2.1. Maximum bHb Adsorption and Adsorption Kinetics
It is known that proteins contain a net positive charge below their isoelectric point (pI) and have an overall negative charge above it. The rate of protein adsorption is likely to be high when the adsorbent surface and protein bear opposite charges. This higher rate of adsorption in these conditions is primarily driven by the electrostatic attractions which accelerates the protein migration towards the oppositely charged surface [
25]. The adsorption studies presented in this study were performed in pH 6 phosphate buffer (0.133 M) due to the overall positive charge on bHb, considering the pI of bHb to be 7.1 [
26] and a negative charge on silica particles such that the ionic interactions could be encouraged. Another reason for the selection of this pH was the higher protein stability at pH 6 compared to pH 7 and 8 under stirring at room temperature [
27,
28]. The bHb adsorption on silica (
Figure 1) increased linearly with the increase in protein concentration, reaching a plateau at 12 mg/mL (225 mg/g) with only a slight increase in protein immobilisation at 15 mg/mL (236 mg/g).
The data for bHb adsorption on silica particles was fitted with both Langmuir and Freundlich models, resulting in R
2 values of 0.7915 and 0.9864, respectively. The Freundlich-type adsorption of bHb on S
FP indicated that it is a multi-layer process in which the amount of adsorbed solute per unit adsorbent mass increases gradually with the increase in bHb concentration [
29].
The adsorption kinetics of bHb adsorption on S
FP presented in
Figure 2 indicate the high affinity of protein to silica particles.
An almost linear increase (
Figure 2, inset) in protein adsorption during the first 60 min confirmed the high affinity of bHb to silica at pH 6. Protein adsorption in excess of 80% at a rate of 0.58 mg/min occurred within 120 min. These studies were restricted to 4 h due to an issue with bHb stability under agitation, as reported by Bhomia et al. [
27,
28]. The adsorption of bHb on S
FP was obtained to be of pseudo first order (R
2 = 0.9532), indicating that protein immobilization was primarily dependent on the adsorbate concentration in the media and its propensity to migrate to the adsorbent surface [
30].
2.2. Effect of Protein Adsorption on Porosity and Pore Volume
Syloid
® excipients are micronized synthetic amorphous silica gels of high purity that are widely used in pharmaceutical formulations [
31,
32]. The S
FP particles are pharmacopeia accepted and FDA-compliant nonordered mesoporous particles with an average size of 3.5 μm [
33]. Micrographs (
Figure 3) of S
FP particles confirmed that they were compacts or agglomerates of irregular shapes with the presence of a large volume of empty spaces and interconnected pores.
The complete nitrogen adsorption and desorption isotherm for S
FP presented in
Figure 4 resembled the BDDT Type IV, indicative of the mesoporous nature of the particles. The hysteresis in this isotherm resembled Type H1, meaning that these particles consisted of agglomerates or compacts of approximately uniform spheres in a regular array [
34]. BET isotherm of bHb-adsorbed silica retained its Type IV nature but showed a decrease in specific surface area. The BET surface area of S
FP decreased from 307 m
2/g to 226 m
2/g after bHb adsorption.
The average pore diameter and cumulative pore volume of S
FP as calculated by the BJH model were 160 Å (16 nm) and 1.16 mL/g, respectively, as shown in
Figure 5 [
35]. A decrease in both the average pore diameter (150 Å) and the cumulative pore volume (0.9 mL/g) was observed after bHb immobilization on silica particles.
The availability of a large available surface area after bHb adsorption may suggest that silica surface or pores within remained unoccupied by protein molecules. However, it is important to remember that the smaller pores which were easily accessible to nitrogen molecules during the BET analysis could be completely inaccessible to large molecules such as bHb. Larger available pores may potentially be able to accommodate more than one molecule within them but pores similar to or smaller than the dimensions of bHb (5.3-5.4-6.5 nm) largely remain unoccupied during the adsorption process. The decrease in the surface area and pore volume clearly indicates the presence of bHb inside the pores or the blockage of pore openings by protein molecules after the adsorption process. Similarly, an increase in the point of zero charge (PZC) for silica from pH 1.6 for bare particles to 6.7 for protein immobilized particles (closer to the IEP of bHb) also indicated coverage of particle surface with bHb molecules.
2.3. SCDDS Preparation
The preparation of SCDDS involved protein immobilisation on a solid surface followed by coating to control the drug release. The protein immobilisation on a solid surface is known to reduce the molecular flexibility, leading to enhanced stability [
22]. Moreover, coating thereafter ensures that these particles have optimal enteric properties to protect the protein from gastric pH and control the release [
36]. The SCDDS formulation process can be summarised as follows:
As illustrated in
Figure 6, the first step in this process allows the bHb molecules from the aqueous solution to be adsorbed on the silica surface which is then collected via centrifugation and dried to obtain bHb-immobilised S
FP particles. The protein adsorbed particles are then coated with MA via melt-deposition method in scCO
2 to limit possible denaturation due to oxidation, hydrolysis and solvent, stress and temperature-induced degradation. The rationale behind using MA coating is also to protect the protein from gastric media due to its limited solubility in acidic conditions and to promote drug release in intestinal environment.
S
FP, S
FP-bHb and SCDDS were analysed using ATR-FTIR to confirm protein adsorption, coating and whether bHb went through any conformational changes due to immobilisation at the given conditions. The spectra presented in
Figure 7 show amid I and amid II peaks at 1644 and 1530 cm
−1, respectively.
Amide I and amide II bands are two major bands of the protein infrared spectrum [
37]. The amide I band (between 1600 and 1700 cm
−1) is directly related to the backbone conformation and it is mainly associated with the C=O stretching vibration. The N-H bending vibration and the C-N stretching vibration results in amide II between 1450 and 1550 cm
−1. This amide II band is conformationally sensitive and can provide information about protein folding/unfolding. Both peaks are still present in bHb immobilised S
FP with reduced peak intensity but without any shift, confirming the lack of changes to protein conformation due to immobilisation. The absence of characteristic peaks (amide I and II) of protein and the emergence of the peak for carboxyl groups at 1701 cm
−1 along with the reduction in the intensity of siloxane (Si-O-Si) band at 1055 cm
−1 demonstrate MA coating of the silica surface.
2.4. In-Vitro Release Studies at pH 1.2 and 6.8
One of the major issues with protein immobilisation on mesoporous surfaces is the inability to obtain sufficient desorption thereafter due to the increase in the surface free energy. Hence, addition of a displacer in the dissolution media was explored to enhance the protein release from these systems. Although protein desorption can be obtained via changes in solvent ionic strength, pH and/or use of surfactants, it is important to carefully consider the choice of the displacement mechanism to ensure conformational integrity [
38,
39,
40,
41,
42]. Pluronics are triblock polymers consisting of polyethylene and polypropylene blocks and act as non-ionic surfactants. Prior to release studies, experiments were performed to determine the efficiency of pluronic F127 (F127) as protein displacer and its impact on the protein conformation.
Figure 8 shows the effect of F217 concentration on the protein desorption from silica surface.
The increase in F127 concentration in the media resulted in higher protein release of up to 1 mg/mL with no real change to the release profile with the further increase of F127 to 5 mg/mL. The effect of F127 in the media on protein conformation was evaluated using CD spectroscopy, which showed no impact on the secondary structures of the bHb, as presented in
Figure 9, in terms of the percentage content of five different secondary structures. These were calculated using CDNN software developed by Dr. Gerald Böhm (Institut für Biotechnologie, Martin-Luther Universität Halle-Wittenberg), which deconvolutes the CD data by cross-referencing the sample spectrum with already installed reference spectra [
28].
Ionic surfactants are known to cause protein unfolding via hydrophobic and ionic interactions [
43], whereas non-ionic surfactants such as pluronics have a limited effect on the protein conformation. Pluronics have also been studied as permeation enhancers which can potentially aid in the absorption of biomolecule after release [
44]. In this study, PF127 was used as a displacer in the media with the aim to formulate SCDDS containing both MA and pluronic in the future to obtain a drug delivery system capable of providing sufficient release at desired conditions. Based on the findings of these experiments, bHb release from SCDDS was determined with 1 mg/mL in the dissolution media.
One of the major goals of this study was to determine whether MA coating can provide sufficient protection to a biomolecule from the harsh gastric pH. The MA coated SCDDS did not show any protein release in the first 120 min at pH 1.2 (
Figure 10), suggesting that these formulations possess required enteric property to provide sufficient protection to protein from the gastric environment.
These findings were similar to the results from Pettit et al. in which systems prepared by solvent coating of MA showed release of adsorbed protein only at alkaline pH and Cerchiara et al. also showed that FA coating imparted enteric properties to vancomycin loaded chitosan particles prepared by freeze drying [
22,
36]. bHb release upon exposure to SIF was immediate, which may be related to the solubility behaviour of MA. The solubility of MA increases at a higher pH, which can subsequently facilitate the continuous erosion of coating and higher protein release at the studied conditions [
45]. bHb release in SIF was the highest (71%) from S
FP:MA0.5-PEN, followed by S
FP:MA0.5-SCF, S
FP:MA1.0-PEN, and S
FP:MA1.0-SCF with a release of 62%, 56% and 48%, respectively.
Although SE coating imparted the required enteric properties to the formulation, SCDDS prepared by SCF processing showed a comparatively lower and slower release in comparison to particles coated by SE. This could be due to the better coating and improved surface coverage of MA when SCF-assisted melt-coating was used in comparison to SE. The protein release was also dependent on the MA/silica ratios, where 0.5:1 showed a higher release compared to 1:1 for SCDDS prepared using both coating methods. For instance, SFP:MA0.5-SCF released a total of 62% bHb, which was 14% higher than the formulation containing a 1:1 (MA:silica) ratio. Similarly, the total release from SCDDS prepared by solvent evaporation was 56% (1:1) and 71% (0.5:1). The increase in FA ratio in the formulation resulted in a decreased bHb release, which is expected to have been due to the slow erosion of the MA layer above the solubility limit. Moreover, SCF processing of MA also allowed the coating to have been performed at 43 °C rather than its actual melting point of 54.5 °C, which occurred due to the dissolution of CO2 in a myristic acid crystalline matrix. This also ensured that bHb was not exposed to comparatively higher temperature which would be the case if melt-coating was performed at atmospheric pressure.
The dissimilarity (F1) and similarity (F2) factors were calculated to determine whether there were any statistical differences (
Table 1) between the bHb release from coated and uncoated particles.
The F1 should be between 0 and 15, whereas F2 should be between 50 and 100 for two dissolution profiles to be considered similar [
46]. For all the above formulations, F1 was higher than 15 and F2 was lower than 50. Hence, it can be concluded that the release from formulations was dissimilar to the control in every scenario.
Improper design or formulation of biologics can result in degradation, denaturation, and/or aggregation of the protein molecules, potentially causing both immunogenic side effects after administration and loss of pharmacological activity. Hence, the conformation of the released bHb was determined to understand whether the immobilisation, coating or release caused any changes to the protein’s secondary structure.
The CD spectra of the untreated and released protein from SCDDS prepared via SCF and SE methods are presented in
Figure 11. The CD spectrum of the released protein shows similar maxima and minima to untreated bHb, indicating that it largely retained its confirmation after release [
47]. The obtained CD spectra were processed by the CDNN software to determine the fraction content of different secondary structures of the protein. The percentage content of secondary structures from CD spectra for protein released from MA coated formulations is shown in
Table 2.
The α-helix content of the protein released from formulations was the same as the freshly prepared bHb solution, i.e., approximately 60%. Similarly, the rest of the secondary structures content was also comparable to the untreated sample, confirming the absence of any conformational changes in the bHb molecules either due to processing or exposure to the release media.