Immobilization of aluminum hydroxide particles on quartz crystal

Subscriber access provided by READING UNIV

Article

Immobilization of aluminum hydroxide particles on quartz crystal microbalance sensors to elucidate antigen-adjuvant interaction mechanisms in vaccines Jean-François Art, Aurélien vander Straeten, and Christine C. Dupont-Gillain Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b03747 • Publication Date (Web): 09 Dec 2017 Downloaded from http://pubs.acs.org on December 11, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Immobilization of aluminum hydroxide particles on quartz crystal microbalance sensors to elucidate antigen-adjuvant interaction mechanisms in vaccines Jean-François Art, Aurélien vander Straeten, Christine C. Dupont-Gillain Université catholique de Louvain, Institute of Condensed Matter and Nanosciences, Bio- and Soft Matter division, Place Louis Pasteur 1/L4.01.10, 1348, Louvain-la-Neuve, Belgium. ABSTRACT: Aluminum hydroxide (AH) salts are the most widely used adjuvants in vaccine formulation. They trigger immunogenicity from antigenic subunits that would otherwise suffer from a lack of efficiency. Previous studies focusing on antigen-AH interaction mechanisms, performed with model proteins, suggested that electrostatic interactions and phosphate-hydroxyl ligand exchanges drive protein adsorption on AH. We however recently evidenced that NaCl, used in vaccine formulation, provokes AH particle aggregation. This must be taken into account to interpret data related to protein adsorption on AH. Here, we report on the successful development and use of a stable AH-coated surface to explore the mechanisms of protein adsorption by means of ultrasensitive surface analysis tools. Bovine serum albumin (BSA) adsorption was studied at different pH and ionic strength (I) using quartz crystal microbalance. The results show that protein adsorption on AH adjuvant cannot be explained solely by electrostatic interactions and ligand exchanges. Hence, a higher adsorption was observed at pH 3 compared to pH 7, although AH and BSA respectively undergo repulsive and attractive electrostatic interactions at these pH values. Almost no effect of I on adsorption was moreover noted at pH 7. These new developments and observations not only suggest that other mechanisms govern protein adsorption on AH, but also offer a new platform for the study of antigen adsorption in the context of vaccine formulation. Immobilizing particles on QCM sensors also enriches the range of applications for which QCM can be exploited, especially in colloid science.

To prevent infectious diseases, vaccines are mostly prepared from antigenic subunits of pathogens, such as proteins or polysaccharides, or from pathogen toxoids, e.g. the tetanus or diphtheria toxoids.1 These compounds may suffer from a lack of immunogenicity 2,3, and adjuvants must be added to trigger sufficient immune response and memory effect.4,5 Many types of vaccine adjuvants have been studied, from bacterial extracts to emulsions and specialty polymers.2,3,6 The more widespread adjuvants currently used in human vaccines are aluminumbased suspensions. Among them, aluminum hydroxide (AH) adjuvants have been extensively studied. Commercial AH adjuvants are produced in a very controlled way 7 to obtain a poorly crystalline AlOOH structure, in which hydroxide groups (OH) are coordinated with aluminum atoms.8,9 The AH particles morphology and aggregation state was studied by Shirodkar et al.9, showing a dynamic system with elemental particles aggregating and disaggregating continuously in suspension, and forming loose aggregates (1-10 µm) considered as the functioning units of adjuvants, i.e. the effective structure for antigen adsorption and immune system stimulation.10 The specific surface area developed by AH adjuvants was found to be around 500 m²/g 11 and their point of zero charge (PZC) is 11.4.12 Stability studies of AH adjuvants showed that more crystalline fractions were developed in AH particles over time or after autoclave treatment. Higher crystallinity led to a lower AH adjuvant specific surface area, leading to a decreased antigen adsorption capacity.13,14 We recently showed that adjusting AH suspension ionic strength (I) with NaCl triggers the aggregation of AH particles. Indeed, adding 150 mM of NaCl to

commercial AH suspensions promoted particles aggregation, thereby decreasing specific surface area.15 In order to produce effective vaccines, the industry has focused on the adsorption of antigenic subunits on adjuvants, as it was shown that many antigenic subunits lead to a higher immune response when adsorbed on aluminum-based adjuvants compared to soluble antigens. The World Health Organization for example recommends for tetanus and diphtheria toxoids that 80% of antigen is adsorbed on adjuvant particles.16 Adsorption mechanisms of model proteins on aluminum-based adjuvants were studied by establishing protein adsorption isotherms on adjuvant particles in suspension. Two major mechanisms were proposed. Electrostatic interactions were thought to be responsible for bovine serum albumin (BSA) and lysozyme adsorption on AH and aluminum phosphate adjuvants at physiological pH (7.4) respectively.8,17 This mechanism was highlighted by a decrease of the adsorptive capacity observed while rising I with NaCl. These authors however did not take into account the effect of NaCl on the adjuvant particles themselves to interpret their data. The second principal driving force for antigen adsorption on adjuvant particles is believed to be a ligand exchange occurring between antigen phosphate groups and adjuvant hydroxyl groups.18,19 It was moreover shown that adsorption by ligand exchange can overcome repulsive electrostatic forces, and lead to less antigen desorption in vitro.19,20 With some antigens however, adsorption on the adjuvant is reported not to be necessary to trigger immunogenicity.21-24 It was also shown that the immune response could even decrease in case of too strong antigen adsorption.25-27 These works highlight the fact

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

that understanding antigen adsorption mechanisms is critical to control and increase the efficacy of vaccines. Other studies addressed the adsorption behavior of antigens used in vaccine formulation in order to develop commercial vaccines, without focusing on adsorption mechanisms.26,28-30 These adsorption studies suffer from several drawbacks. First, adsorption is performed in an adjuvant suspension with a complex composition, and the exact role of each component in the adsorption process is difficult to identify. Secondly, the adsorbed amount is determined indirectly by measuring the remaining protein amount in the supernatant after centrifugation. Furthermore, the effect of salt, acid or base addition on adjuvant structure is not taken into account in the adsorption studies or proposed mechanisms. We however showed that NaCl provokes the aggregation of AH particles 15, and the latter is as well dependent on pH. Using advanced characterization methods, for which the effect of NaCl, pH or excipients is controlled, and which allow the amount of adsorbed antigen to be determined directly, should lead to a better knowledge of antigen adsorption mechanisms on aluminum-based adjuvants. Controlling the aggregation and disaggregation of AH particles during the adsorption process would also help to obtain more representative adsorption results. In this context, quartz crystal microbalance (QCM) emerged over the past few years as a powerful tool to directly study adsorption from bulk solutions onto solid surfaces, without requiring labelled molecules.31 QCM has already been used for a wide range of applications such as monitoring layer-by-layer constructions with polymers and proteins 32, characterizing and understanding the behavior of stimuli-responsive surfaces 33,34, and measuring protein/particle adsorption from a solution onto a surface-modified QCM sensor.35,36 The chemical functionalization of the QCM sensor surface, and the deposition of polymer coatings or nanoparticles were also monitored by QCM to create antibacterial surfaces.37,38 The study of protein adsorption or even of cell adhesion on tailored interfaces was also probed by QCM measurements.39,40 QCM was as well transformed in a gas chromatograph for the quantification of controlled substances at high temperature, by modifying sensor surfaces with layers of PDMS or carbon nanotubes.41 Owing to its high sensitivity, QCM was also used to characterize the interactions between lipodisks grafted on QCM sensors and peptides in solution, with an accuracy similar to the one obtained with fluorescent labelling.42 Beside all these applications, researchers have been interested in the past few years in the comprehension of the effects of proteins, viruses or nanoparticles adsorption or adhesion on the observed frequency and dissipation shifts in QCM.31 Johannsmann et al. investigated the effect of the distribution of adsorbed globular proteins on dissipation.43 They concluded that the recorded dissipation is due to the geometry of the attachment rather than to the internal properties of the attached particles. Grunewald et al. aimed at determining the coverage of monodisperse functionalized silica nanoparticles on a gold surface.44 The use of scanning electron microscopy images was needed to determine the coverage and to understand the obtained QCM results. The weak binding of these rigid particles on the sensor was responsible for the high dissipation recorded by QCM, and the low surface coverage was attributed to repulsion between particles. The design and use of particle-covered QCM sensors is very challenging. Particle deposition needs to be controlled, and layer properties must be understood before using these modified sensors for adsorption studies.

Page 2 of 10

This work first aims at developing stable and reproducible quartz crystal microbalance sensor surfaces coated with immobilized AH particles. Once immobilized, AH particles will be less sensitive to changes in the surrounding conditions (especially to aggregation in presence of NaCl). AFM and QCM, which were not used before in the field of vaccine formulation, will be used to characterize immobilized aluminum hydroxide adjuvant particles. The second aim of the present work is to further use QCM to directly monitor the adsorption behavior of BSA, taken as model antigen, on AH particles in different adsorption conditions. The development of AH adjuvant-covered QCM sensors is expected to bring significant advances for the in situ and in real time study of antigen adsorption. This will open the possibility to directly determine antigen-AH adsorption mechanisms, which is a key for a more efficient vaccine formulation process.

EXPERIMENTAL SECTION Two commercial aluminum hydroxide adjuvants were used. ALHYDROGEL “85” 2% - Ph. Eur. (AH1) was kindly provided by BRENNTAG (Brenntag Biosector, Denmark), and REHYDRAGEL LV (AH2) was purchased from General Chemical (NJ, USA). Both adjuvants consist in a suspension of particles in water. Bovine serum albumin (BSA) was selected as model protein antigen (≥96%; Sigma Aldrich, St. Louis, MO, USA) and was diluted at 200 µg/ml in ultrapure water whose pH and I were adjusted. I was adjusted with NaCl (Merck, Darmstadt, Germany) and pH was adjusted with HCl or NaOH and measured with pH-indicator strips (pH 1-14 and pH 2-9, Merck, Darmstadt, Germany). Immobilization of AH particles on QCM sensors. AH-based surfaces were prepared by coating quartz crystal microbalance sensors (4.95 MHz AT-cut gold-coated quartz sensors, QSense, Stockholm, Sweden) with adjuvant particles from suspension. QCM sensors were washed with a 2:1 (v:v) mixture of H2SO4 95% and H2O2 30% w/w (VWR International, Leuven, Belgium), rinsed with ultrapure water, and exposed during 15 min to UV/O3 (UVO cleaner, Jelight Company, USA) before coating. Gold surfaces were coated by spin-coating of 160 µl of pH-adjusted AH suspension (pH 4 for AH1 suspension and pH 6 for AH2 suspension) with a WS-400B6NPP/Lite spin-coater (Laurell 8 technologies corporation, North Wales, USA) during 30 s at 10,000 rpm. Surface chemical composition determined by X-ray photoelectron spectroscopy (XPS). Atomic surface composition of AH-coated sensors was determined using a SSX 100/206 photoelectron spectrometer from Surface Science Instruments (USA), as described previously.15 Morphology revealed by atomic force microscopy (AFM). Topographic images of AH-coated sensors were recorded using a Nanoscope V multimode microscope (Digital Instruments, Santa Barbara, USA), as described previously.15 AH coating and BSA adsorption assessed by quartz crystal microbalance (QCM). QCM measurements were performed on a Q-Sense E4 device (Q-Sense Technologies, Stockholm, Sweden). Solutions were degassed under vacuum before experiments. The flow was set at 50 µl/min and the temperature was fixed at 25°C. The resonance frequency of cleaned unmodified quartz sensors, as well as of sensors after coating


Page 3 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


with AH particles were measured under ultrapure water. Every AH-covered sensor was submitted two times to successively (i) NaCl 750 mM solution for 15-20 min, and (ii) ultrapure water for 15-20 min to obtain stable AH-modified surfaces (see details in the “Results and discussion” section). Sensors rinsed according to this procedure were characterized by XPS and AFM. The resonance frequency of NaCl-rinsed AHcoated sensors was also recorded under ultrapure water and subtracted to the one obtained for unmodified sensors, to obtain a frequency shift (∆fAH) that gives an estimation of the amount of immobilized AH particles. BSA adsorption experiments were then conducted as schematically described in Figure S-1 (Supporting Information), at pH 3, 5 and 7 under I of ~0 mM (ultrapure water), 60, 250 and 750 mM in NaCl. These I values were chosen according to previous adsorption studies from the literature.8,17 The “Q-stomize” software, developed in the laboratory by K. Mc Evoy 45, enabled to accurately extract ∆f and dissipation shifts (∆D) over selected time periods, with respect to the baseline recorded in water. The average of the normalized 3rd, 5th and 7th overtones was used to compute BSA adsorption frequency shift (∆fBSA-AH in Fig. S-1). BSA adsorption results were expressed under the form of ∆f values, without conversion into mass, and were normalized by the frequency shift representing the immobilized AH amount (∆fAH). This enables the comparison of results from different experiments, i.e. obtained on different sensors. An isoresponse map of ∆fBSA-AH/∆fAH ratios presented as a function of pH and the square root of I was finally drawn with the JMP software (JMP 12.0, SAS Institute, NC, USA). This contour plot was generated by a triangulation method and linear interpolation of experimental data. BSA adsorption assessed on AH in suspension. BSA adsorption isotherms on AH adjuvant particles in suspension were determined experimentally at pH 3, 5 and 7 and for final I of ~0, 60, 250 or 750 mM. Therefore, the residual protein concentration was measured in the supernatant of the adjuvant suspension after adsorption and subtracted from the initial concentration. Adsorption experiments in suspension were only conducted on AH1 adjuvant because of the complete gelation of AH2 adjuvant suspension while adding NaCl.15 Suspensions (2 ml) were prepared by mixing pH-adjusted AH suspension containing 3.4 mg aluminum, which corresponds to the maximum authorized aluminum amount in a vaccine 17, pH-adjusted NaCl to set I and pH-adjusted BSA solution as a final step. Incubation was achieved at 200 rpm on an orbital shaker (Infors HT, Bottmingen, Switzerland) for 1 h. Samples were centrifuged at 136×g (4000 rpm on Heraeus Multifuge XIR, Thermo Scientific, Osterode, Germany) and the supernatant was collected, then diluted 0, 2 and 5 times in glass tubes. Protein concentration was measured using the standard protocol of the BCA protein assay procedure (Pierce BCA Protein Assay, Thermo Fisher Scientific Inc., Rockford, IL, USA). A calibration curve was included in each experiment. BSA adsorptive capacities were extracted from adsorption isotherms in every condition of pH and I (Figure S-2 in Supporting Information). An adsorptive capacity isoresponse map presented as a function of pH and √I was finally drawn with the JMP software (JMP 12.0, SAS Institute, NC, USA).

RESULTS AND DISCUSSION Preparation and characterization of AH-coated QCM sensors. In order to study BSA adsorption on AH-coated

QCM sensors, the deposited adjuvant layer must be thin and homogeneous to keep a sensitive sensor, and to avoid BSA adsorption on uncovered gold areas of the sensor respectively. Furthermore, it is important to check that AH particles keep the same morphology and chemistry as in suspension. NaCl treatment on immobilized AH adjuvant particles. QCM sensors spin-coated with AH particles were mounted in the QCM device. The AH-covered sensors suffered from signal drifts and frequency shifts upon changing medium composition during QCM experiments (Figure S-3 in Supporting Information). To obtain a better signal stability, before any protein adsorption experiments, AH-modified QCM sensors were submitted to NaCl treatment as described in experimental section. Figure 1 A presents a typical result of this NaCl treatment applied to a AH1 spin-coated layer. The first rinsing step with NaCl decreases the resonance frequency. This decrease was also observed while treating naked gold sensors with NaCl solutions (data not shown) and is attributed to a bulk effect due to the presence of ions in the rinsing solution. The rinsing step with water provokes a large increase of ∆f, attributed to the partial release of the adjuvant layer. The second treatment with NaCl and water does not modify the deposited AH1 layer further.

Figure 1. NaCl 750 mM treatment applied to AH-spin-coated layers and monitored by QCM. A: AH1, B: AH2. Treatment consists into two cycles of successive rinsing steps with NaCl 750 mM and water. The 3rd, 5th and 7th overtones are presented.

Figure 1 B presents the same NaCl treatment applied to a AH2 spin-coated layer. The first rinsing step with NaCl seems to alter the deposited AH2 layer as the frequency shift increases. The rinsing step with water shows a strong discrepancy between the different overtones, together with an increase of the dissipation factor (data not shown). This observation suggests that the mechanical properties of the deposited layer are changing to a more viscoelastic behavior. This increase in dissipation observed during the NaCl treatment of AH2 immobilized particles could be linked to the weakening of the bonds between AH particles and the gold surface, as explained by Johannsmann.43 After the second treatment with NaCl and



water, the AH2 layer has the same normalized frequency shift value and same overtone discrepancy than after the first NaCl rinsing step. Coverage and morphology of AH particles immobilized on QCM sensors. Figure 2 shows atomic force microscopy images of samples prepared by spin-coating of AH particles on QCM sensors before (Figure 2 - A, B) and after NaCl 750 mM treatment (Figure 2 - C, D). The comparison with a naked sensor surface (Figure S-4 in Supporting Information) shows that surfaces spin-coated with AH suspensions are almost completely covered by adjuvant particles. Before NaCl 750 mM treatment, surfaces show different morphologies depending on the aluminum hydroxide adjuvant (AH1 vs AH2). AH1 surface is covered with homogeneously spread micrometer scale clusters of particles, with a height of ~100 nm. These clusters are surrounded by smaller round-shaped or elongated particles considered as the elemental units of AH adjuvants.10 These particles (Fig. 2 A inserts) keep, after spin-coating, the morphology and size already described in the literature.9,15 AH2 surface is mostly covered by adjuvant particles showing the expected shape and size for the elemental units of this adjuvant (Fig. 2 B - inserts).9,15 In some places, adjuvant particle aggregates of micrometer scale in height and length are observable as well. These differences in AH layer morphology explain the different behavior of AH1 and AH2 upon NaCl treatment.

Page 4 of 10

After NaCl treatment in the QCM flow cells, both surfaces still present a very high coverage with AH particles (Fig. 2 C, D). Moreover, AH1 and AH2 surfaces now present similar morphologies.The AH1 particle clusters were rinsed or degraded by exposure to NaCl 750 mM. The positive frequency shift observed after the first NaCl rinsing step (Fig. 1 A) is consistent with this loss of AH mass. AH1 particle layers are now smooth and homogeneous. AH2 micrometer scale aggregates were also rinsed off or degraded by exposure to NaCl. Only a small positive frequency shift was however observed by QCM (Fig. 1 B). It seems that there is almost no mass loss during NaCl treatment. The overtone discrepancy observed in Fig. 1 B could be explained by a loose adsorption of AH particles coming from the degraded aggregates on the goldimmobilized AH particles layer. The weaker bonds between AH2 particles could be responsible for the dissipation increase and the overtone discrepancy, as proposed by Johannsmann et al.43 and Grunewald et al.44 This explanation is consistent with the overall higher layer roughness observed on Fig. 2 D, compared to Fig. 2 C. Furthermore, in both cases, the morphology of AH elemental units, i.e. the smaller observed particles, is not altered, and particle size is not decreased by the saline treatment (Fig. 2 - C, D inserts). NaCl was previously shown to promote AH particles aggregation in AH suspension.15 Here, NaCl removes AH aggregates deposited from suspension onto gold surfaces.

Figure 2. AH1-coated (A, C) and AH2-coated (B, D) QCM sensors imaged by AFM. Images were recorded before (A, B) or after (C, D) NaCl 750 mM treatment. Inserts: images recorded at higher magnification on the same sample.


Page 5 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Two different mechanisms are thus triggered by NaCl. In AH suspension (pH around 5-6), AH particles are positively charged (PZC around 11.4) and are repelling each other. NaCl addition to the suspension screens AH positive charges, reduces repulsion and makes aggregation possible. Moreover, at acidic pH, ligand exchange between protonated AH particles and Cl- anions can occur, leading to a less positive particle electrical charge, which further promotes aggregation.46 For surface-immobilized AH adjuvant, ligand-directed labilization may be invoked to explain the observed behavior. Adsorbed ligands, such as chloride ions, might enter the coordination sphere of metallic atoms and weaken bond strength. This is especially the case for aluminum-oxygen bonds that are weak and easily undergo ligand-directed labilization.47 Al-O-Al structures are then evolving into Al-OH + Cl-Al, which indeed promotes the dissolution and desorption of AH particle aggregates.47 Al–O bonds partial rupture may explain the disappearance of clusters and aggregates, which are rinsed with the liquid flow in the QCM cells. Osmotic effect may also explain the removal of aggregates from the surface. Water is trapped in the spin-coated adjuvant particles layer. Treating this layer with a saline solution provokes the expulsion of trapped water, thereby expelling AH aggregates. Aggregates removal from AH-deposited layers is probably due to a combination of ligand-directed labilization and osmotic effect. Finally, the obtained AH layers highly cover the gold surface, which will reduce unwanted protein-gold interaction events during QCM experiments, and AH particles keep the same size and shape as in adjuvant suspension. Surface chemistry of AH particles immobilized on QCM sensors. XPS analysis of AH powders obtained by freezedrying of adjuvant suspensions were performed in a previous work.15 When treating XPS data, the oxygen peak was divided into oxygen linked to aluminum (O–Al, from the adjuvant) and oxygen from environmental organic contamination (O– C).15 Figure 3 compares the aluminum to O-Al ratio obtained for powders and for AH-spin-coated QCM sensors before and after treatment with NaCl. All samples present an Al/O-Al ratio between 0.33 and 0.38 as expected for aluminum hydroxide adjuvants.8,15 Indeed, AH adjuvants were characterized and described as a mixed structure composed of AlOOH crystallites (Al/O-Al = 0.5) and amorphous Al(OH)3 areas (Al/O-Al = 0.33). The Al/O-Al ratio is similar on the different samples. It means that the AH adjuvant chemistry (atomic composition and structure formula) is not modified upon spin-coating on QCM sensors and after NaCl 750 mM treatment. This result confirms that AH layers, used for adsorption experiments in QCM, keep the same surface chemistry and formula than AH particles in adjuvant suspension. AH-coated surface elemental composition obtained by XPS, before and after NaCl 750 mM treatment, is presented in Table S-1 (Supporting Information). Aluminum and oxygen linked to aluminum (O-Al) are the major detected constituents, confirming the modification of gold sensors by AH particles. Sensors coverage with AH particles is almost complete, as the signal of gold is low for both surfaces (2.4 % for AH1, 0.2 % for AH2). The effect of NaCl treatment on the morphology of AH1-coated surfaces, i.e. degradation of AH1 particle clusters (Fig. 2), is also evidenced by XPS, through the observed slight increase of gold fraction, concurrently with the decrease of Al and O-Al fractions (Table S-1).

Figure 3. Aluminum to oxygen ratio obtained by XPS analysis of AH1 and AH2 adjuvants in powders or spin-coated on QCM sensors. From left to right, Al/O-Al ratio for powders, Al/O-Al ratio for AH spin-coated sensors and Al/O-Al ratio for spin-coated sensors treated by two rinsing steps with 750 mM NaCl. Error bars are standard deviation (n≥3).

Almost no change was observed in the elemental composition of NaCl-treated AH2-coated sensors, which are still considered as fully covered after treatment with NaCl 750 mM (0.3 % of gold). This is consistent with a re-adsorption of AH2 particles coming from aggregate degradation upon exposure to NaCl. Note that there is no surface contamination with Na or Cl after NaCl treatment. Reproducibility of AH coating procedure and evaluation of the AH spin-coated layer. The determination of the frequency shift attributed to AH adjuvant particles is shown in Figure S-5 (Supporting Information). Figure S-6 (Supporting Information) presents the frequency shifts ∆fAH, obtained for more than seventy AH1 (Figure S-6 A) and AH2 (Figure S-6 B) spin-coated QCM sensors. The first two columns of Table 1 present the average and standard deviation obtained for ∆fAH1 and ∆fAH2. Before NaCl treatment, the deposited amount is of the same order of magnitude, and the dispersion of data is comparable for both AH. The NaCl treatment of AH1-coated sensors strongly decreases the deposited amount and the deviation between the different sensors. On AH2-covered sensors, NaCl treatment does neither significantly alter the deposited amount nor the deviation. Table 1. Average and standard deviation of ∆fAH, estimated mass and thickness of AH adjuvant layers, before and after treatment with NaCl. Average of ∆fAH [Hz]

Standard deviation of ∆fAH [Hz]

∆m obtained by Sauerbrey31 [ng cm-2]

Estimated thickness [nm]

AH1 Before NaCl

703

118

12,441

51

AH1 After NaCl

261

30

4,618

19

AH2 Before NaCl

1,173

188

20,757

86

AH2 After NaCl

1,008

202

17,836

74



To quantify the amount of AH adjuvant particles immobilized on QCM sensors, the behavior of AH layers under water must be evaluated. Figure 4 represents the ratio between the dissipation shift ∆D and the frequency shift ∆f obtained under water for the 3rd, 5th and 7th overtones, on uncoated, AH-coated and NaCl-treated QCM sensors (AH1 in Figure 4 A, AH2 in Figure 4 B). Ratios calculated for the uncoated QCM sensors are inconstant, because dissipation and frequency shifts are both very low and close to zero. Ratios obtained for AHcoated surfaces and NaCl 750 mM-treated surfaces are below the value of |∆D/∆fAH | = 4x10-7 Hz-1, proposed by Reviakine et al. as the threshold between rigid and soft films.31 Asdeposited and NaCl-rinsed AH layers can thus be considered as rigid. A second distinction is made between laterally homogeneous films or films made of discrete nanosized objects, because the evaluation of immobilized amount is then different.31,44 Considering images obtained by AFM (Fig. 2), AH adjuvant layers probably constitute an intermediate state between these film types. Since the dissipation shifts are low (not shown), and the layer can be considered as rigid, the frequency shifts ∆fAH are directly representative of the areal mass density (mass per unit of surface) when the overtone dispersion is low 31 (which is especially the case for AH1 adjuvant layers). In this case, the Sauerbrey relation can be used to obtain a deposited mass.

Figure 4. Typical ratios between ∆D and ∆f obtained under water for uncoated QCM sensors, AH-coated sensors and NaCl 750 mM-treated AH-coated sensors (A: AH1, B: AH2). For each sensor treatment, frequency and dissipation were recorded during 4 min under a flow of ultrapure water.

The third column in Table 1 presents the areal mass density of immobilized AH particles obtained by application of the Sauerbrey relation on the average of ∆fAH for both adjuvants. The obtained deposited mass is then used to roughly estimate the thickness of the adjuvant layer (fourth column of Table 1) based on AH density. AH adjuvant being a mixed structure composed of amorphous Al(OH)3 and poorly crystalline AlOOH 8, the density of aluminum hydroxide was used (2.42

Page 6 of 10

g/cm³, The National Institute for Occupational Safety and Health – NIOSH, Atlanta, USA). Considering AH particles observed in inserts of Fig. 2, the estimated thickness of NaCl-rinsed AH1 layer (~19 nm) is very close to the size of elemental particles (~20 nm). AH1 layer is thus probably made of a homogeneous monolayer of AH particles. For NaCl-rinsed AH2 adjuvant layers, the estimated thickness (~74 nm) quite well corresponds to the hypothesis of a first layer of adjuvant particles topped with several layers of loosely adsorbed particles. However, the overtone discrepancy observed in Fig. 1 B for AH2 layers treated with NaCl theoretically prevents an exact determination of AH2 immobilized mass through Sauerbrey modeling. The Voigt-Voinova model 48, which takes the viscoelastic behavior of films into account, was also applied to our results (data not shown). Obtained mass and thickness values with this model are ~1.5 times higher for AH1 and ~1.7 times higher for AH2 adjuvant compared to Sauerbrey model. The determination of the deposited AH layer thickness was needed to check that the QCM sensors keep their sensitivity after being coated. The sensitivity of the QCM sensor is determined by the penetration depth of the shear wave in the liquid over the sensor. With 4.95 MHz QCM sensors in water, this depth is about 250 nm.31 Whatever the model used to estimate the AH layer thickness on QCM sensors, it is reasonable to assume that NaCl-rinsed AH-coated sensors will be sensitive to the further adsorption of BSA. Considering the different sources of error in modeling deposited mass and layer thickness (choice of model, film roughness 45 , etc.), the use of the frequency shift ∆fAH to represent immobilized AH amount was preferred to masses extracted from modeling of QCM raw data. In conclusion, QCM sensors showing a very high surface coverage with a thin and homogeneous layer of AH particles were obtained without any detectable change in adjuvant morphology nor chemistry compared to the state in suspension. These AH particle layers are moreover insensitive to further I change, as well as to pH changes in the range of 3 to 7 (see Table S-2 and Figure S-7 in Supporting Information). BSA adsorption on AH adjuvant can then be studied by QCM measurements on AH-modified sensors. BSA adsorption monitored by quartz crystal microbalance. NaCl-treated AH-coated QCM sensors were used to monitor BSA adsorption at pH 3, 5 and 7, and under I of ~0, 60, 250, 750 mM. In the studied pH range, AH adjuvant particles are positively charged (PZC around 11.4).12 BSA is itself positively charged at pH 3, close to neutral at pH 5 and is negatively charged at pH 7 (IEP around 4.7 – 4.9 49). ∆fBSA-AH/∆fAH ratios obtained by QCM in each condition of pH and I are presented in Figure 5 (A for AH1, B for AH2) while values with associated standard deviation are reported in Table S-3 and S-4 respectively for AH1 and AH2 adjuvants. The Debye length (κ-1), which indicates the distance over which electrostatic interactions are at play, is inversely proportional to √I. Isoresponse maps of Fig. 5 are presented as a function of pH and of √I, to directly interpret results in terms of Debye length and electrostatic interactions. BSA adsorption on AH1 is the highest in ultrapure water (~0 mM) around pH 5 (Fig. 5 A and Table S-3). Being polyampholytes, proteins are often presenting an adsorption maximum near their IEP, because of the decrease of protein-protein repulsion.50


Page 7 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. BSA adsorption frequency shifts normalized by AH1 adjuvant frequency shift (A) or AH2 adjuvant frequency shift (B) measured on NaCl-treated AH-coated QCM sensors and presented as a function of pH and square root of ionic strength. Black dots are means of experimental data (n=5).

At pH 3, when BSA and AH1 particles are both positively charged, BSA adsorption increases with increasing I. At high I, the positive charges of protein and adjuvant are screened, the Debye length is reduced and adsorption can occur by a mechanism which is not related to electrostatic interactions (e.g. through ligand exchange or van der Waals interactions). Moreover, patches of negatively charged residues may also contribute to the adsorption mechanism, and their action will only become efficient at high ionic strength, when the repulsion operated by the predominant positive residues is strongly decreased. BSA adsorption at pH 3 and 750 mM in NaCl actually reaches almost the same level as adsorption at pH 5 in ultrapure water, i.e. the highest amount of adsorbed BSA. The most interesting observation on AH1 particles is made at pH 7, with BSA and AH1 respectively negatively and positively charged. At this pH, in ultrapure water, BSA poorly adsorbs on AH1 adjuvant. Furthermore, when a low salt concentration is added, BSA adsorption is higher, and does not vary further with I (considering standard deviations presented in Table S-3), excluding a major contribution of coulombic interactions in adsorption. On AH2-coated QCM sensors, BSA adsorption is globally lower than with AH1-coated sensors (Fig. 5 B and Table S-4). The highest amount of adsorbed BSA is measured at pH 5. The effect of rising I at pH 5 seems greater than for AH1, but considering standard deviations in Table S-4, it can be concluded that this effect is quite limited. At pH 3, BSA adsorption is low in ultrapure water, and also reaches a higher level at 750 mM in NaCl, when electrostatic repulsion is better screened. At pH 7, the conclusion is the same as for AH1, i.e. BSA adsorption is the lowest in ultrapure water, while in other saline conditions, the adsorption is slightly higher and independent of I. In conditions of pH at which BSA and AH adjuvants are oppositely charged, adsorption mechanism seems thus not related to electrostatic interactions. Adsorption in spite of electrostatic repulsion is furthermore recorded at pH 3. This is in contradiction with previously published works of Hem and coworkers related to the adsorption of BSA on AH adjuvants at physiological pH (7.4), in which electrostatic interactions were identified as the main mechanism of adsorption.8,17

It should be noted that the BSA adsorption results obtained here by QCM confirm that the NaCl-rinsed AH-coated sensors we developed are sensitive enough, i.e. AH deposited layer is sufficiently thin, to monitor the adsorption of proteins on AH adjuvant particles by QCM. The morphology of the AHcoated surfaces was moreover stable upon BSA adsorption (see Figure S-8 in Supporting Information). BSA adsorption on AH particles in suspension. For the sake of comparison, Figure 6 presents the adsorptive capacity isoresponse map obtained for BSA adsorption on AH1 particles in suspension (this time not on AH particles immobilized on QCM sensors), as a function of pH and √I.

Figure 6. Adsorption capacity isoresponse map obtained from BSA adsorption isotherms on AH1 particles in suspension. BSA adsorbed amount was determined by difference and is expressed in mg BSA/mg Al. Black dots are means of experimental data (n≥2).

At pH 3, BSA adsorption behavior is the same as the one obtained by QCM on AH immobilized particles, i.e. adsorption capacity increases with I. BSA adsorbed amount is the highest at pH 5 in ultrapure water (~0 mM), i.e. near the IEP of the protein as expected from the literature.50 At this pH, BSA adsorbed amount decreases with increasing I, while it was constant at the same pH in QCM experiments performed with immobilized adjuvant particles. Results obtained at pH 7 are consistent with those presented in the literature at pH 7.4



Page 8 of 10

and tentatively explained by electrostatic interactions.17 However, these results are the opposite of those obtained by QCM on immobilized AH particle layers, which showed an increase in BSA adsorption from experiments in ultrapure water and at 60 mM in NaCl and then, the same adsorbed amount whatever the I. For the sake of comparison, ∆fBSA-AH1/∆fAH1 values obtained by QCM were converted into mg BSA/mg Al, based on the Sauerbrey model. Adsorptive capacities obtained by QCM were two to five times lower than the ones obtained in suspension. A lower adsorptive capacity was expected in QCM, considering the immobilization of AH particles, and thus a decrease of accessible particle surface for BSA. The adsorptive capacities are however of the same order of magnitude in QCM and after measurements on particles in suspension, further supporting the fact that the AH particle layers enable a quantitative detection of adsorbed proteins by QCM. The values of I chosen in the present work, in line with previous reports from the literature will now be further discussed. The Debye lengths (κ-1), calculated 51 at pH 7 for each I (Table S-5 in Supporting Information), are really short and close to each other, in the range of 0.4 to 1.3 nm, between 60 and 750 mM in NaCl. It means that differences in the surface charge screening effect due to I adjustment are very low, whatever the NaCl concentration. Comparatively, κ-1 in ultrapure water is really high (~800 nm). If BSA adsorption was only driven by electrostatic interactions, its adsorption capacity in suspension should be drastically decreased between experiments at 0 and 60 mM in NaCl, while it should be almost constant between 250 and 750 mM in NaCl. However, this trend is not observed experimentally in suspension. This means that the conclusions drawn by Al-Shakhshir et al.17, based on BSA adsorption isotherms constructed at these same I values, need to be nuanced. The BSA adsorption capacity decrease, observed in suspension at pH 7 in previous works 17, is more correctly explained by the aggregation effect of NaCl and the concomitant specific surface area decrease 15 than by the screening of electrostatic interactions. By immobilizing AH adjuvant particles on a surface, and treating the adjuvant layer with two rinsing steps of NaCl 750 mM and water, no aggregation and no differences in developed specific surface area are observed whatever the NaCl concentration used for subsequent adsorption. BSA adsorption at pH 7 on AH-coated QCM sensors has shown to be the lowest in ultrapure water and higher and almost constant at higher I. Moreover, in AH1 suspension, BSA adsorption at pH 5 decreases with rising I while it is constant in QCM experiments. When the adjuvant is stable upon ionic strength changes, the effect of I on BSA adsorption vanishes. This is a proof that the effect of I observed in suspension is related to AH adjuvant aggregation state alteration more than to the screening of electrostatic interactions, and that BSA adsorption can obviously not be considered as only driven by electrostatic interactions.

thus developed through the use of quartz crystal microbalance. More generally, this opens perspectives for the analytical use of QCM in many applications in colloid science. The other advantage of this QCM-based method for the study of antigen adsorption on adjuvants is that adsorbed amounts can be determined directly and in real time during adsorption experiments. The elaborated layers will further allow other advanced surface analysis tools to be applied, such as AFM. The newly developed platform was used to monitor BSA adsorption on AH-based surfaces by QCM in a range of pH and I conditions. Results are not consistent with those obtained in suspension, and the electrostatic interaction-based mechanism suggested in the literature. Considering the used NaCl concentrations and their effect on charge screening, the decrease of BSA adsorption with rising I at physiological pH may be more convincingly explained by AH particle aggregation and the concomitant decrease of specific surface area due to NaCl addition 15. BSA adsorption on AH adjuvant is thus driven by another mechanism than electrostatic interactions. The fact that electrostatic interactions are weakly, or even not, involved in BSA adsorption on AH adjuvant revives the interest in antigen adsorption mechanisms on aluminum-based adjuvants and is of great importance in the vaccine formulation industry. Indeed, vaccine produced from new protein antigens are firstly formulated by trial and error experiments on the basis of electrostatic interactions. Protein adsorption on AH layer monitored with the help of the QCM technique could bring new information about antigen adsorption in order to

CONCLUSION

REFERENCES

Homogeneous layers of AH adjuvant particles were created without any change of surface chemical composition nor morphology compared to AH particles in suspension. These reproducible thin films present the advantage of being stable under ionic strength modification, while AH particles in suspension aggregate in contact with NaCl.15 A novel platform for studying antigen adsorption mechanisms on AH adjuvants was

(1) Perrie, Y.; Mohammed, A. R.; Kirby, D. J.; McNeil, S. E.; Bramwell, V. W. International Journal of Pharmaceutics 2008, 364, 272280. (2) Mbow, M. L.; De Gregorio, E.; Valiante, N. M.; Rappuoli, R. Current Opinion in Immunology 2010, 22, 411-416. (3) Eppstein, D. A.; Byars, N. E.; Allison, A. C. Advanced Drug Delivery Reviews 1989, 4, 233-253.

reduce time and costs of vaccine formulation.

ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge on the ACS Publications website.

AUTHOR INFORMATION Corresponding Author * E-mail: [email protected] Université catholique de Louvain, Institute of Condensed Matter and Nanosciences, Bio- and Soft Matter division, Place Louis Pasteur 1/L4.01.10, 1348, Louvain-la-Neuve, Belgium.

Author Contributions The manuscript was written through contributions of all authors.

ACKNOWLEDGMENT The authors thank the BRENNTAG company for kindly providing the ALHYDROGEL® ‘85’ material. The work was supported by the Belgian Science Policy through the Interuniversity Attraction Pole Program (P07/05) and by the Belgian National Foundation for Scientific Research (FNRS). Aurélien vander Straeten is a PhD researcher supported by a Research Foundation for Industry and Agriculture (FRIA) fellowship.


Page 9 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(4) O'Hagan, D. T.; MacKichan, M. L.; Singh, M. Biomolecular Engineering 2001, 18, 69-85. (5) Aguilar, J. C.; Rodríguez, E. G. Vaccine 2007, 25, 3752-3762. (6) Sivakumar, S. M.; Safhi, M. M.; Kannadasan, M.; Sukumaran, N. Saudi Pharmaceutical Journal 2011, 19, 197-206. (7) Lindblad, E. B. Vaccine 2004, 22, 3658-3668. (8) Hem, S. L.; White, J. L. In Vaccine Design : The Subunit and Adjuvant Approach, Powell, M. F.; Newman, M. J., Eds.; Plenum Press: New York, 1995, pp 249-276. (9) Shirodkar, S.; Hutchinson, R. L.; Perry, D. L.; White, J. L.; Hem, S. L. Pharmaceutical Research 1990, 7, 1282-1288. (10) Morefield, G. L.; HogenEsch, H.; Robinson, J. P.; Hem, S. L. Vaccine 2004, 22, 1973-1984. (11) Johnston, C. T.; Wang, S.-L.; Hem, S. L. Journal of Pharmaceutical Sciences 2002, 91, 1702-1706. (12) Rinella Jr, J. V.; White, J. L.; Hem, S. L. Journal of Colloid and Interface Science 1998, 205, 161-165. (13) Burrell, L. S.; Lindblad, E. B.; White, J. L.; Hem, S. L. Vaccine 1999, 17, 2599-2603. (14) Burrell, L. S.; White, J. L.; Hem, S. L. Vaccine 2000, 18, 21882192. (15) Art, J.-F.; vander Straeten, A.; Dupont-Gillain, C. C. International Journal of Pharmaceutics 2017, 517, 226-233. (16) Clausi, A.; Cummiskey, J.; Merkley, S.; Carpenter, J. F.; Braun, L. J.; Randolph, T. W. Journal of Pharmaceutical Sciences 2008, 97, 5252-5262. (17) Al-Shakhshir, R. H.; Regnier, F. E.; White, J. L.; Hem, S. L. Vaccine 1995, 13, 41-44. (18) Iyer, S.; HogenEsch, H.; Hem, S. L. Pharmaceutical Development and Technology 2003, 8, 81-86. (19) Morefield, G. L.; Jiang, D.; Romero-Mendez, I. Z.; Geahlen, R. L.; HogenEsch, H.; Hem, S. L. Vaccine 2005, 23, 1502-1506. (20) Jiang, D.; Morefield, G. L.; HogenEsch, H.; Hem, S. L. Vaccine 2006, 24, 1665-1669. (21) HogenEsch, H. Frontiers in Immunology 2013, 3. (22) Romero Méndez, I. Z.; Shi, Y.; HogenEsch, H.; Hem, S. L. Vaccine 2007, 25, 825-833. (23) Berthold, I.; Pombo, M.-L.; Wagner, L.; Arciniega, J. L. Vaccine 2005, 23, 1993-1999. (24) Noh, J. Y.; Song, J. Y.; Choi, W. S.; Lee, J.; Seo, Y. B.; Kwon, Y. J.; Ko, G. J.; Cha, D. R.; Kang, Y. S.; Lee, Y.-K.; Cheong, H. J.; Kim, W. J. Human Vaccines & Immunotherapeutics 2016, 12, 29022908. (25) Hansen, B.; Sokolovska, A.; HogenEsch, H.; Hem, S. L. Vaccine 2007, 25, 6618-6624. (26) Hansen, B.; Belfast, M.; Soung, G.; Song, L.; Egan, P. M.; Capen, R.; HogenEsch, H.; Mancinelli, R.; Hem, S. L. Vaccine 2009, 27, 888-892. (27) Hansen, B.; Malyala, P.; Singh, M.; Sun, Y.; Srivastava, I.; Hogenesch, H.; Hem, S. L. Journal of Pharmaceutical Sciences 2011, 100, 3245-3250. (28) DePaz, R. A.; Henderson, I.; Advant, S. J. Vaccine 2005, 23, 4029-4035. (29) Iyer, S.; Robinett, R. S. R.; HogenEsch, H.; Hem, S. L. Vaccine 2004, 22, 1475-1479. (30) Ljutic, B.; Ochs, M.; Messham, B.; Ming, M.; Dookie, A.; Harper, K.; Ausar, S. F. Vaccine 2012, 30, 2981-2988. (31) Reviakine, I.; Johannsmann, D.; Richter, R. P. Analytical Chemistry 2011, 83, 8838-8848. (32) Mauquoy, S.; Dupont-Gillain, C. Colloids and Surfaces B: Biointerfaces 2016, 147, 54-64. (33) Osypova, A.; Magnin, D.; Sibret, P.; Aqil, A.; Jerome, C.; Dupont-Gillain, C.; Pradier, C. M.; Demoustier-Champagne, S.; Landoulsi, J. Soft Matter 2015, 11, 8154-8164. (34) Delcroix, M. F.; Demoustier-Champagne, S.; Dupont-Gillain, C. C. Langmuir 2014, 30, 268-277. (35) Delcroix, M. F.; Laurent, S.; Huet, G. L.; Dupont-Gillain, C. C. Acta Biomaterialia 2015, 11, 68-79. (36) Raj, G.; Swalus, C.; Delcroix, M.; Devillers, M.; Dupont-Gillain, C.; Gaigneaux, E. M. Journal of Colloid and Interface Science 2015, 445, 24-30.

(37) Yang, Y.; Wikieł, A. J.; Dall’Agnol, L. T.; Eloy, P.; Genet, M. J.; Moura, J. J. G.; Sand, W.; Dupont-Gillain, C. C.; Rouxhet, P. G. Biofouling 2016, 32, 95-108. (38) Kruk, T.; Szczepanowicz, K.; Kręgiel, D.; Szyk-Warszyńska, L.; Warszyński, P. Colloids and Surfaces B: Biointerfaces 2016, 137, 158-166. (39) Dolatshahi-Pirouz, A.; Jensen, T. H. L.; Kolind, K.; Bünger, C.; Kassem, M.; Foss, M.; Besenbacher, F. Colloids and Surfaces B: Biointerfaces 2011, 84, 18-25. (40) Modin, C.; Stranne, A.-L.; Foss, M.; Duch, M.; Justesen, J.; Chevallier, J.; Andersen, L. K.; Hemmersam, A. G.; Pedersen, F. S.; Besenbacher, F. Biomaterials 2006, 27, 1346-1354. (41) Benz, M.; Benz, L.; Patel, S. V. Analytical Chemistry 2015, 87, 2779-2787. (42) Agmo Hernández, V.; Reijmar, K.; Edwards, K. Analytical Chemistry 2013, 85, 7377-7384. (43) Johannsmann, D.; Reviakine, I.; Richter, R. P. Analytical Chemistry 2009, 81, 8167-8176. (44) Grunewald, C.; Schmudde, M.; Noufele, C. N.; Graf, C.; Risse, T. Analytical Chemistry 2015, 87, 10642-10649. (45) Mc Evoy, K. Improving quantification of proteins at interfaces through a better and practical modeling of Quartz Crystal Microbalance response. Université catholique de Louvain2012. (46) Burrell, L. S.; Johnston, C. T.; Schulze, D.; Klein, J.; White, J. L.; Hem, S. L. Vaccine 2000, 19, 275-281. (47) Wingrave, J. A. Oxide Surfaces; Marcel Dekker, Inc., 2001; Vol. 103. (48) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Physica Scripta 1999, 59, 391. (49) Salis, A.; Boström, M.; Medda, L.; Cugia, F.; Barse, B.; Parsons, D. F.; Ninham, B. W.; Monduzzi, M. Langmuir 2011, 27, 1159711604. (50) Norde, W. Colloids and Surfaces B: Biointerfaces 2008, 61, 1-9. (51) Myers, D. In Surfaces, Interfaces, and Colloids; John Wiley & Sons, Inc., 2002, pp 79-96.




Page 10 of 10

10

Immobilization of aluminum hydroxide particles on quartz crystal

Recommend Documents