Elucidating the porous structure of aluminum adjuvants via in-situ small-angle scattering technique

Abstract

Aluminum-based adjuvants are widely used in vaccine formulations due to their immunostimulatory properties and strong safety profile. Despite their effectiveness and safety, the exact mechanisms by which they enhance vaccine efficacy remain unclear. One proposed mechanism is that aluminum adjuvants form a depot that gradually releases antigens, thereby improving antigen uptake by antigen-presenting cells. This study investigates the porous structures of two commonly used aluminum adjuvants, aluminum hydroxide (AH) and aluminum phosphate (AP), using small-angle X-ray scattering (SAXS) and small-angle neutron scattering (SANS). Our measurements reveal that AH nanoparticles, with their needle-like morphology, form smaller, interconnected pores within the aggregated architecture. In contrast, AP nanoparticles, with a plate-like shape, forms more discrete, isolated porous structures. Both adjuvants have pore sizes within the range of commonly used vaccine antigens, supporting the depot theory. Our findings also reveal that antigen retention is prolonged when the antigen size is comparable to the average pore size of the adjuvant. This study highlights the utility of SAXS and SANS for in-situ characterization of adjuvant porosity and provides insights into how nanoparticle morphology affects antigen retention and release. By elucidating these structural details, our research underscores the importance of porous structure in adjuvant function and offers potential pathways for improving vaccine formulations through tailored adjuvant design.

INTRODUCTION

The discovery of vaccines stands as one of the greatest milestones in the history of medicine¹. Vaccines primarily aim to induce a protective immune response against pathogenic organisms, which is achieved by introducing antigens into the body². Antigens can be in the form of live or attenuated bacteria and viruses, as well as inactivated versions of these pathogens³. However, the use of such antigens carries the inherent risk of reversion and subsequent infection⁴. To mitigate this risk, subunit vaccines have been developed and employed⁵. Subunit vaccines utilize specific fragments of pathogens that possess immunogenic properties as antigens. These fragments can consist of amino acids, polysaccharides, nucleic acids, or proteins of varying sizes⁶. Despite their advantages, one notable drawback of subunit vaccines is their limited immunogenicity, which means they may not generate sufficient immune responses to provide complete protection⁷. To address this challenge, adjuvants are included in the formulation to enhance the effectiveness of subunit vaccines by boosting immune responses^{8, 9}. The first adjuvant invented for use in tetanus and diphtheria vaccines was aluminum salt^{10, 11}. While numerous other adjuvants have been researched for potential use, aluminum-based adjuvants have proven to be highly effective and possess a longstanding safety record spanning several decades¹². Among the commercially available aluminum-based adjuvants, aluminum hydroxide (AH) and aluminum phosphate (AP) are the most commonly used.

Since the discovery of aluminum adjuvants, significant research has been dedicated to elucidating the mechanisms underlying their adjuvant effect^{13, 14}. Early studies highlighted the importance of antigen adsorption to aluminum adjuvants before injection, a process crucial for promoting antigen uptake by antigen-presenting cells (APCs)^{15, 16}. This phenomenon laid the foundation for the “depot theory,” proposed by Glenny et al.¹⁷, which is now widely accepted as a plausible explanation for the adjuvant action of aluminum salts. According to this theory, aluminum adjuvants form antigen depots that allow for a controlled release of the antigen, giving immune cells sufficient time to activate and respond effectively¹⁷. Following the WHO recommendations that at least 80% of antigens in vaccines containing aluminum adjuvant should be adsorbed¹⁸, research has intensified into the interaction mechanisms between the aluminum-based adjuvant and antigens, particularly the roles of pH¹⁹, ionic strength²⁰, point zero charge (PZC) of the aluminum adjuvant, and the isoelectric point of the antigens^{21, 22}. These studies seek to understand how electrostatic interactions between the aluminum adjuvant and antigens contribute to a regulated and comparatively prolonged binding of the antigen, preventing it from dissolving in body fluids after administration^{15, 23–28}. Such slow release allows immune cells sufficient time to activate cytokines, subsequently activating dendritic cells²⁹. Building on this foundational understanding, current research efforts are now focused on a more detailed examination of the physical and chemical properties of aluminum adjuvants, aiming to tailor adjuvant structures for optimized vaccine efficacy^{27, 30, 31}.

To characterize the physical and chemical properties of aluminum adjuvants, previous investigations have primarily focused on elucidating the relationship between the morphology, crystallinity, and hydroxyl content of these adjuvants and their role in augmenting antigen-specific immune responses^32–34. Research has shown that AH and AP are hydrated nanoparticles with distinct morphologies—AH exhibits a needle-like structure, while AP has a plate-like structure^{35, 36}. Although AH and AP are individual nanoparticles, they are known to spontaneously form aggregates that can reach micron sizes, which are the actual functional forms of adjuvants in solution^{35, 36}. These aggregates create porous structures that potentially offer sites for antigen adsorption and retention, contributing to the slow release of antigens and prolonged immune stimulation. Understanding the porosity of these aggregated aluminum adjuvants is crucial for controlling the release of antigens. Traditionally, the porous structure of materials has been analyzed using electron microscopy and nitrogen adsorption, which require drying the samples and can damage the hydrated shell and overall structure of the aggregates^37–40. Therefore, the ability to characterize the porous structures within aluminum adjuvants in-situ is crucial for a comprehensive understanding of antigen/adjuvant interactions. Achieving this understanding will allow researchers to fine-tune antigen binding and release mechanisms, thereby optimizing vaccine efficacy based on detailed knowledge of the porous structure.

In this study, we utilized small-angle scattering (SAS) to characterize the microstructure of two commercial aluminum adjuvants—Alhydrogel (AH) and Adjuphos (AP)—in their native solution states, focusing on how AH and AP nanoparticles aggregate into porous structures and the impact of these structures on protein binding. Our results reveal that AH and AP nanoparticles not only differ in morphology but also form aggregates with varying porosities. Detailed analysis of the scattering profiles collected from small-angle X-ray scattering (SAXS) technique shows that there are significant differences in size, surface roughness, and the arrangement of the porous structures within the adjuvants for both AH and AP, suggesting that the distinct morphologies of AH and AP nanoparticles can lead to their different aggregation behaviors and subsequently, the porosities within the aggregated structures. The small-angle neutron scattering (SANS) results, though not providing additional structural information beyond SAXS, validates the observed features in SAXS profiles are due to the presence of porous structure within the aluminum nanoparticle aggregates. Furthermore, SAXS data collected from adjuvant/antigen complexes at various ratios reveal that antigen proteins are adsorbed within the porous structures of the adjuvants. The size disparities between the antigens and pores play a significant role in the retention of these adsorbed antigens. These findings advance our understanding of the porous structure of aluminum adjuvants in solution, addressing a notable gap in current research. Ultimately, our results emphasize the importance of evaluating and modulating the porous structure of aluminum adjuvants for the design and development of vaccines, aiming to enhance their immunogenicity and efficacy.

MATERIALS & METHODS

Materials.

In this study, commercial aluminum adjuvants, Alhydrogel^® and Adju-Phos^®, were obtained from InvivoGen (San Diego, CA, USA) for further characterization. Alhydrogel^® primarily consists of aluminum hydroxide, while Adju-Phos^® is composed of aluminum phosphate. For ease of reference within this manuscript, Alhydrogel^® and Adju-Phos^® is referred to as AH and AP, respectively. Bovine serum albumin (BSA, Cat. No. BP9701–100) was procured from Fisher Scientific (Hampton, NH, USA), lysozyme from MP Biomedicals (Cat. No. 100831), and beta-lactoglobulin (BLG, Cat. No. L3908–1G) was sourced from Sigma Aldrich (St. Louis, MO, USA). The National Institute of Standards and Technology (NIST) monoclonal antibody reference material (NISTmAb, reference material RM 8670) was received as 10 mg/mL protein solution prepared in 25 mM histidine buffer at pH 6. All other chemicals were of analytical grade and used as received without further purification.

SAXS measurements.

SAXS measurements were performed at the 7A1 beamline of the Cornell High Energy Synchrotron Source (CHESS) located in Ithaca, NY. Aluminum adjuvant samples were prepared at a concentration of 2 mg/mL in a 10 mM HEPES buffer. Adjuvant/protein complexes were also measured by SAXS. To ensure complete saturation of the adjuvants and to prevent the presence of unbound proteins in the solution, adjuvant/protein complexes were prepared according to the binding capacity of each adjuvant. For SAXS analysis, a 50 μL aliquot of each sample was placed into a flow cell containing a 1.5 mm inner diameter quartz capillary with wall thickness of 10 μm. The energy of the X-ray utilized is 11.3 eV. To minimize radiation damage during data acquisition, the fluid sample was gently oscillated. Scattering data were recorded across a q-range of 0.009 ~ 0.5 Å⁻¹. For each sample, a series of ten scans were taken, each with an exposure time of one second. The collected data were then averaged using the BioXTAS RAW software⁴¹. Data analysis was subsequently carried out with the ATSAS package in conjunction with the NCNR analysis macros embedded within IgorPro42, 43.

SANS measurements.

SANS experiments were conducted at the Bio-SANS beamline (CG-3) at the High Flux Isotope Reactor (HFIR) at Oak Ridge National Laboratory (ORNL) located in Oak Ridge, TN. The aluminum adjuvant samples were diluted to match the concentration used for SAXS measurements, using a 10 mM HEPES buffer made with 100% D₂O to minimize incoherent background signals caused by hydrogen in the high-q region. Prepared adjuvant samples were loaded into Banjo cells with a 2 mm path length. To ensure an even distribution of adjuvant samples during SANS measurements, the cells were gently rotated in-situ in the radial direction using a tumbling sample stage. All measurements were performed using one of two different configurations of the Bio-SANS instrument. First configuration, with the main detector at 15.5 m and the curved wing detector rotation at 1.4°, provided a wide q-range of 0.003 Å⁻¹ to 0.8 Å⁻¹ but lower-flux, while the second covered a smaller range of q of 0.007 Å⁻¹ to 0.8 Å⁻¹ but with higher flux by placing the main detector at 7 m and the curved wing detector rotation at 3.2°. All measurements were performed with a neutron wavelength of 6 Å. Data reduction was performed using the facility-developed software, drtsans, to obtain the scattering intensity I(q) on an absolute scale as a function of q⁴⁴. The reduced data from the experiment was subsequently analyzed by using the NCNR analysis macros embedded within IgorPro software pacakge^{42, 43}.

Scattering Data Analysis.

During the process of a SAS measurement, suspended aluminum adjvuants particles are subjected to a well-collimated beam, which could be either X-rays in the case of SAXS or neutrons in the case of SANS experiments. The beam used has a specific wavelength, denoted by λ. The scattered intensity I(q) is recorded as a function of scattering vector q. The scattering vector q is expressed as 4πsinθ/λ, where 2θ is the scattering angle between incident and scattered beam. For a system containing monodisperse, homogeneous and isotropic dispersion of spherical particles, I(q) can be expressed as:

$$I(q)=\varphi V_p{(\Delta \rho )}^{2}P(q)S(q)$$ Eq. 1

where ϕ is the volume fraction of scattering particle with volume V_p, Δρ is the difference in scattering length densities (SLD) between the particles being measured and the solvent background. P(q) and S(q) are the form factor and structure factor of the scattering particle, respectively. In SAS techniques, SLD is a critical parameter that quantifies how a material scatters X-rays or neutrons. X-ray SLD (xSLD) is primarily determined by the electron density of the material, as X-rays interact with the electron cloud surrounding the nuclei; regions with higher electron densities scatter X-rays more strongly^{45, 46}. In contrast, neutron SLD (nSLD) is determined by the nuclear scattering lengths of the atoms within the material. Neutrons interact directly with atomic nuclei, making their scattering dependent on the specific nuclear properties of the atoms^{47, 48}, particularly noteworthy in the case of hydrogen and deuterium. The significant difference in the nSLD values of hydrogen and deuterium makes neutrons especially useful for studying hydrated materials. The scattering vector q is inversely proportional to the real-space distance d:

$$q=\frac{2\pi }{d}$$ Eq. 2

Therefore, the low-q region of the scattering profile measures structures that are large in size, whereas the high-q region measures structures with smaller length scale. In this research, two-power law model: a low-q power law regime and a high-q power law regime, is applied to fit the SAXS profiles obtained from aluminum adjuvants to conduct a comprehensive analysis of aluminum nanoparticles in their solution state⁴⁹. Generally, power-law scattering is expressed as:

$$I(q)\propto q^{-P}$$ Eq. 3

The exponent P in the low-q region provides information on the fractal network formed by the structure, while the exponent in the high-q regime provides insights into the surface characteristics of the structure. The Igor Pro software packages were used for this analysis, enabling us to characterize both the low-q and high-q power law behaviors of the scattering profile. The crossover point q^*, where the low-q and high-q power law regions intersect, is indicative of a characteristic length scale within the analyzed sample, i.e. the average size of a measurable structure within the sample.

Elution of bound protein in response to decreasing pH.

Adjuvant/protein complexes were prepared at certain aluminum to protein ratio r to ensure that a minimum of 80% of the added protein was bound to the adjuvants. Adjuvant/protein pairs chosen for this study include AH/BSA, AH/BLG, AP/NISTmAb, and AP/lysozyme. At first, adjuvant/protein complexes were prepared at around pH 7 to promote electrostatic interactions between the adjuvant and protein. These complexes were stored at 4 °C overnight to ensure the thorough binding between adjuvant and protein molecules. Five aliquots were then taken from the adjuvant/protein complexes after overnight incubation, and each of them was subjected to pH adjustment so that the final pH of the solution was in the range from 3–7. After pH adjustment, each sample was subjected to centrifugation to separate the adjuvant/protein complexes (as pellets) and unbound protein (in the supernatant). The protein concentration of the supernatant was measured using a UV-Vis spectrophotometer. The molar extinction coefficient of BSA, BLG, NISTmAb, and lysozyme at 280 nm is 43,824 cm⁻¹M⁻¹, 17,600 cm⁻¹M⁻¹, 213,000 cm⁻¹M⁻¹, and 38,940cm⁻¹M⁻¹ respectively^50–53. The percentage of protein remaining bound to the adjuvant was then calculated using Eq. 4.

$$\%\mathit{\text{Bound}}=\frac{M_{\mathit{\text{add}}}-M_{\mathit{\text{supernatant}}}}{M_{\mathit{\text{add}}}}\times 100\%$$ Eq. 4

where M_add represents the amount of protein initially added, and M_supernatant represents the amount of unbound protein that remained at the supernatant.

RESULTS

SAXS measurements reveal the porous architecture within the aggregated networks of aluminum adjuvant nanoparticles.

In this study, we performed SAXS measurements to investigate the microstructures of aluminum adjuvants. According to Eq. 2, the low-q region of the scattering profile reflects larger structures, while the high-q region captures structures with a smaller length scale. The scattering profiles for both AH and AP exhibit two distinct power-law regions, as illustrated in Figure 1. Scattering in the low-q region for both adjuvant nanoparticles displays an upturn in I(q), indicative of interparticle interactions. Such a result is not surprising since aluminum adjuvant nanoparticles are known to form aggregated networks. The transition point between the two discrete power-law regions is represented as q^* for both adjuvant nanoparticles. The position of q^* suggests a repeating distance within the sampled structures. Given that this crossover occurs at 0.095 Å⁻¹ for AH and 0.056 Å⁻¹ for AP, the repeating distances can be estimated at approximately 6.6 nm and 11.2 nm, respectively, using Eq.2. Prior research has indicated that the average sizes of both AH and AP nanoparticles far exceed these measured repeating distances, reaching around 50 nm⁵⁴. Hence, the emergence of q^* likely results from structures significantly smaller than the individual AH and AP nanoparticles. This size discrepancy suggests that the measured repeating distances of 6.6 nm and 11.2 nm could stem from the presence of void spaces within the aggregated networks formed by the AH and AP nanoparticles. In addition to the q^* feature, scattering profiles measured from AH and AP both demonstrate two power law regions with distinct exponents. The power law region in the high-q (q > 0.1 Å⁻¹) region corresponds to surface scattering from the scattering object⁵⁵, in this case, the pores within the aggregated network, whereas the power law region in the low-q region (q < 0.08 Å⁻¹) corresponds to the arrangement of the pores within the aggregated network⁵⁶. In scattering profiles measured from AH, the high-q exponent was 3.5, suggesting that the pores within the aggregated AH network have a relatively rough surface. The high-q exponent was measured to be 4 for AP, suggesting that the pores formed within the AP network have smooth surfaces. The fractal dimension, which provides insight into the connectivity of the pores for both adjuvants, is related to the power law exponent in the low-q region (q < 0.08 Å⁻¹)⁵⁵. The fractal dimension of AH and AP was determined to be 2.2 and 3, respectively (Figure 1). The fractal dimension measured from AP suggests that the pores within AP aggregates were densely packed, forming a more compact structure. In contrast, the pores in AH aggregates exhibited a loosely connected arrangement, possibly forming a network of channels within the AH structure^57–59.

Figure 1.

SAXS profiles measured from AH (a) and AP (b) samples prepared in 10 mM HEPES buffer at pH 7.4. Dotted lines represent the two-power law model used to fit the experimental data.

SANS measurements confirm the presence of nanometer-sized pores within native aluminum adjuvant samples.

SAXS and SANS both provide valuable structural insights at the nanometer length scale (from a few to hundreds of nanometers), but they reveal different structural entities. SAXS measurements depend on interactions with electrons, while SANS relies on interactions with atomic nuclei, resulting in distinct scattering contrasts compared to X-ray measurements^{60, 61}. Consequently, SAXS and SANS offer complementary structural information⁶². Building on our SAXS results, we proceeded to conduct SANS measurements on our adjuvant samples. Our SAXS analysis revealed unique crossover points in the intermediate q-range for both AH and AP samples. However, these crossover points were absent from the SANS profiles of the same samples (Figure 2), suggesting that the porous structures were either not present or not visible to neutrons. According to Eq. 1, the scattering intensity of a sample depends on various factors, including the scattering length density (SLD) difference between the scattering object and the buffer background (Δρ). Since AH and AP were prepared from the same stock for both SAXS and SANS measurements, it is unlikely that the porous structures would suddenly disappear when switching from hydrogenated (for SAXS measurements) to deuterated buffer (for SANS measurments). To better understand these differences, we compared the SLD values of aluminum adjuvants and the buffer background. The X-ray SLD (xSLD) values for AH and AP were estimated to be approximately 21.25 × 10⁻⁶/Ų and 21.72 × 10⁻⁶/Ų, respectively, while the xSLD value of the H₂O buffer used for SAXS measurements was about 9.47 × 10⁻⁶/Ų.⁶³ Such significant difference in xSLD values between the solvent-filled pores and the adjuvants allowed the porous structures to contribute to features in the SAXS scattering profiles (Figure 2). For SANS measurements, our previous research found the neutron SLD (nSLD) values of AH and AP in D₂O buffer to be approximately 4.56 × 10⁻⁶/Ų and 6.13 × 10⁻⁶/Ų, respectively, while the nSLD value of D₂O was around 6.36 × 10⁻⁶/Ų.⁶⁴ Compared to the xSLD values, the difference in nSLD values between both adjuvants and the deuterated buffer is small, suggesting that the solvent-filled pores in AH or AP may not be visible to neutrons, the crossover point in the intermediate q-range was not observed from SANS profiles. Although SANS did not reveal additional structural details about the porous structures within the aluminum adjuvants, the absence of the crossover point in the SANS profiles supports our SAXS-based hypothesis that the crossover point arises from the presence of porous structures in the native aluminum adjuvant samples.

Figure 2.

SAXS and SANS profiles measured from AH and AP samples. The illustrations demonstrate the contrast between aluminum adjuvants and their corresponding buffer backgrounds in H₂O (for SAXS) and D₂O (for SANS). The solvent-filled pores and aggregated aluminum adjuvant structures are color-coded based on their respective xSLD and nSLD values. Due to the small difference in nSLD values between the solvent and adjuvants in SANS measurements, the pores are less visible to neutrons compared to X-rays.

Porous structures within aluminum adjuvant aggregates as depot for proteins.

Scattering results lead us to hypothesize that AH and AP nanoparticles could form aggregates with varied pore sizes around the 10 nm size range. Following this hypothesis, we expect that the electrostatic interactions between the model protein molecules and the pore surfaces will lead to the occupation of these pores by the proteins, thereby altering the distribution of the porous structure within the aggregated adjuvant networks. Such structural changes could alter the scattering profile, especially affecting the position or magnitude of the crossover point q^*, which emerges from the porous structure. To validate our hypothesis, we prepared a series of adjuvant/protein complexes at different protein to adjuvant mass ratios (r) and measured them using SAXS (Figure 3). SAXS profiles collected from AH/BSA complexes with various r values are presented in Figure 3a to demonstrate the transition of the internal structure of the complexes as more BSA was added into the system. One noticeable change among various scattering profiles is that the magnitude of q^* peak became less pronounced with increasing r values, suggesting that the amount of pores within the nanoparticle aggregates were reduced, possibly due to protein deposition within these pores. Different from the AH/BSA system, scattering profiles from AP/lysozyme complexes (Figure 3b) demonstrate a different trend where the crossover points q^* remains unchanged across different complexes.

Figure 3.

SAXS profiles measured from adjuvant and adjuvant/protein complexes prepared at varying r values: AH/BSA (a), AP/lysozyme (b) and AP/NISTmAb (c). Arrows in the figure highlight the disappearance of q* with increasing r values observed from AH/BSA and AP/NISTmAb systems. Illustrations demonstrates the size differences between model proteins and porous structures within different aluminum adjuvants.

In order to better understand the change in scattering profiles, we also measured the binding capacities of BSA to AH and lysozyme to AP: approximately 4.5 mg of BSA binds to 1 mg of aluminum in AH, and 0.7 mg of lysozyme binds to 1 mg of aluminum in AP (Supporting Information Figure S1). Scattering data in Figure 3 indicates that porous structures within AH was largely diminished when complexes achieved an r value of 2.3 (about half of the maximum binding capacity of BSA to AH). In contrast, the porous structures within AP aggregates persisted even at an r value of 1.14, surpassing the maximum binding capacity of lysozyme by AP. To interpret these findings, we compared the sizes of BSA and lysozyme molecules to the measured pore sizes within AH and AP aggregates. The radius of gyration for BSA in HEPES buffer is approximately 2.9 nm (SASBDB⁶⁵ entry: SASDA32), comparable to the averaged pore size within AH aggregates, which is at 6.6 nm. Such similarity in size might explain the disappearance of q^* in the scattering profiles measured from complexes prepared with increasing r values. As more proteins diffused into the porous structure within AH aggregated networks, BSA molecules fully occupied the pores, leaving minimal free space (Figure 3a). Different from the AH/BSA system, scattering profiles measured from AP/lysozyme complexes suggest porous structures remained present even after the maximum binding capacity of lysozyme was met. This observation could be attributed to the significant size difference between lysozyme (radius of gyration around 1.5 nm in HEPES buffer, SASBDB⁶⁵ entry: SASDA96) and the average 11.2 nm pore size within AP aggregates (Figure 3b). Therefore, although lysozyme molecules can deposit into the surface of the porous structures within AP aggregated networks, these pores could still appear vacant due to the smaller size of lysozyme molecules.

Size difference between the porous structure and protein determines the internal structure of aggregated adjuvant network after protein-binding.

Considering the size difference between the porous structures within adjuvants and the model proteins, proteins residing inside the pores are anticipated to experience varying environments. The average pore size within AH is comparable to that of BSA molecules. Thus, when BSA molecules diffuse and bind to the pores formed by aggregated AH nanoparticles, the proteins largely occupy the void space within the pores, resulting in a less solvated environment. In contrast, the average pore size within aggregated AP nanoparticles is significantly larger than that of lysozyme. Consequently, even as proteins diffuse into and interact with the pore surfaces, the pores remain substantially unfilled, providing a more open environment to the external solution. In the same vein, we also performed SAXS measurements on complexes formed between AP and NISTmAb at various protein to adjuvant ratios. We chose NISTmAb because it is larger than lysozyme with a reported radius of gyration of around 5 nm⁶⁶, aligns with the average pore size within AP aggregated networks. The isoelectric point of NISTmAb is at pH 9.3⁶⁷, therefore, the protein molecule is positively charged in HEPES buffer at pH 7.4, and therefore the interactions between NISTmAb and AP were mainly driven by electrostatic interactions. While other interactions, such as van der Waals forces, hydrogen bonding, and hydrophobic interactions may be present, their contributions are minimal in the context of Adjuvant/protein interactions. We expect that with NISTmAb molecules occupying the porous structures within aggregated AP networks, the crossover point observed in scattering profiles will disappear, as noted in the AH/BSA case. SAXS profiles from AP/NISTmAb complexes (Figure 3c) reveal that the magnitude of the crossover point becomes less pronounced with the addition of more NISTmAb molecules to the system, suggesting that the presence of the porous structures was reduced as more NISTmAb molecules filled the pores within the AP adjuvant. In summary, the scattering profiles from adjuvant/protein complexes confirm that the structural features observed at c.a. 10 nm length scale in aluminum adjuvants are attributable to the presence of porous structures within the networks formed by aluminum adjuvant nanoparticles. Furthermore, different adjuvants display varying porosity, which dictates the internal structure of aggregated adjuvant network after protein-binding.

Effect of porosity of aggregated adjuvant network on the retention of antigen molecules.

We have demonstrated that the amount of porous structures within the aggregated adjuvant network can be reduced when proteins of similar sizes are deposited within the pores. It is intuitive to think that in systems with a substantial size disparity between the protein and the porous structures, such as the AP/lysozyme system, the porous structures would likely remain relatively open. This is because the adsorbed lysozyme is small compared to the average pore size within the AP network, leading to proteins being situated in a relatively open and hydrated environment. Conversely, in systems where the protein and the porous structures are of similar sizes, such as AH/BSA and AP/NISTmAb, the proteins are expected to fill most of the pore space, resulting in a more confined environment that limits solvent exposure. Consequently, we hypothesize that the size discrepancy between the porous structures and proteins is a crucial factor determining the environment of the adsorbed proteins within the porous network formed by aluminum nanoparticles. To verify our hypothesis, we measured protein retention within the porous structures under varying solution pH levels. We anticipate that if the pores remain significantly open, any changes in solution pH would induce corresponding pH shifts within the porous structure. Conversely, if the porous structures are fully occupied by protein molecules, with limited exposure to the solvent, then changes in the external solution pH will not affect the pH within the porous structures.

To this end, we have measured the amount of protein that remained bound to the adjuvant as the solution pH decreased from 7.4 to 3 (Figure 4). Four different adjuvant/protein pairs were prepared, including AH/BSA, AH/BLG, AP/lysozyme, and AP/NISTmAb. In systems containing AH, two different proteins were used to create the complexes: BSA, which is of a size comparable to the porous structure within the AH network, and BLG, which is significantly smaller than BSA (Rg = 2.4 nm⁶⁸). For systems containing AP, two different proteins were also utilized to form the complexes: lysozyme, which is significantly smaller than the pore size within the AP network, and NISTmAb, which is of a size comparable to the porous structures. By comparing the retention of protein in different systems with pH changes, we can gain a better understanding of the role porosity plays in determining the microstructure of adsorbed proteins experience as they are located within the porous structures. Results shown in Figure 4 suggest that among the four adjuvant/protein systems, the AH/BSA and AP/NISTmAb pairs exhibit relatively stable protein retention across the examined pH range, with only slight decrease in protein binding as the pH is decreased from 7.4 to 3. Given that the sizes of the porous structures from AH and AP are similar to those of BSA and NISTmAb, respectively, these results imply that the proteins bound within the porous structures of adjuvants are situated in a confined environment, which reduces their exposure to the external solution. Conversely, the adjuvant/protein pairs with larger size differences, namely AH/BLG and AP/lysozyme, demonstrate significant decreases in protein retention with decreasing pH. This suggests that proteins within the porous structures are in a more open and hydrated environment, making them more prone to pH-induced desorption from the porous structures. Therefore, we can conclude that the size of the protein relative to the pore size plays a crucial role in defining the environment in which proteins are exposed to as they reside within the porous structures. It is the degree of this confinement that governs the extent of solvent exchange and protein retention when the buffer conditions are altered from the external environment.

Figure 4.

Effect of pH variation on the percentage of protein remaining bound to adjuvant nanoparticles. Datasets are colored differently to represent different adjuvant/protein pairs. Dotted lines are drawn between data points to guide the eyes. Error bars represent the standard deviation among three replicates.

To determine if changes in pH influence the porous structure of aluminum adjuvant aggregates and subsequently affect the antigen retention, we performed in-situ SAXS measurements on both AH and AP prepared under various pH conditions. Scattering profiles were recorded across a q-range of 0.009–0.5 Å⁻¹ for each adjuvant at different pH levels. These profiles were compared to the scattering profile of the adjuvant sample prepared at pH 7.4. As shown in Figure 5, the scattering profiles measured from both adjuvant samples across different pH values are consistent, indicating negligible change in the porous structure. This suggests that the porous architecture adopted by the aggregated adjuvant nanoparticles remains unchanged despite variations in solvent pH. Therefore, the observed changes in antigen retention are not due to modifications in the porous structure within the adjuvants but must be attributed to other factors unrelated to the structural integrity of the adjuvant aggregates.

Figure 5.

SAXS profiles measured from AH (a) and AP (b) at various pH conditions to assess the impact of solvent pH on the porous structure of aluminum adjuvants.

DISCUSSION

Understanding how aluminum salts function as adjuvants is fundamental to vaccinology and immunology. Glenny et al. introduced the ‘depot theory’ to describe the mechanism of action of aluminum-based adjuvants, suggesting that these salts enhance immunogenicity by promoting the sustained release of antigens through adsorption^{17, 69}. This slow release results in prolonged antigen exposure following injection, leading to improved immunization⁷⁰. Subsequent research has supported these findings, showing that antigen adsorption onto aluminum salts creates a high local concentration, which enhances uptake by antigen-presenting cells (APCs)^70–72. Collectively, these studies underscore the importance of depot formation in strengthening the immune response induced by aluminum-adjuvanted vaccines.

It is widely recognized that aluminum adjuvants are nanoparticles that can form large aggregates with porous structures²⁷. It is intuitive to think that the antigens can not only bind to the surface of aluminum nanoparticles but can also diffuse and stay in the porous structures within their aggregates. As a result, it is anticipated that these porous aggregates could play a crucial role in the depot formation for antigens, a concept that is central to enhancing vaccine efficacy. Our research aims to explore this fundamental characteristic of aluminum adjuvants, which could pave the road for the control and customization of adjuvant aggregates in vaccine formulation. Despite its importance, the in-situ characterization of the porous structure of aluminum adjuvants presents a significant challenge. Traditional characterization methods like scanning electron microscopy (SEM), transmission electron microscopy (TEM), and atomic force microscopy (AFM) are extensively used to characterize porous materials. However, these characterization methods require prior preparation of the samples which can change the arrangements of aluminum nanoparticles and thus the porous structures within them^73–75. Consequently, there is a pressing need for robust and reliable characterization tools capable of probing the microstructure of aluminum adjuvants in their natural, hydrated state to ensure accurate characterization of their porous structure and aggregate formation.

In this study, we demonstrated the use of SAXS and SANS for the in-situ characterization of the porous structures within two commonly used aluminum adjuvants: AH and AP. The SAXS experiments provided crucial information about the porosity of these adjuvants, including the size and arrangement of the pores within the aggregates. We found that the average pore size within AH is approximately 6.6 nm, while the pores in AP nanoparticles average around 11.2 nm. The observed difference in pore size between AH and AP may be related to their respective morphologies. AH has a needle-like shape, resulting in a higher surface area compared to spherical particles of similar volume. This increased surface area facilitates closer packing and smaller interstitial spaces within AH aggregates. In contrast, AP nanoparticles have a plate-like morphology, which leads to less efficient packing and larger void spaces within AP adjuvant aggregates. To assess this, we analyzed newly purchased adjuvant samples and compared them to aged samples from the same batch, stored for less than one year—within the manufacturer-recommended shelf life. The results, shown in Supporting Information Figure S2, confirm that the porous structure remains stable in both fresh and aged samples from the same batch. These findings demonstrate the reproducibility of the porous features under controlled preparation and storage conditions, underscoring that the morphology of the NPs significantly impacts the average pore size. SAXS results indicate that the porous structures within AH and AP adjuvants are in the size range of approximately 10 nm. Since antigens in subunit vaccines typically fall within a similar size range^76–78, we hypothesize that the size difference between the pores and antigens could be a key factor influencing the depot effects of these adjuvants.

The formation of an antigen depot involves several key steps, including trapping antigens within adjuvants and gradually releasing them into interstitial fluid¹². Our results show that antigen molecules, such as proteins, can infiltrate the porous structures within aluminum adjuvant nanoparticles. In this study, we investigated several model antigens, including BSA, BLG, lysozyme, and NISTmAb, which vary in size and charge properties. Our results reveal a clear relationship between the size of these antigen and the pore size of the adjuvant aggregates. Importantly, while this study was limited to relatively small protein antigens, the findings suggest that this relationship may extend to larger antigens, such as virus-like particles (VLPs). However, it would be necessary to engineer aluminum nanoparticles with specific shapes and sizes, resulting in aggregates with pore dimensions that align with the physical characteristics of larger antigens. This insight opens exciting opportunities for adjuvant formulation engineering, where tailoring the size and morphology of primary nanoparticles could enable the design of adjuvants optimized for specific antigen types. For example, by manipulating the reaction conditions during nanoparticle synthesis—such as temperature or reactant concentrations—larger primary nanoparticles could be produced, leading to aggregates with larger pore sizes suitable for macromolecular antigens like VLPs

Furthermore, our study highlights that the extent to which antigens occupy these pores is largely determined by the size difference between the pores and the antigens: larger size discrepancies result in greater unoccupied spaces. This creates a more open, solvated environment upon diffusion, which can reduce antigen retention due to the rapid exchange with the solvent, potentially leading to dissociation of antigens from the adjuvants upon injection. SAXS analysis further reveals that AH nanoparticles form tightly packed aggregates with smaller, rough-surfaced pores that are interconnected, creating channel-like features (Figure 2). In contrast, AP nanoparticles form more loosely connected aggregates with isolated, smoother pores, resulting in a more open and permeable network (Figure 2). While not directly examined in this study, the arrangement of these porous structures is anticipated to impact antigen retention, with more open networks facilitating easier antigen escape. Therefore, both the size and arrangement of porous structures play a crucial role in how antigens diffuse and are retained within adjuvants. The ability to measure and control these porous structures is crucial for more effective vaccine development, as the native porosity of adjuvants is closely related to their role as antigen depots.

CONCLUSION

In this study, we performed SAXS and SANS measurements to investigate the porous structures within native aluminum adjuvants in solution, focusing specifically on AH and AP, the two most commonly used licensed vaccine adjuvants. Our measurements revealed distinctive features in the scattering patterns, indicating differences in the porous structures of AH and AP. AH, with its needle-shaped morphology, exhibited a smaller average pore size compared to the plate-shaped AP. We found that the shape of the nanoparticles significantly influences their aggregation behavior and, consequently, the size and arrangement of the resulting porous structures. Despite variations in nanoparticle shape, both adjuvants had pores within the size range of commercial vaccine antigens, with an average pore size of approximately 10 nm. Our findings suggest that the porous structure of aluminum adjuvants functions as an antigen depot, facilitating controlled release and retention of antigens. We observed that model antigen proteins can diffuse into the porous networks of both AH and AP, with differences in retention attributed to the size disparities between the antigens and pores, and potentially the interconnectivity of the porous network. These insights provide opportunities for tailoring adjuvant porous structures to optimize vaccine formulations for improved immunogenicity and efficacy. Overall, our research demonstrates the utility of small-angle scattering as a powerful technique for in-situ characterization of aluminum adjuvant aggregates, offering valuable insights into their porous structure and interactions with antigens. This underscores the importance of porous structure as a critical property influencing vaccine efficacy and highlights avenues for the design of novel adjuvants.