune response, nor be frozen or lyophilized due to aggregation and diminished adjuvant efficacy. This study developed a series of lyophilizable nano‐aluminum adjuvants by coating nanoscale aluminum adjuvant (Adju‐Phos) with glycol chitosan (GCS, 20 kDa) using microfluidics technology. When the GCS/alum weight ratio was higher than 4:1, the nano‐aluminum adjuvants withstood autoclaving and freeze‐drying treatments in the presence of trehalose without significant physicochemical changes. Moreover, the lyophilized nano‐aluminum adjuvants remained stable for 15 months at room temperature. When combined with HPV16 L1 virus‐like particles (VLPs), the nano‐aluminum adjuvant with 7:1 GCS/alum weight ratio (G7‐A1) elicited stronger adaptive immunity, as evidenced by higher neutralizing antibodies, balanced IgG1/IgG2c ratio, elevated IFN‐γ/TNF‐α/IL‐4/IL‐10 secretion, and increased Th1/CTL/memory T cell frequencies. Furthermore, combining G7‐A1 with MF59‐like emulsion or C‐di‐AMP enhanced antigen‐specific humoral and cellular immunity, memory T cell responses, and dendritic cell activation. These findings demonstrate the great potential of nano‐aluminum adjuvants in improving vaccine efficacy, and suggest that they may serve as alternatives suitable for vaccines requiring balanced Th1/Th2 immunity or strong cellular immunity.

Microfluidic engineering of glycol chitosan‐coated nano‐aluminum generates a lyophilizable, thermostable adjuvant platform with enhanced formulation stability. After loading HPV16 L1 VLPs, the formulation promotes dendritic‐cell activation and T‐cell priming, eliciting robust neutralizing antibodies, balanced Th1/Th2 immunity, cellular responses, and memory T cells. Combination with MF59‐like emulsion or C‐di‐AMP further strengthens antigen‐specific immunity. Created with MedPeer (www.medpeer.cn).

1. Introduction

Aluminum adjuvants have been used in many approved human vaccines, and have displayed an acceptable safety profile and great potential to enhance antigen immunogenicity in real‐world applications [1, 2]. However, conventional aluminum adjuvants exist as heterogeneous microparticle aggregates (1‐10 µm) composed of nanoscale aluminum particles [3, 4]. These bulky gels primarily elicit a Th2‐biased immune response, and therefore cannot meet the needs of vaccines requiring cellular immune responses [5, 6]. Moreover, traditional aluminum adjuvants cannot withstand freeze‐drying treatments and require strict cold‐chain storage, because freezing leads to irreversible particle aggregation and potency loss [7, 8]. To overcome the above limitations, combining aluminum adjuvants with other immunostimulants or adjusting their physicochemical properties, such as downsizing conventional aluminum adjuvants to the nanoscale or engineering their surface, have emerged as promising strategies [4, 9, 10, 11, 12, 13]. Nano‐aluminum adjuvants, in particular, have been shown to promote more efficient antigen uptake and presentation, thereby stimulating T cells and B cells to enhance adaptive immune responses [2, 14]. Additionally, nano‐aluminum adjuvants can modulate immune response bias and reduce injection‐site reactions [2]. For example, nano‐aluminum demonstrates significantly enhanced dendritic cell (DC) uptake and elevated antibody titers compared to conventional aluminum adjuvants, while concurrently reducing injection‐site inflammation after ultrasonication treatment [15]. Nevertheless, nano‐aluminum systems remain inherently unstable and prone to rapid re‐aggregation without appropriate surface stabilization. This highlights the importance of integrating size reduction of aluminum adjuvants with robust surface engineering strategies [16].

Several methods have been used to prepare nano‐aluminum adjuvants, such as ultrasonic dispersion [16], hydrothermal synthesis [17], laser ablation [18] and high‐pressure microfluidics [4]. Polyvinylpyrrolidone‐stabilized nano‐aluminum prepared by ultrasonication treatment exhibited superior adjuvant activity to micron‐sized aluminum adjuvants and maintained stability for three months at 37°C [19]. Nano‐aluminum prepared by hydrothermal synthesis exhibited superior suspension stability after antigen adsorption when compared to Alhydrogel [17]. Nano‐aluminum adjuvants produced by laser ablation showed anticancer and antipathogenic activities, but their adjuvant activity was not tested [18]. Among the above approaches, microfluidic processing offers distinct advantages from a materials‐engineering standpoint, including precise control over shear forces, mixing kinetics, and particle size distribution, as well as scalability and reproducibility. Importantly, microfluidization enables the integration of stabilizing polymers during particle formation, allowing simultaneous control of both nanoscale structure and surface functionality. However, the success of such systems critically depends on the selection of suitable stabilizers that can prevent aggregation, maintain mixture stability during processing and storage, and ideally introduce additional functional properties. For instance, microfluidization of the mixture of Alhydrogel and polymers like poly(acrylic acid) (PAA) resulted in freeze‐thaw resistant nano‐aluminum adjuvants with enhanced Th1/Th2 immunity [4].

Chitosan, a natural cationic polysaccharide, has been used in FDA‐approved wound dressing products, and exhibits excellent biocompatibility, antibacterial activity, and biodegradability. Similar to chitosan, its derivative glycol chitosan (GCS) also displays good biocompatibility, high biodegradability, and mucosal adhesion [20, 21]. From a materials‐engineering perspective, GCS is particularly attractive as a coating material because its hydrophilic ethylene glycol side chains render it soluble over a broad pH range, whereas unmodified chitosan is typically restricted to acidic aqueous conditions. This broad solubility window facilitates integration with aluminum adjuvants and is advantageous for scalable processing and formulation control. Both chitosan and GCS can be used to deliver small molecules, proteins, and genes in the form of nanoparticles or micelles [22, 23, 24, 25, 26]. Recently, several studies have demonstrated that chitosan promotes the production of type I interferons, facilitates DC maturation, and augments Th1‐mediated cellular immune responses by activating the cGAS‐STING pathway [27, 28]. In another study, GCS has been shown to enhance the durability of antigen‐specific immune response when used as an adjuvant for respiratory syncytial virus (RSV) F protein subunit vaccine [29]. In addition, low‐molecular‐weight chitosan is reported to result in more robust adjuvant activity via greater penetration into APCs [30, 31]. These biological findings are relevant and also support a viable concept: low‐molecular‐weight GCS is not only biocompatible but also a rational stabilizing and surface‐engineering component for constructing positively charged, safe, and efficient nano‐aluminum adjuvants.

Small‐molecule immunostimulants play an important role in helping traditional adjuvants enhance the induction of the desired antigen‐specific immune response. Among them, Toll‐like receptor (TLR) agonists have received increasing attention due to their clear structure‐activity relationships and acceptable safety profiles [5]. For instance, imiquimod (IMQ), as a TLR7/8 agonist, activates NF‐κB to produce type I interferons, pro‐inflammatory cytokines and chemokines [32, 33], resulting in a robust cellular immune response [34, 35]. Besides, immunostimulants targeting other pattern recognition receptors (PRRs) have also displayed potential to modulate innate and adaptive immune systems, including NOD‐like receptors (NLRs) [36], C‐type lectin receptors (CLRs) [37], retinoic acid‐inducible gene I‐like receptors (RLRs) [38], and STING ligands [39]. C‐di‐AMP (CdA), a cGAS‐STING agonist, activates innate immunity to initiate type I interferon production and elicits antigen‐specific Th1/Th2/Th17 responses [39, 40, 41, 42]. MF59, a squalene‐based oil‐in‐water emulsion adjuvant, is the first non‐alum adjuvant applied in approved human vaccines and still in use today. MF59 can create a transient immune microenvironment at injection sites through promoting various cytokine and chemokine secretion and immune cell recruitment, followed by the infiltration of antigen/adjuvant‐positive APCs to draining lymph nodes (dLNs) [43, 44, 45]. MF59 can also induce a high frequency of germinal centers (GC) and follicular helper T cells (TFH) in the lymphoid follicles [46, 47], resulting in Th1/Th2 type responses [6, 44, 48].

In this study, we developed a series of nano‐aluminum adjuvants by coating aluminum adjuvant (Adju‐Phos) with low‐molecular‐weight GCS at different ratios using microfluidic technology. The microfluidic process enabled precise regulation of particle size and uniform GCS coating, yielding well‐defined nanoscale structures with stable positive surface potential. GCS provides a hydrated, sterically protective polymer layer around aluminum cores while also introducing protonatable amino groups that alter particle charge. The strong electrostatic interaction between the positively charged GCS layer and the negatively charged aluminum cores stabilizes the particles under near‐neutral pH conditions. Compared with laser‐ablated nano‐aluminum particles and hydrothermally prepared nano‐aluminum adjuvant (the latter optimizes intrinsic colloidal stability by reducing surface free energy), the surface engineering‐based design of GCS‐coated aluminum adjuvant offers not only nanoscale size control but also tunable surface charge, improved formulation stability, resistance to autoclaving and freeze‐drying, and long‐term room‐temperature stability after lyophilization. These properties are of central importance to nano‐aluminum engineering, as formulation performance is closely dictated by surface chemistry and charge. The resultant lyophilized nano‐aluminum adjuvants showed excellent stability during 15‐month storage at room temperature. When combined with HPV16 L1 virus‐like particles (VLPs), nano‐aluminum adjuvant with GCS/alum weight ratio of 7:1 (G7‐A1) displayed superior adjuvant activity as evidenced by enhanced antigen‐specific IgG antibody and neutralizing antibody (nAb) levels and balanced IgG1/IgG2c ratio. To further enhance G7‐A1 adjuvant activity, we attempted to combine it with MF59‐like emulsion, IMQ, or CdA. Among these complex adjuvant systems, the combination of G7‐A1 with MF59‐like emulsion or CdA significantly enhanced HPV16 L1 VLP‐specific humoral and cellular immunity, memory T cell responses, and the activation of DCs in dLNs. These results highlighted the critical role of surface‐engineered nano‐aluminum in linking physicochemical properties to immunological outcomes and demonstrate the potential of this materials‐engineering strategy for next‐generation vaccine adjuvant design.

2. Material and Methods

2.1. Synthesis of Nano‐Aluminum Adjuvants

Glycol chitosan (GCS, degree of polymerization ≥ 200, degree of acetylation = approximately 80%) was purchased from FUJIFILM Wako Pure Chemical Corporation, Japan. Low‐molecular‐weight GCS was prepared according to the previous report [49]. In brief, 0.8 g GCS was dissolved in 60 mL of 4 M HCl. The solution was magnetically stirred at a speed of 600 rpm at 50°C for 96 h and then dialyzed in deionized (DI) water for 24 h until pH reached 7.0. The resultant solution containing low‐molecular‐weight GCS was freeze‐dried, and the molecular weights of both the original GCS and the resultant low‐molecular‐weight GCS were determined using Viscotek Gel Permeation Chromatography with integrated differential refractive index detector and 270 dual detectors (Malvern Panalytical, UK). GPC curves, the number‐average molecular weight (Mn), weight‐average molecular weight (Mw), z‐average molecular weight (Mz), peak molecular weight (Mp), and polydispersity (Mw/Mn) of the original material GCS and the low‐molecular‐weight GCS (derived from three independent degradation batches) are shown in Figure S1 and Table S1 of Supplementary material. The original GCS showed a Mn of 203.55 kDa, a Mw of 323.62 kDa, and a polydispersity of 1.59, while the degraded GCS gave a Mn of 15.29 ± 0.68 kDa, a Mw of 21.4 ± 1.38 kDa, and a polydispersity of 1.40 ± 0.05. Subsequently, the resultant low‐molecular‐weight GCS was added into Adju‐Phos (InvivoGen, San Diego, CA, USA) at a weight ratio of 1:1 to 7:1, respectively (Table 1). The resultant mixtures were diluted in DI water and were stirred overnight at 4°C. A high‐shear homogenizer (MICCRA, Germany) was used to break up the large particles at 11 000 rpm for 5 min. To further reduce particle size, the above GCS‐aluminum adjuvant mixtures were microfluidized in M‐110L microfluidizer (Microfluidics Inc., USA) at 50 psi for 35 passes. The particle size and zeta potential of the resultant nano‐aluminum adjuvants were determined with dynamic light scattering (Malvern Panalytical, UK), and the pH was measured using a pH meter (Mettler‐Toledo, CH).

TABLE 1.

Formulations of Gn‐A1 nano‐aluminum adjuvants.

Name	G0‐A1	G1‐A1	G2‐A1	G3‐A1	G4‐A1	G5‐A1	G6‐A1	G7‐A1
GCS (mg/mL)	0	0.5	1	1.5	2	2.5	3	3.5
Al3+ (mg/mL)	0.5	0.5	0.5	0.5	0.5	0.5	0.5	0.5

2.2. Protein Expression, Purification and Assembly

pET30a‐HPV16 L1 recombinant plasmid (obtained from BCHT, Changchun, China) and pTf16 molecular chaperone expression plasmid (Takara, Japan) were co‐transfected into ER2566 competent cells (Weidi, Shanghai, China). The cells were cultured in Lysogeny Broth (LB) medium at 37°C for 4 h. When the optical density at 600 nm (OD 600) reached 0.6, isopropyl‐β‐d‐thiogalactoside was added at a final concentration of 0.1 mM to induce L1 protein expression. Subsequently, the culture was further incubated at 25°C for 16 h. Cells were harvested by centrifugation and subjected to ultrasonication lysis. The supernatant was supplemented with 20 mM dithiothreitol (DTT) for protein reduction and subsequently purified through cation exchange chromatography using POROS XS resin (Thermo Fisher Scientific, Eugene, OR, USA). The purified protein was subjected to a disassembly/reassembly process to generate uniform HPV16 L1 VLPs.

2.3. Adsorption Capacity of Nano‐Aluminum Adjuvants for HPV16 L1 VLPs

HPV16 L1 VLPs (20 µg/mL) were adsorbed on the above nano‐aluminum adjuvants (20, 200 or 400 µg/mL) by continuously stirring the mixture of antigen and adjuvants at a desired mass ratio in 8 mM phosphate buffer (PB, pH 6.0) containing 0.15 M NaCl and 0.05% polysorbate 80 at 4°C for 1 h. The supernatant was collected by centrifugation at 100 000 g for 16 min at 4°C and then 10‐fold concentrated using a 10 kDa ultrafiltration tube. Considering that the hydroxyl groups of GCS may reduce Cu^{2 +} under alkaline conditions, potentially leading to false‐positive outcomes in Bicinchoninic Acid (BCA) assay, and given the limited availability and high cost of commercial ELISA kits for detecting the HPV16 L1 VLPs, we opted to use SDS‐PAGE for quantifying the antigen in the supernatant. A direct correction of BCA interference was not considered because the GCS/alum ratio differs among formulations, and the residual GCS in processed samples is formulation‐dependent and cannot be predetermined; therefore, a universal blank subtraction or correction factor would be unreliable. As a result, the protein concentration in the concentrated supernatant was analyzed by SDS‐PAGE using 12% preformed gels, and bovine serum albumin (BSA) solutions with a concentration range of 25–300 µg/mL were used to generate the working standard curve for quantification. Both the supernatant and the BSA standards were loaded at 10 µL per lane. Following electrophoresis, proteins were visualized by immersion in Coomassie brilliant blue solution (30 min) and subsequently destained in a methanol‐acetic acid solution [45% (v/v) methanol, 10% (v/v) acetic acid] with gentle agitation overnight. Gel images were digitally captured using an Odyssey infrared imaging system (Gene Company Limited, Hong Kong China), and protein concentration in the supernatant was calculated according to the standard curve. The adsorption rate was calculated according to the following equation and therefore represents an apparent adsorption efficiency under the present assay conditions.

$$\begin{array}{cc}& \mathrm{The}\mathrm{adsorption}\mathrm{rate}\left(\%\right)\hfill \\ & =\left(1-\frac{\mathrm{the}\mathrm{amout}\mathrm{of}\mathrm{protein}\mathrm{in}\mathrm{the}\mathrm{supernatant}}{\mathrm{the}\mathrm{total}\mathrm{amout}\mathrm{of}\mathrm{protein}}\right)\times 100\%\hfill \end{array}$$

2.4. Autoclaving and Lyophilization of Gn‐A1 Nano‐Aluminum Adjuvants

Sterilization of nano‐aluminum adjuvants was performed via autoclaving at 0.1 MPa (121°C) for 20 min. As for freeze‐drying, 2%‐10% (w/v) cryoprotectant trehalose was added to nano‐aluminum adjuvant suspension at different ratios, and the resultant mixtures were lyophilized using a freeze‐dryer (Christ, Germany). The lyophilized powder was kept at room temperature in a desiccator. At days 193 and 458, the size and zeta potential of nano‐aluminum adjuvants obtained through rehydration of lyophilized powder were measured by dynamic light scattering (Malvern Panalytical, UK).

2.5. Transmission Electron Microscopy (TEM)

Carbon‐coated 200‐mesh grids were immersed in a 10 µL droplet of freshly prepared samples with/without antigen or GCS for 10–15 min, and were stained with 2% (w/v) phosphotungstic acid for 30 s. Grids were finally re‐wicked and allowed to dry overnight, and then were analyzed via Talos L120C TEM (120 kV, Thermo Fisher, Eugene, OR, USA).

2.6. Preparation of Complex Adjuvant Systems

ZMF59, a homemade MF59‐like squalene‐based oil‐in‐water emulsion, was prepared as previously described via our vaccine adjuvant platform technology [50, 51]. ZMF59 has the same composition as commercial MF59, consisting of 3.9% (w/v) squalene, 0.47% (w/v) Tween 80, 0.47% (w/v) Span 85, and citrate buffer (10 mM, pH 6) [45]. ZMF59 showed a uniform particle size of 150–160 nm, and its polydispersity index (PDI) was less than 0.15. Three complex adjuvants were prepared through mixing G7‐A1 nano‐aluminum adjuvant with ZMF59, IMQ (Sigma‐Aldrich, Saint Louis, MO, USA), or CdA disodium (‌MedChemExpress, New Jersey, USA), respectively.

2.7. Mice, Immunizations and Reproducibility Statement

C57BL/6 mice (6–8 weeks old, female) were purchased from Liaoning Changsheng Biotechnology Co. Ltd (Liaoning, China) and were maintained under specific pathogen‐free (SPF) conditions. For the animal studies, the nano‐aluminum adjuvant formulations used for immunization were sterilized by autoclaving prior to administration. To illustrate different adjuvant activity of nano‐aluminum adjuvants, the immunization was performed by intramuscular injection of 50 µL vaccine into each hind leg at weeks 0 and 2. All vaccines contained 2 µg HPV16 L1 VLPs, but were formulated with different adjuvants: 50 µg nano‐aluminum adjuvants (Gn‐A1 groups), 350 µg GCS (20 kDa) (GCS group), 50 µg Adju‐Phos (Adju‐Phos group) or 50 µg Adju‐Phos with 5 µg MPLA (Avanti Research, Alabaster, Bama, USA) (ZAS04 group). PBS and 2 µg HPV16 L1 VLPs were used as controls. Blood samples were collected at weeks 2 after the primary and boost immunizations, respectively.

To evaluate complex adjuvant systems, female C57BL/6 mice (n = 5) aged 6–8 weeks were intramuscularly injected with 4 µg HPV16 L1 VLPs formulated with various adjuvants in 100 µL in the hind legs twice at a 2‐week interval: 19.73 µg G7‐A1 plus 59.2 µL ZMF59, 50 µg G7‐A1 plus 7.5 µg IMQ, 50 µg G7‐A1 plus 7.5 µg CdA. In addition, 50 µg G7‐A1 and 50 µg Adju‐Phos were used as controls. Serum samples were collected at days 14, 28, 56 and 84, and spleens were collected at day 28 after primary immunization. The above studies were conducted following the Guide for the Care and Use of Laboratory Animals (National Research Council). All procedures were reviewed and approved by the Animal Welfare and Research Ethics Committee at Jilin University (Ethical approval number 2022YNPZSY0708). Mice were euthanized using CO₂ asphyxiation at the end of the experiment.

Animals were randomly assigned to experimental groups before immunization, and biologically independent mice were used as the unit of analysis. Where applicable, repeated measurements or replicate wells derived from the same sample were treated as technical replicates. Blinding was not formally implemented during all experimental procedures; however, sample processing, data acquisition, and analysis were performed using the same predefined workflow across all groups.

2.8. Determination of HPV L1 VLP‐Specific Antibody Titers

Mouse blood was centrifuged twice at 3,000 rpm for 10 min. Mouse sera were collected and treated by thermal inactivation for 30 min at 56°C before detection. The ELISA plates were coated with 1 µg/mL HPV16 L1 VLPs overnight at 4°C. The plates were then washed three times with PBS containing 0.05% Tween‐20 (PBST) and blocked with 200 µL/well of 5% BSA in PBST for 1–2 h at 37°C. After washing with PBST for another three times, two‐fold serial dilutions were prepared in 1% BSA for each serum sample in eight consecutive wells of one column (100 µL/well) and then incubated for 1 h at 37°C. Plates were then washed three times with PBST. Due to the fact that C57BL/6 mice secrete IgG2c antibodies and do not produce IgG2a isotypes [52, 53, 54, 55], horseradish peroxidase (HRP)‐conjugated goat anti‐mouse immunoglobulin reagents specific for mouse IgG (Jackson ImmunoResearch, Lancaster, PA, USA), IgG1 or IgG2c (Abcam, Cambridge, UK), were diluted in 1% BSA (1:10,000) and added to the plates (100 µL/well). Following incubation at 37°C for 1 h, the plates were washed three times and then 50 µL/well of 3,3’,5,5’‐Tetramethylbenzidine (TMB) substrate was added to each well at room temperature. The reaction was stopped with 2 M H₂SO₄ after 15–30 min. OD450 was recorded using a microplate reader (Biotek, Vermont, USA). Antibody titers were defined as the highest serum dilution yielding an OD450 value at least 2.1‐fold higher than that of the negative serum.

2.9. Pseudovirus‐Based Neutralization Assay

HPV16 pseudovirions (PsV) were generated by co‐transfecting two plasmids encoding HPV16 L1 and red fluorescent protein (RFP) into 293FT cells (Thermo Fisher, Eugene, OR, USA). The TCID₅₀ (tissue culture infective dose) of the PsVs was calculated according to the classical Reed‐Muench method. For neutralization assay, 293FT cells were plated at 1.5×10⁴ cells/well in 96‐well plates in Dulbecco's modified Eagle medium (Gibco, CA, USA), supplemented with 10% fetal bovine serum (FBS) (Gibco, CA, USA) and 1% penicillin‐streptomycin, and incubated for 6 h at 37°C with 5% CO₂. Each serum sample was serially diluted two ‐ fold into seven or eight gradients, and PsVs were diluted into 200 TCID₅₀/50 µL. Equal volumes of the PsVs and the serially diluted sera were mixed and incubated at 4°C for 1 h. 100 µL of these mixtures was transferred to plates seeded with a monolayer of 293FT cells and incubated at 37°C for 72 h. Cells alone and cells incubated with PsVs were employed as negative and positive controls, respectively. RFP expression spots of 293FT cells were counted using a Cytation 3 imaging reader (BioTek, Vermont, USA). The endpoint titers were calculated as the log₂ of the highest serum dilution with a percent infection inhibition higher than 50%.

2.10. Preparation of Single Cell Suspensions From Spleens

Mice were euthanized at day 28 after boost immunization and soaked in 75% ethanol for 5 min. The spleens were isolated and processed into a single‐cell suspension by gentle grinding and passage through a 70 µm filter. Cells were pelleted by centrifugation at 350 g for 15 min, and the resultant cell precipitate was resuspended in 1 mL RPMI‐1640 medium followed by the addition of 5 mL ACK lysis buffer (Solarbio, Beijing, China). Cells were then centrifuged at 350 g for 10 min, and the precipitate was washed with 10 mL RPMI‐1640 medium followed by centrifugation at 350 g for 5 min. Finally, cells were resuspended in 1 mL of RPMI‐1640 medium supplemented with 10% FBS and 1% penicillin‐streptomycin (complete medium) and diluted to 10⁷ cells per mL after counting.

2.11. ELISPOT Assay for IFN‐γ Production

Pre‐coated IFN‐γ ELISpot plates (Mabtech, Sweden) were washed 4 times with sterile PBS (200 µL/well), and then blocked with complete medium for at least 30 min at room temperature. After removing the medium, the stimuli (20 µg/mL HPV16 L1 VLPs or medium alone) were added. As a positive control, 10 µL solution containing PMA (50 ng/mL) and ionomycin (1 µg/mL) was added to one well for each group. After that, spleen cells prepared as described above were seeded into ELISpot plates at a final concentration of 8×10⁵ per well in duplicate. After incubation at 37°C and 5% CO₂ for 24 h, the cells were washed off five times with PBS. Plates were then incubated with biotinylated anti‐mouse IFN‐γ (Mabtech, Sweden) for 2 h at room temperature, followed by washing five times with PBS. Then, the plates were incubated with HRP‐conjugated streptavidin (Mabtech, Sweden) for 1 h at room temperature. After washing five times with PBS, spots were developed with 100 µL/well TMB (Mabtech, Sweden). Finally, plates were air‐dried for 24 h after terminating the chromogenic reaction with deionized water and spots were counted in an ELISpot reader (Cellular Technology Ltd, Cleveland, OH, USA).

2.12. Flow Cytometry Analysis

10⁶ splenocytes per well, prepared as described in section 2.10, were stimulated in 96‐well plates with 2 µg HPV16 L1 VLPs in the presence of CD28 and CD49d antibodies (Biolegend, San Diego, CA, USA) for 2 h at 37°C and 5% CO₂ in an incubator, and then Brefeldin A protein transport inhibitor (Invitrogen, Carlsbad, CA, USA) was added at 1:1000 dilution and incubated for 6 h. After incubation, plates were centrifuged at 4°C, 600 g for 5 min and the cell pellet was resuspended in 300 µL cell staining buffer (Biolegend, San Diego, CA, USA) and was washed followed by centrifugation again. The supernatant was discarded and Fc receptors on the splenocytes were blocked with anti‐mouse CD16/32 antibody (Biolegend, San Diego, CA, USA) on ice for 10 min.

To detect multifunctional T cell response, surface staining was conducted with a cocktail including CD3‐FITC (Biolegend, San Diego, CA, USA), CD4‐APC‐eFluor780 (eBioscience, San Diego, CA, USA) and CD8‐PE (BD Pharmingen, San Diego, CA, USA) for 40 min in the dark at 4°C. Then, cells were washed with staining buffer, and were resuspended in 200 µL of fixative solution (eBioscience, San Diego, CA, USA) per well followed by incubation for 20 min. After centrifuging at 800 g for 5 min, cells were resuspended in 200 µL of permeabilization buffer (eBioscience, San Diego, CA, USA) per well and then were centrifuged again followed by removing the supernatant. The resultant cells were stained intracellularly with TNF α‐eFluor 450 (eBioscience, San Diego, CA, USA), IL‐2‐APC (Biolegend; San Diego, CA, USA) and IFN‐γ‐PerCP‐Cy5.5 (BD Pharmingen, San Diego, CA, USA) in permeabilization buffer for 30 min. Finally, cells were washed once and resuspended with permeabilization buffer before performing flow cytometry assay using CytoFLEX instrument (Beckman Coulter, Indianapolis, IN, USA). The gated populations were further analyzed with CytExpert 2.4 software. The gating strategy is shown in Figure S4.

As for the assay of memory T cells, surface staining was performed the same way as described previously. Besides, the splenocyte suspensions were labeled with CD3‐FITC, CD4‐APC‐eFluor780, CD8‐APC, CD44‐PE and CD62L‐BV421. The gating strategy is shown in Figure S5.

2.13. Measurement of Cytokine Secretion in Single Cell Suspensions From Spleens

Splenocytes prepared as described in section 2.10 were cultured at a final concentration of 5×10⁶ cells/well in 0.5 mL in flat‐bottom 24‐well plates and incubated in complete medium containing HPV16 L1 VLPs at the final concentrations of 10 µg/mL at 37°C in a 5% CO₂ incubator. Besides, an equal volume of complete medium or 10 µL of PMA/ionomycin (50 ng/mL and 1 µg/mL, respectively) was added into corresponding wells as the negative control or positive control, respectively. Supernatants were harvested after 48 h and tested for the presence of IFN‐γ, IL‐2, IL‐4 and IL‐10 using a mouse Th1/Th2 uncoated ELISA kit (Thermo Fisher, Eugene, OR, USA).

2.14. Histological Examination

C57BL/6 mice were injected into the hind leg with HPV16 L1 VLPs formulated with varied adjuvants, respectively. Additionally, sterile PBS and ZAS04 were used as controls. Mice were euthanized on days 0, 1 and 7 post‐immunization. Muscles at the injection site were removed and fixed with 4% paraformaldehyde. Fixed tissues were embedded in paraffin blocks, sectioned, and stained with hematoxylin and eosin (H&E). H&E‐stained slides were pictured and analyzed with SLIDEVIEW VS200 (OLYMPUS, Tokyo, Japan). The images were generated using OLYMPUS OlyVIA software.

2.15. Immunofluorescence Analysis

C57BL/6 mice were injected into the hind leg with HPV16 L1 VLPs formulated with different adjuvants. In addition, sterile PBS and ZAS04 were used as controls. Mice were euthanized on days 0, 1 and 7 after injection, and the inguinal dLNs from each mouse were removed and fixed with 4% paraformaldehyde. Fixed tissues were embedded in paraffin blocks, sectioned, and stained with anti‐CD11c (Servicebio, Wuhan, China), re‐stained with peroxidase‐labelled goat anti‐rabbit IgG antibody (SeraCare, Milford, MA, USA), and thereafter incubated for 20 min at room temperature with Cyanine3‐labelled tyramine signal amplification (TSA) reagent. This was followed by a second round of staining using the next primary antibody, anti‐CD86 antibody (Cell Signaling Technology, Danforth, MA, USA), stained again with peroxidase‐labelled goat anti‐rabbit IgG secondary antibody (SeraCare, Milford, MA, USA), and then incubated with Alexa Fluor 488 labelled TSA reagent for 20 min. Finally, the nuclei were labelled with DAPI. Stained slides were scanned and analyzed with SLIDEVIEW VS200 (OLYMPUS, Tokyo, Japan). The images were generated using OLYMPUS OlyVIA software.

2.16. Statistics Analysis

No data transformation or normalization was performed unless otherwise stated. No animals or samples were excluded from the analysis unless a predefined technical failure occurred, such as failed sample collection or assay failure. Data are presented as mean ± SD unless otherwise indicated. Statistical analyses were performed by GraphPad Prism V.10.1.2 software (San Diego, CA, USA). Comparisons between two groups were performed using an unpaired t‐test. Comparisons among multiple groups were performed using one‐way analysis of variance (ANOVA) followed by an appropriate multiple‐comparisons test when applicable. For experiments involving two independent variables, two‐way ANOVA followed by an appropriate multiple‐comparisons test was used. A p‐value of ≤ 0.05 was considered statistically significant. Statistical tests were selected according to the experimental design and are specified in the corresponding figure legends. The number of biologically independent mice used for each experiment is indicated in the corresponding figure captions.

3. Results

3.1. Fabrication and Characterization of Gn‐A1 Nano‐Aluminum Adjuvants

Adju‐Phos is primarily composed of aggregated plate‐like entities with a size of 15–50 nm, which assemble into large and heterogeneous aggregates of variable overall size through the random association and aggregation of these plate‐like structures [3]. These particles can be temporarily dispersed through sonication or other high‐shear methods but will rapidly re‐aggregate shortly afterward [2]. To obtain monodisperse particles, we introduced a low‐molecular‐weight GCS as a stabilizer. The naive GCS showed a Mw of 323 kDa (Figure S1A), while the degraded GCS gave an average Mw of 20 kDa (Figure S1B‐D). After microfluidization of the GCS‐aluminum adjuvant mixture, a stable, uniformly milky nano‐aluminum suspension was generated (Figure 1A). The average particle size of G0‐A1 and G1‐A1 ranged between 190 and 195 nm, while the remaining nano‐aluminum adjuvants produced with higher concentrations of GCS showed a diameter of approximately 178 nm. The PDI of all nano‐aluminum adjuvants ranged from 0.1 to 0.148, indicating a monodisperse distribution (Figure 1C). G0‐A1, without GCS coating, exhibited a potential of ‐21.8 mV. With increasing GCS/AP ratios, the zeta potentials of nano‐aluminum adjuvants increased from 21.17 mV for G1‐A1 to 33.1 mV for G7‐A1 (Figure 1E). Concomitantly, the pH values decreased from 6.88 to 5.293 (Table S2). TEM analysis showed that Adju‐Phos has an average size of about 1–3 µm as previously reported, and displays a visibly higher particle concentration, which can be attributed to its more pronounced aggregation of plate‐like microstructures (Figure 1G) [3, 56]. Gn‐A1 nano‐aluminum adjuvants displayed irregular shapes with an average size of 100–200 nm, and no significant difference was observed across varying GCS ratios (Figure S2). After staining G7‐A1 with 2% phosphotungstic acid, a shell surrounding the nano‐aluminum core was observed, which was distinct from the appearance of G0‐A1 (Figure 1H and Figure S3). This shell is presumed to be composed of GCS, as only the hydroxyl and amide bonds present in GCS can interact with phosphotungstic acid, leading to the deposition of the dye.

FIGURE 1.

Characterization and autoclave tolerance of Gn‐A1 nano‐aluminum adjuvants. Appearance of G0‐A1 to G7‐A1 before (A) and after autoclaving (B). Particle size distribution before (C) and after autoclaving (D). Zeta potential (E) and pH values (F) before and after autoclaving. Data are mean ± SD from three independent repeated experiments (n = 3). TEM images: (G) unstained Adju‐Phos; (H) G0‐A1 and G7‐A1 stained with 2% (w/v) phosphotungstic acid. Scale bars: 200 nm and 100 nm in (G); 200 nm in (H). No statistical comparison was performed unless otherwise indicated.

After autoclaving, nano‐aluminum adjuvants of G1‐A1 to G3‐A1 exhibited varying degrees of precipitation at the bottom of the tubes (Figure 1B). Although no precipitate was found in G4‐A1, its particle size was significantly larger than those of G5‐A1 to G7‐A1 (Figure 1D). When GCS/alum weight ratio was higher than 4:1, all nano‐aluminum adjuvants displayed particle sizes ranging from 207 to 211 nm, with PDI values of 0.131 to 0.136, indicating a monodisperse system (Figure 1D and Table S3). We speculated that autoclaving may lead to GCS detachment and slight degradation of GCS at low GCS/alum weight ratios, and therefore insufficient GCS coating reduced electrostatic repulsion between nano‐aluminum particles. As a result, aggregation happened to G1‐A1‐G3‐A1 after autoclaving. With increasing GCS/alum weight ratios, the high content of GCS could ensure sufficient GCS coating and therefore strong electrostatic repulsion even though the detachment and degradation of a small quantity of GCS may occur in G5‐A1‐G7‐A1. The above rational speculation was supported by the results of zeta potential and pH determination (Figure 1E,F, and Table S3). In fact, GCS detachment and breakage resulting from autoclaving led to a decrease in the zeta potential of all nano‐aluminum adjuvants to varying degrees, due to the increased pH, compared with that of the adjuvants without autoclaving. (Tables S2 and S3). In summary, G5‐A1 to G7‐A1 can withstand standard autoclave without significant changes in physicochemical characteristics owing to high GCS/alum weight ratios.

3.2. Antigen Adsorption Evaluation

To investigate the adsorption capacity of various Gn‐A1 nano‐aluminum adjuvants to HPV16 L1 VLPs, we first expressed the major capsid protein L1 and then prepared HPV16 L1 VLPs via self‐assembly process [57]. The resultant HPV16 L1 VLPs showed a size of approximately 60 nm and a zeta potential of 1.5 mV (Figure 2A and Table S6), which agrees well with its isoelectric point of 7.95 [58]. As expected, when adding HPV16 L1 VLPs to G0‐A1, precipitation happened. This can be explained by VLP adsorption‐mediated flocculation of nano‐aluminum particles. It is well known that antigen can adsorb onto Adju‐Phos adjuvant via electronic interaction. This will also happen between positively charged VLPs and negatively charged G0‐A1 adjuvant, although HPV16 L1 VLPs carry weak positive charges. The rational speculation is supported by the increasing zeta potential with increasing antigen/G0‐A1 weight ratios. For GCS‐coated nano‐aluminum adjuvants, VLPs were observed within the GCS shell of Gn‐A1 as evidenced by the results of TEM analysis (Figure 2A). This observation indicates that antigen association in Gn‐A1 is not limited to a simple outer‐surface adsorption mode. Because the adsorption efficiency was quantified from the residual antigen remaining in the supernatant after ultracentrifugation, the reported values should be interpreted as apparent adsorption efficiencies. For GCS‐coated particles, antigen may penetrate the hydrated polymer shell, and some shell‐associated antigen may not sediment completely, leading to a conservative estimate of actual adsorption. The penetration of VLPs into the GCS shell may be attributed to hydrogen bonding, van der Waals forces, and hydrophobic interaction, although weak electrostatic repulsion may undermine their interactions [59, 60]. However, it should be noted that HPV16 L1 protein is a large polymer‐like molecule, and both negatively charged and positively charged amino acids exist on its surface [58]. Therefore, electrostatic interaction might still exist between GCS and HPV16 L1 VLPs.

FIGURE 2.

Characterization of HPV16 L1 VLPs adsorption onto Gn‐A1. (A) Representative TEM images of HPV16 L1 VLPs, antigen‐absorbed G0‐A1 and G7‐A1 stained with 2% (w/v) phosphotungstic acid. Scale bar indicates 200 nm. (B) Binding capacity of different aluminum adjuvants to HPV16 L1 VLPs at a ratio of 1/20 (VLPs/alum, w/w). The zeta potential (C), size (D) and PDI (E) of antigen‐adsorbed Gn‐A1 adjuvants. The data are shown as mean ± SD of three independent repeated experiments (n = 3). No statistical comparison was performed unless otherwise indicated.

With respect to adsorption capacity, as expected, a higher percentage of HPV16 L1 VLPs can be loaded onto Gn‐A1 adjuvants with increasing alum/antigen weight ratios (Table S4). Interestingly, with increasing GCS/alum weight ratios, the adsorption capacity of Gn‐A1 nano‐aluminum adjuvants decreased (Tables S4 and S5), which is probably due to the shielding of negatively charged particle surfaces of nano‐aluminum adjuvants and the limited loading capacity of GCS shell. At the 1/20 weight ratio of HPV to alum, only about 50% antigen was loaded onto G5‐A1 to G7‐A1 adjuvants (Figure 2B and Table S5). This relatively constant adsorption rate of G5‐A1 to G7‐A1 can be explained by the complete coating of nano‐aluminum adjuvant particles by GCS and the saturated penetration of antigens in the GCS shell. The rational speculation is supported by their comparable particle size and zeta potential (Figure 2C–E). Notably, when antigen/alum weight ratio is close to 1:1, antigen adsorption‐mediated flocculation of Gn‐A1 nano‐aluminum particles occurred as evidenced by the increasing particle sizes and PDI values (Figure 2D,E). In addition, their zeta potential was also closer to that of HPV16 L1 VLPs (Figure 2C).

3.3. Freeze‐Drying Feasibility

It is well known that traditional aluminum adjuvants can be neither frozen nor freeze‐dried. Here, the above nano‐aluminum adjuvants were lyophilized with cryoprotectant trehalose at various concentrations, and then rehydrated and characterized to illustrate the effect of freeze‐drying on their physicochemical properties. For all nano‐aluminum adjuvants, when trehalose concentration was kept at 2%, various degrees of cake shrinkage happened due to low solid content, especially for G0‐A1 (Figure 3). However, with increasing GCS/alum weight ratios, cake appearance became better. When trehalose concentration was higher than 4%, no cake defects, such as collapse or shrinkage, were found in all freeze‐dried nano‐aluminum adjuvants. However, compared with freshly prepared nano‐aluminum adjuvants, freeze‐dried products with GCS/alum ratios lower than 2:1 showed larger particle sizes and higher PDI values after rehydration even in the presence of 10% trehalose (Figure 3). In contrast, nano‐aluminum adjuvants with GCS/alum ratios higher than 3:1, freeze‐dried with 6% trehalose or higher, displayed comparable particle sizes, PDI and zeta potentials in comparison with those of freshly prepared nano‐aluminum adjuvants. Similar phenomena were also noted for freeze‐dried nano‐aluminum adjuvants which were stored for 15 months (Figure S6). The above results suggest that high contents of GCS and trehalose can efficiently prevent particle aggregation of freeze‐dried nano‐aluminum adjuvants after rehydration. The excellent freeze‐drying resistance conferred by GCS and trehalose may facilitate future storage and transportation of nano‐aluminum adjuvants.

FIGURE 3.

Appearance, hydrodynamic diameter, and zeta potential of rehydrated nano‐aluminum adjuvants which were freeze‐dried and stored for 193 days at room temperature. Data are representative of three independent repeated experiments, and are shown as mean ± SD (n = 3). No statistical comparison was performed unless otherwise indicated.

3.4. Nano‐Aluminum Adjuvants Induced Balanced Humoral and Cellular Immunity

To investigate the adjuvant activity of the above nano‐aluminum adjuvants, C57BL/6 mice were immunized intramuscularly twice at a 2‐week interval with HPV16 L1 VLPs formulated with various adjuvants. As shown in Figure 4, nano‐aluminum adjuvants with GCS/alum ratios higher than 5:1 elicited significantly higher antigen‐specific IgG than antigen‐alone group 2 weeks after primary immunization (Figure 4B). A similar trend was also observed for nAb titers, although only G7‐A1 showed statistically significant difference (Figure 4F). After booster injection, markedly increased IgG and nAb titers were noted (Figure 4C,G). In addition, all groups vaccinated with adjuvanted vaccines, but not the GCS group, showed significantly higher antibody levels than those of antigen‐alone group, indicating strong adjuvant activity. Interestingly, with increasing GCS/alum ratios, IgG1/IgG2c ratio decreased (Figure 4D,E), suggesting more balanced humoral and cellular immunity. The above trend of immune bias is largely attributed to the ability of GCS to induce a balanced immune response. Moreover, antigen formulated with nano‐aluminum adjuvants with GCS/alum ratios higher than 5:1 led to slightly higher nAb levels than those of Adju‐Phos and ZAS04 (Figure 4G). Interestingly, Adju‐Phos outperformed G0‐A1 in generating both antigen‐specific and neutralizing antibodies. This may stem from their distinct antigen distribution patterns. In Adju‐Phos, antigens likely reside on the surface of its micron‐sized particles. In contrast, G0‐A1's nanoparticles first adsorb antigens and then aggregate into larger particles via antigen bridging, trapping some antigens internally. This internalization could lead to more complex desorption and slower antigen release in vivo, which is generally less favorable for immune induction [61]. These hypotheses require further experimental validation. In summary, nano‐aluminum adjuvants with high GCS/alum ratios are comparable or superior to Adju‐Phos and ZAS04 in eliciting early antibody production, protective nAb responses, and a balanced Th1/Th2 immune response.

FIGURE 4.

Schematic diagram of mouse immunization and antibody level evaluation. (A) C57BL/6 mice were injected at days 0 and 14, and serum samples were collected at days 14 and 28, respectively. Antibody titers of HPV16 L1 VLP‐specific IgG (B and C), IgG1 and IgG2c (D), IgG1/IgG2c ratio (E), and nAb (F and G) (n = 5). The significance was evaluated using one‐way or two‐way analysis of variance (ANOVA), as appropriate. Unless otherwise indicated, significance markers denote comparisons versus the HPV16 L1 VLPs alone group. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p < 0.0001.

3.5. Physicochemical Characterization of G7‐A1/ZMF59 Complex Adjuvants

To further enhance the adjuvant activity of nano‐aluminum adjuvants, G7‐A1 was chosen to be combined with ZMF59. The resultant G7‐A1/ZMF59 complex adjuvants with different ratios were first characterized in terms of appearance, particle size, and zeta potential to optimize their combination. ZMF59 was a homemade MF59‐like squalene‐based emulsion adjuvant, and gave a milk‐like appearance, a particle size of 152.8 nm, and a zeta potential of ‐25.3 mV (Table S11). As expected, all G7‐A1/ZMF59 complex adjuvants with different volume ratios showed an appearance similar to ZMF59 (Figure 5A). With increasing G7‐A1/ZMF59 ratios, the complex adjuvants showed increased particle sizes, PDI, and zeta potentials, especially when G7‐A1/ZMF59 ratio was higher than 5:5 (Figure 5C,D, and Table S11). No apparent precipitation was noted for all combinations except for the combination with a 9:1 G7‐A1/ZMF59 ratio (Figure 5A). The above phenomena can be explained by the interaction of ZMF59 and G7‐A1 at different ratios. At the low G7‐A1/ZMF59 ratios, positively charged G7‐A1 would be adsorbed onto the surface of negatively charged ZMF59 emulsion droplets. As Adju‐Phos is composed of 15–50 nm plate‐like entities and G7‐A1 has a high GCS/alum ratio, the adsorption of these GCS‐coated plate‐like entities onto the surface of ZMF59 led only to a slightly increased size at the low G7‐A1/ZMF59 ratios. However, zeta potential can be affected to a greater extent as part of the negative charge at the interface of ZMF59 is neutralized. With increasing G7‐A1/ZMF59 ratios, the surface of ZMF59 may be saturated, and the weak interactions between GCS‐coated plate‐like entities would dominate and therefore led to markedly increased particle size. With further increasing G7‐A1/ZMF59 ratios, ZMF59 would adsorb onto the irregular surface of G7‐A1, and therefore lead to flocculation of nano‐aluminum particles. Because ZMF59 cannot cover the entire surface of G7‐A1, a positive zeta potential was observed in this case (Figure 5D). The phenomenon is similar to that of VLP adsorption‐mediated flocculation of G0‐A1. We further characterized HPV16 L1 VLP‐loaded G7‐A1/ZMF59 complex adjuvants. With increasing G7‐A1/ZMF59 ratios, HPV16 L1 VLP‐loaded G7‐A1/ZMF59 showed similar trends to those of complex adjuvants without antigen absorption in terms of size, PDI, and zeta potential, which may be due to the limited dose of antigen (Figure 5E–G). In the subsequent experiment, G7‐A1/ZMF59 with 4/6 ratio was used as HPV16 L1 VLP's adjuvant because of its smaller size and lower PDI (Figure 5C,F).

FIGURE 5.

Characterization of G7‐A1/ZMF59 complex adjuvants with and without HPV16 L1 VLPs. (A) The appearance of G7‐A1/ZMF59 adjuvants with different ratios. The size distribution (B and C) and zeta potential (D) of G7‐A1/ZMF59 adjuvants. Particle size (E), PDI (F), and zeta potential (G) of HPV16 L1 VLP‐loaded G7‐A1/ZMF59 adjuvants. The buffer of HPV16 L1 VLPs solution was 8 mM PB, and contained 0.15 M NaCl and 0.05% polysorbate 80, and has a pH of 6.0. The data are shown as mean ± SD of three independent repeated experiments (n = 3). No statistical comparison was performed unless otherwise indicated.

3.6. G7‐A1‐Based Complex Adjuvants Elicited Strong and Balanced Humoral and Cellular Immunity

To evaluate the adjuvant activity of the above nano‐aluminum complex adjuvant, mice were intramuscularly immunized with HPV16 L1 VLPs formulated with G7‐A1/ZMF59 complex adjuvant at a volume ratio of 4:6 on days 0 and 14, respectively (Figure 6A). In addition, two additional G7‐A1‐based complex adjuvants were prepared by mixing G7‐A1 with molecular immunostimulant IMQ or CdA. The G7‐A1‐based complex adjuvants were prepared immediately before use, and no obvious physical instability was noted during routine short‐term handling (Figure S7). As shown in Figure 6B, both the G7‐A1 and Adju‐Phos triggered significantly higher IgG antibody levels than the antigen‐alone group at 6 weeks post‐boost, and there was no statistical difference between G7‐A1 and Adju‐Phos groups. Moreover, G7‐A1‐based complex adjuvants showed higher IgG, IgG1 and IgG2c titers than those of ZMF59 or the corresponding immunostimulant alone, respectively, after boost inoculation (Figure 6B–D). Furthermore, in comparison to Adju‐Phos group, G7‐A1‐based complex adjuvants induced a more balanced IgG1/IgG2c ratio at week 2 post‐boost (Figure 6E). With respect to nAbs, G7‐A1‐based complex adjuvants elicited remarkably higher antibody titers than ZMF59 or immunostimulant alone after boost injection (Figure 6F). Although no statistically significant difference in nAb titers was observed between G7‐A1 and Adju‐Phos groups, G7‐A1 demonstrated superior modulation of IgG subclass profiles, resulting in a more favorable IgG1/IgG2c balance compared to Adju‐Phos. Interestingly, the Adju‐Phos, G7‐A1, and ZAS04 groups achieved their peak nAb levels at week 6 post‐boost, while the inclusion of IMQ or CdA accelerated the peak nAb response to week 2 post‐boost. This suggests that the incorporation of TLR7/8 or STING agonists facilitates a faster onset of the antibody response. Notably, both G7‐A1/CdA and G7‐A1/ZMF59 maintained superior nAb responses even at week 10 post‐boost. These findings highlight that G7‐A1/CdA or G7‐A1/ZMF59 induced faster and more durable antibody responses, along with more balanced Th1 and Th2 immunity.

FIGURE 6.

Antibody response to HPV16 L1 VLPs formulated with various adjuvants. (A) Schematic diagram of vaccination and serum sampling. Mice (n = 4–5) were intramuscularly injected twice at a 2‐week interval. Serum samples were collected at weeks 2, 4, 8, and 12 after primary immunization. HPV16 L1 VLP‐specific IgG (B), IgG1 (C), IgG2c (D) antibody titers, IgG1/IgG2c ratio (E), and nAb titers (F). The significance was evaluated using a t test or one‐way analysis of variance (ANOVA), as appropriate. Unless otherwise indicated, significance markers denote comparisons against HPV16 L1 VLPs alone or the corresponding single‐component adjuvant control (IMQ, CdA, or ZMF59 alone). ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p < 0.0001.

3.7. G7‐A1‐Based Complex Adjuvants Induced Potent Cellular Immunity

To illustrate the effect of complex adjuvants on cellular and humoral immunities, mice were immunized via intramuscular injection with various vaccines formulated with either G7‐A1 plus IMQ (7.5 µg per dose), CdA (7.5 µg per dose), or ZMF59 (at a 4/6 volume ratio) at a 2‐week interval, and G7‐A1 and Adju‐Phos‐adjuvanted vaccines were used as controls (Figure 7A). As shown in Figure 7C–E, G7‐A1 induced significantly higher levels of TNF‐α, IL‐4, and IL‐10 compared with the Adju‐Phos group. Notably, when G7‐A1 was combined with IMQ, CdA or ZMF59, the resultant complex adjuvants promoted the production of IFN‐γ, TNF‐α, IL‐4, and IL‐10 to varying extents (Figure 7B–E). It is well established that CD4⁺ Th1 cells and CD8⁺ CTLs mainly produce IFN‐γ and TNF‐α, while Th2 cells produce IL‐4 [62, 63]. IL‐10, mainly derived from CD4⁺ Treg cells, is an anti‐inflammatory cytokine that enhances B cell survival, prevents GC (germinal center) B cell apoptosis, promotes their development into plasma cells, and synergistically acts with IL‐4 while antagonizing TNF‐α [64, 65, 66, 67]. The above results suggested that G7‐A1‐based complex adjuvants induced not only robust cellular immunity but also humoral immunity.

FIGURE 7.

Schematic diagram of vaccination and evaluation of cellular immune responses. (A) C57BL/6 mice were intramuscularly injected at days 0 and 14, and the cytokine profiles of splenocytes were detected on day 28. The levels of IFN‐γ (B), TNF‐α (C), IL‐4 (D), and IL‐10 (E) in splenocyte supernatants were determined by ELISA (n = 3–5). Frequency of HPV16 L1 VLP‐specific IFN‐γ‐secreting splenocytes (F) and representative images (G) of IFN‐γ ELISPOT assay (n = 4–5). Frequency of HPV16 L1 VLP‐specific CD4⁺ (H) and CD8⁺ T cells (I) expressing IFN‐γ, IL‐2, and TNF‐α determined by flow cytometry (n = 4–5). Representative FACS gating is shown in Figure S4. The significance was evaluated using a t test or one‐way analysis of variance (ANOVA), as appropriate. Unless otherwise indicated, significance markers denote comparisons against Adju‐Phos or the corresponding single‐component adjuvant control (IMQ, CdA, or ZMF59 alone). ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, ^**** p < 0.0001.

To further illustrate the potential of G7‐A1‐based complex adjuvants, cellular immune responses were evaluated by measuring the frequency of IFN‐γ‐secreting splenocytes using the ELISPOT assay. As shown in Figure 7F,G, the frequency of IFN‐γ⁺ splenocytes induced by G7‐A1 was 2.73‐fold higher than that in the AP group (p = 0.0031). Furthermore, the G7‐A1/IMQ group exhibited 7.23‐fold higher IFN‐γ⁺ splenocytes compared to the IMQ group. A similar trend was observed in the G7‐A1/ZMF59 group, with the frequency of IFN‐γ⁺ splenocytes being 4.49‐fold higher than that in the ZMF59 group. Interestingly, CdA induced more IFN‐γ spots than IMQ or ZMF59. Moreover, when combined with G7‐A1, the resultant complex adjuvant led to a 2.32‐fold higher frequency of IFN‐γ⁺ splenocytes (p = 0.0189) than the CdA‐alone group. These results confirm that the combination of CdA with G7‐A1 can significantly enhance antigen‐specific cellular immune responses.

Subsequently, the effect of G7‐A1‐based complex adjuvants on antigen‐specific multifunctional T cell responses was also investigated. As shown in Figure 7H, the G7‐A1/CdA adjuvant induced the highest percentage of multifunctional CD4⁺ T cells among all adjuvants followed by G7‐A1/ZMF59. In the case of CD8⁺ T cells, G7‐A1 alone induced the highest percentage of antigen‐specific multifunctional CD8⁺ T cells, followed by G7‐A1/CdA (Figure 7I). Notably, both G7‐A1/CdA and G7‐A1/ZMF59 were superior to Adju‐Phos and ZAS04 in both CD4⁺ and CD8⁺ multifunctional T cell induction. In summary, the above results further confirmed the advantages of G7‐A1 and G7‐A1‐based complex adjuvants (especially G7‐A1/CdA and G7‐A1/ZMF59) over Adju‐Phos alone and immunostimulants, respectively, in inducing cellular immunity.

3.8. G7‐A1‐Based Complex Adjuvants Enhanced Antigen‐Specific Memory T Cell Responses

To evaluate the effect of G7‐A1‐based complex adjuvants on antigen‐specific memory T cell responses, CD44 and CD62L were used as markers to identify central memory T cell (Tcm) and effector memory T cells (Tem). As shown in Figure 8A, the G7‐A1/CdA group significantly enhanced the CD4⁺ Tcm response compared to other groups, followed by G7‐A1/ZMF59. In contrast, G7‐A1/IMQ induced a higher frequency of CD8⁺ Tcm cells (Figure 8D). With respect to Tem, G7‐A1 elicited more robust CD4⁺ and CD8⁺ Tem responses than Adju‐Phos. Additionally, G7‐A1/CdA significantly augmented the frequency of CD4+ and CD8+ Tem cells compared to CdA alone, while G7‐A1/ZMF59 also induced slightly higher CD4⁺ and CD8⁺ Tem responses than ZMF59 (Figure 8B,E). Since Tcm and Tem represent the predominant types of memory T cells in peripheral circulation [68], we further evaluated the overall memory T cell responses. As shown in Figure 8C,F, G7‐A1 showed superior performance to Adju‐Phos in inducing CD4⁺ and CD8⁺ memory T cell responses, while G7‐A1/CdA significantly amplified the antigen‐specific memory T cell responses when compared with CdA alone. The above results suggested that G7‐A1 and G7‐A1/CdA may be more potent in maintaining long‐term immune responses.

FIGURE 8.

Frequency of HPV16‐specific memory T cells determined by flow cytometry. Frequencies of HPV‐specific CD4⁺ Tcm (A), CD4⁺ Tem (B) and total memory T cells (C) in T cells. Frequencies of CD8⁺ Tcm (D), CD8⁺ Tem (E) and total memory T cells (F) in T cells (n = 4–5). Representative FACS gating is shown in Figure S5. The significance was evaluated using a t test or one‐way analysis of variance (ANOVA), as appropriate. Unless otherwise indicated, significance markers denote comparisons against HPV16 L1 VLPs alone, Adju‐Phos or the corresponding single‐component adjuvant control (IMQ or CdA alone). ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

3.9. G7‐A1‐Based Complex Adjuvants Promote the Activation of Dendritic Cells in Draining Lymph Nodes

DCs are considered one of the most functional APCs that can bridge innate and adaptive immunity. Adjuvants can stimulate the innate immune system and trigger the recruitment and activation of APCs, and then APCs loaded with the antigen migrate to dLNs where they play a key role in activating both T and B cell responses [69]. CD86, as a costimulator, is primarily expressed on the surface of APCs such as DCs, macrophages and B cells. When APCs interact with T cells, CD86 binds to CD28 on the surface of T cells, delivering a co‐stimulatory signal that is essential for T cell activation [70]. To explore how nano‐aluminum adjuvants trigger a robust immune response, immunofluorescence assay was used to assess the infiltration of CD11c⁺ CD86⁺ cells in the dLNs of mice immunized with various vaccines. We found significantly higher expression of CD11c⁺ CD86⁺ cells in the G7‐A1 group compared with the Adju‐Phos group at all time points (Figure 9). In addition, when G7‐A1 was combined with IMQ, CdA or ZMF59, the number of CD11c⁺ CD86⁺ cells was significantly higher compared with those in the Adju‐Phos, IMQ, CdA or ZMF59, respectively. These results suggest that the G7‐A1 enhances the proliferation and activation of DCs in dLNs, which can potentially explain the improved antibody levels and stronger cellular immune responses mentioned before following immunization with the nano‐aluminum adjuvanted vaccines.

FIGURE 9.

Immunofluorescence analysis of activated DCs in dLNs of mice at 1, 3 and 7 days after immunization. Representative images of one mouse per group at each time point are shown. Paraffin sections of dLNs were stained with anti‐CD11c antibody (green), anti‐CD86 antibody (red) and DAPI (blue). Scale bar indicates 1 mm.

3.10. In Vivo Safety of G7‐A1‐Based Complex Adjuvants

To assess the safety of nano‐aluminum adjuvants, the body temperature and weight of mice were monitored after immunization with various vaccines (Figure 10A, B and Figure S8). The body weight of mice increased, and no significant difference was found among all groups during 28 days post‐prime. Mouse body temperature in all groups showed limited fluctuation no more than 0.9°C within 4 h post‐immunization. Besides, mice were subjected to histological examination of the injection site at 1, 3 and 7 days post‐injection. As illustrated in Figure 10C, no inflammatory cell infiltration was found in the HPV16 L1 VLP group, and only a slight inflammatory reaction happened at day 1 in the IMQ group. G7‐A1 and CdA groups showed transient inflammatory cell infiltration and peaked at day 3 post‐injection, while MF59 led to mild inflammatory reactions on both day 3 and day 7. In contrast, Adju‐Phos led to remarkable inflammatory reactions at all sampling time points compared with the other adjuvants, characterized by a pronounced accumulation of inflammatory cells at the inoculation site. Inflammatory reactions caused by ZAS04 were mild, but persisted during day 1 to day 7, and similar phenomena were also observed in G7‐A1/IMQ, G7‐A1/CdA, and G7‐A1/ZMF59 groups. The findings indicated that G7‐A1 and its combined adjuvants did not induce any apparent systemic toxicity or severe local adverse reactions. The superior performance of G7‐A1 compared with Adju‐Phos may be due to its smaller particle size, and therefore less local retention and more rapid diffusion and drainage into the lymph nodes.

FIGURE 10.

Safety evaluation of G7‐A1‐based complex adjuvants in mice. (A) The body weight of mice was monitored for 28 days after immunization. (B) The body temperature of mice was monitored at 0, 0.5, 1, 2, and 4 h after immunization. (C) Representative H&E histological images of the muscle at the injection sites. C57BL/6 mice were intramuscularly injected with HPV16 L1 VLP‐adsorbed adjuvants on day 0. Sterile PBS or HPV16 L1 VLPs alone were used as negative controls, and ZAS04 was used as a positive control. On days 1, 3, and 7, mice were euthanized, and the muscle samples at the injection sites were collected for H&E staining (scale bar, 1 mm).

4. Discussion

In this study, we successfully synthesized a series of positively charged nano‐aluminum adjuvants through the microfluidization of low‐molecular‐weight GCS and Adju‐Phos. Compared with previously reported nano‐aluminum, the novelty of this system lies not merely in reducing aluminum adjuvant to the nanoscale, but in combining microfluidic fabrication, cationic surface engineering, and enhanced process robustness within a single formulation platform. The resultant nano‐aluminum adjuvants with GCS/alum ratios higher than 2 showed an excellent batch‐to‐batch reproducibility and gave particle size around 170–180 nm, demonstrating that the microfluidic process provided robust control over particle refinement and dispersion. Antigen adsorption studies revealed that at a 1:20 antigen/adjuvant ratio, the nano‐aluminum system was still stable and achieved an adsorption rate of approximately 50%. Although this adsorption was lower than that typically observed for conventional aluminum adjuvants, it remained substantial and functionally sufficient to support strong immunogenicity, suggesting that complete antigen adsorption was not required for adjuvant efficacy in the present system. It should also be noted that the immunization formulations were administered without removing the non‐associated antigen fraction. Therefore, the in vivo data reflect the biological performance of the final vaccine formulations as administered, rather than the isolated effect of only the particle‐associated antigen fraction. In addition, while PAA‐stabilized microfluidized nano‐aluminum established the value of polymer‐assisted high‐pressure processing, the current GCS‐coated system differs fundamentally in its use of a biocompatible cationic polymer shell, and this distinction is directly relevant to antigen interaction. In the PAA‐coated nano‐aluminum, adsorption of negatively charged antigen was strongly hindered by electrostatic repulsion from the carboxyl‐rich anionic shell [4]. The GCS shell in this study contains protonatable amino groups as well as hydroxyl and hydroxymethyl groups, providing a more chemically diverse interface for antigen interaction. Accordingly, the present results support the view that GCS coating does not simply stabilize the particles, but also creates a more versatile adsorption interface than single‐side‐chain stabilizer shell.

Aluminum‐adjuvanted vaccines require strict maintenance of a 2–8°C cold chain during transportation and storage, and accidental freezing events frequently compromise antigen immunogenicity and cause vaccine loss [7, 71, 72]. One of the major practical contributions of this study is that the GCS‐coated nano‐aluminum formulations exhibited clear resistance to both autoclaving and freeze‐drying under appropriate GCS/alum ratios. Specifically, nano‐aluminum adjuvants with sufficiently high GCS/alum ratios tolerated standard autoclaving without marked physicochemical change, and lyophilized formulations with GCS/alum ratios above 3:1 containing 6% (w/v) trehalose retained particle size and polydispersity under ambient storage conditions for >12 months. Importantly, the in vivo immunization studies in this work were performed using sterilized nano‐aluminum formulations, indicating that adjuvant activity was retained after autoclaving. The excellent freeze‐drying resistance conferred by GCS and trehalose may facilitate future storage and transportation of nano‐aluminum adjuvants. At the same time, the data also clarify the distinct roles of GCS and trehalose. Trehalose has superior hydration properties and ice inhibition capacity relative to other sugars. As a glass‐forming excipient, it reduces the freeze‐concentration of adjuvants, thus minimizing or avoiding aggregation mediated by modifications to the adjuvant's surface chemistry during freezing [73, 74, 75]. In addition, its polyhydroxy structure can form hydrogen bonds with hydroxyl groups on GCS, which could help protect the particles from surface damage. Current evidence does not support GCS alone as an effective lyoprotectant, because aggregation still occurred when trehalose content was inadequate. However, GCS remains indispensable as it prevented nano‐aluminum adjuvant particle aggregation via electrostatic repulsion during rehydration. Besides, GCS can serve as an efficient bulking agent during lyophilization, which has been demonstrated by GCS concentration‐dependent improvement in lyophilized cake appearance. The experimentally supported conclusion is that trehalose is required for effective freeze‐drying protection, whereas GCS contributes critically to maintaining particle integrity during lyophilization and rehydration. Microfluidization with GCS surface coating is critical, as the key challenge in aluminum nano‐engineering extends beyond particle‐size reduction to maintaining the resultant particles in a stable, functional state. Further work will therefore be needed to determine whether the reconstituted lyophilized formulations fully preserve in vivo adjuvant activity.

Effective pathogen clearance typically requires a balanced interplay between humoral and cellular immunity. However, traditional aluminum adjuvants such as Alhydrogel and Adju‐Phos typically induce weak cellular immune responses, a critical limitation for combating intracellular pathogens. In this study, G7‐A1 showed improved immunomodulatory performance relative to Adju‐Phos, as evidenced by stronger antigen‐specific immune responses, more balanced IgG1/IgG2c ratios, and higher levels of selected cytokines. Importantly, these immunological improvements paralleled clear physicochemical changes introduced by GCS coating. Compared with G0‐A1, which remained negatively charged and performed poorly immunologically, increasing GCS content progressively shifted the particle surface to positive zeta potentials, while high‐GCS formulations also improved stability and antibody responses. Therefore, G7‐A1 combined three experimentally demonstrated features: a stable nanoscale structure, a GCS surface shell, and a strongly positive surface charge. Specifically, these changes were associated with enhanced antigen‐specific IgG antibody and nAb responses, a reduced IgG1/IgG2c ratio indicative of more Th1‐skewed immunity, elevated multifunctional T‐cell and memory T‐cell responses, and higher frequencies of CD11c⁺CD86⁺ cells in dLNs than Adju‐Phos. Taken together, the results support a direct relationship between GCS coating, positive zeta potential, and immune enhancement. While the precise intracellular mechanisms require further study, the present work directly demonstrates that modifying aluminum adjuvants with a cationic GCS shell changes the particle interface in a way that is associated with improved DC activation and stronger downstream immune outcomes. Moreover, G7‐A1 induced significantly attenuated inflammatory infiltration compared to Adju‐Phos formulations in safety assessment, indicating that smaller size and surface engineering enhanced localized dispersion‌ and reduced inflammatory cell recruitment [2, 15, 76]. Collectively‌, our findings identify G7‐A1 as a next‐generation nano‐aluminum adjuvant capable of inducing balanced Th1/Th2 immunity and potent memory T cell responses while maintaining favorable safety profiles, positioning it as a promising candidate for pan‐pathogen vaccine development.

Besides physicochemical modification, combining TLR agonists with traditional aluminum adjuvants represents a highly effective strategy for enhancing antigen‐specific immune responses, especially cellular immunity [77, 78, 79]. The combination of aluminum adjuvants with the TLR4 agonist MPL (AS04 adjuvant system) has been successfully implemented in approved HPV and hepatitis B vaccines [77]. Meanwhile, aluminum‐adjuvanted formulations incorporating either the TLR9 agonist CpG 1018 or the TLR7/8 agonist imidazoquinoline‐class compound (represented by IMDG) have been deployed in authorized COVID‐19 vaccines [78, 79]. Prior studies reported that CdA/Adju‐Phos effectively overcame the inherent limitations of conventional alum adjuvants, achieving coordinated enhancement of both humoral and cellular immune responses, though memory T cell responses and long‐term antibody persistence remain uncharacterized [12]. Building on these findings, G7‐A1 was further evaluated as a platform component in complex adjuvant systems containing IMQ or CdA. Both combinations outperformed the corresponding single‐immunostimulant formulations in sustaining antibody responses, amplifying cellular immunity, and promoting memory T cell generation. Furthermore, incorporating CdA into G7‐A1 amplified Th1 and Th2 immunity compared to G7‐A1 or ZAS04 controls. These outcomes are directly supported by the observation that G7‐A1/CdA produced more balanced antibody subclass profiles and more durable neutralizing‐antibody responses, enhanced multifunctional CD4⁺ T cell responses and memory T cell responses, promoted stronger DC activation in dLNs than the corresponding single‐component controls. A more cautious interpretation is that the GCS‐coated nano‐aluminum platform is compatible with additional immunostimulants and can function as an effective co‐formulation scaffold, improving overall immune performance relative to either conventional aluminum or the single immunostimulants used alone. In this sense, G7‐A1 is not only a standalone adjuvant candidate but also a modular materials platform whose positively charged and stable nanoscale interface can support complex‐adjuvant design.

We systematically investigated the combination of G7‐A1 with ‌an MF59‐like adjuvant (ZMF59)‌, which was mediated by electrostatic forces and provided an additional opportunity to assess how the engineered surface properties of G7‐A1 influence formulation compatibility. Emerging evidence confirms that cationic nano‐aluminum particles exhibit electrostatic adhesion to the emulsion surface [80]. The G7‐A1/ZMF59 system showed composition‐dependent changes in particle size and zeta potential, indicating definite physicochemical interaction between the positively charged nano‐aluminum particles and the negatively charged emulsion droplets. On the basis of these measurements, the 4:6 (G7‐A1:ZMF59) formulation was selected because of its smaller size and lower PDI, and this optimized complex adjuvant subsequently showed stronger humoral and cellular immunity than ZMF59 alone. Specifically, G7‐A1/ZMF59 elicited higher antibody titers, enhanced cytokine secretion, increased IFN‐γ‐secreting splenocytes, and elevated memory T‐cell frequencies. In addition, the superior performance of G7‐A1/ZMF59 may partly result from the improved antigen delivery mediated by aluminum and the lipid affinity of the emulsion [10, 81]. The above speculation is also supported by the stronger DC activation mediated by G7‐A1/ZMF59.

In summary, this study demonstrated that the integration of microfluidic fabrication and polymer‐based surface modification effectively addressed key limitations of conventional aluminum adjuvants. The present work established that the novelty of this system lies not only in generating nano‐aluminum, but in producing a GCS‐coated, positively charged, lyophilizable and sterilizable‌ nano‐aluminum adjuvant that remained stable for more than one year at room temperature after lyophilization, whose surface engineering differentiates it from previously reported nano‐aluminum adjuvants. Importantly, G7‐A1 enhanced antigen‐specific immune responses with a more balanced Th1/Th2 profile compared with traditional aluminum adjuvants. This study established a clear link between GCS coating, particle surface charge, and immunological performance. These findings highlighted the value of controlling particle size and interfacial properties in aluminum adjuvant engineering‌. In addition, G7‐A1 remained compatible with ZMF59 or CdA, and these combinatorial formulations further improved immune responses compared to the corresponding single‐component controls, while maintaining comparable local reactogenicity. These findings highlight the engineering significance of G7‐A1 as a thermostable, lyophilization‐compatible and adaptable nano‐aluminum adjuvant platform for next‐generation vaccine formulation. To facilitate clinical translation, future investigations will focus on mechanistic studies to elucidate the adjuvant's immunomodulatory pathways and potential cross‐talk with innate immune signaling networks. Collectively, our findings validate G7‐A1‐based formulations as clinically translatable candidates that may redefine vaccine design paradigms through rational engineering of multifunctional adjuvant systems.

Author Contributions

Conceptualization, Y.Z.; methodology, Y.Z., L.C., J.Z., L.B., X.Z., D.L., J.X., M.S., G.L., X.W., J.Z., L.X., B.S., C.J., and Y.C.; formal analysis, Y.Z., L.C., J.Z., L.B., X.Z., D.L., J.X., M.S., G.L., X.W., J.Z., L.X., B.S., C.J., and Y.C.; writing – original draft preparation, Y.Z., L.C., J.Z., L.B., and D.L.; writing – review and editing, Y.Z. and L.C.; supervision, Y.Z. and Y.C.; research funding acquisition, Y.Z.; project administration, Y.Z. All authors have read and agreed to the published version of the manuscript.

Funding

This study was partly funded by the National Key Research and Development Program of China (Grant No. 2025YFC2311600).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adhm71359‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data presented in this study are available upon request from the corresponding author.