Effects of buffers on spray-freeze-dried/lyophilized high concentration protein formulations

Abstract

Solid-state protein formulations are known to exhibit enhanced storage stability compared to their liquid dosage form counterparts. pH is one of the factors affecting the stability of protein formulations. The pH of protein formulations in the solution is influenced by the buffer used, directly impacting their solid-state stability. During lyophilization, buffer components may interact with other formulation components present in the protein formulations, causing a pH shift. This study aimed to investigate the effects of phosphate buffer and amino acid buffers (such as histidine and/or arginine) on the physical properties and accelerated storage stability of spray freeze-dried or lyophilized protein formulations. A model protein, bovine serum albumin (BSA), was used to prepare high-concentration protein formulations. The formulations consisted of BSA, trehalose, and mannitol in an 80:15:5 ratio (w/w), respectively. Various buffers were utilized in the preparation of protein formulations, and the resultant solid formulations underwent screening via accelerated stability study using size exclusion chromatography (SEC). The combination of phosphate and arginine buffers resulted in increased monomer loss in the accelerated storage stability study. Additional characterizations, including solid-state Fourier transform infrared spectroscopy (ssFTIR) and powder X-ray diffraction (PXRD), were conducted. While these analyses did not definitively elucidate the mechanism behind the observed instability, their outcomes provide valuable insights for further investigation, highlighting the need for future research in this area.

1. Introduction

Biopharmaceutical research has been evolving after the 1950s, including protein formulations, gene therapies, cell therapies etc. [1]. Biopharmaceuticals are often made up of proteins with complex structures. These proteins in biopharmaceuticals are necessary for their functionality and are prone to degradation by several pathways like physical degradation involving aggregation or phase separation, and chemical degradation involving several hydrolytic and redox reactions [2, 3]. Depending on the type of physical/chemical changes protein denaturation may be reversible or irreversible [4], either of which may change the efficacy of the protein formulations.

Aggregation of protein is a type of physical instability where the single complex structure forms an oligomer, either through covalent or non-covalent bond formation [5]. This aggregation process is due to interactions among the hydrophobic and hydrophilic functional groups present on the exterior or interior of the protein structures [6]. The change in protein’s native state (partial unfolding) has also been observed as a driving force for aggregation process [7]. Several factors such as change in temperature and pH, or stresses induced during manufacturing or storage can promote protein instability [8, 9]. These challenges can be addressed by formulating biologics with various stabilizing excipients such as buffers, cryoprotectants, bulking agents, surfactants, and other additives [10, 11].

Solid-state protein formulations are favored over their liquid counterparts due to their capacity to decrease molecular mobility and water activity, thereby extending the shelf life of proteins [12, 13]. Various drying techniques are employed, such as lyophilization, spray drying, and spray freeze-drying [14]. Spray drying is a continuous heat assisted drying process which enables the formation of dry powder biologics [14, 15]. Lyophilization (or freeze-drying) and spray freeze-drying employ ultra-low temperatures and pressures to sublimate water to manufacture solid-state formulation [16]. Spray freeze-drying (SFD), a relatively novel technique, is particularly valuable for producing powders of temperature-sensitive biopharmaceuticals ensuring lesser heat stress on proteins than the spray drying process. In contrast to lyophilization, SFD enables the creation of spherical and porous particles with controlled sizes. During SFD, the liquid feed is sprayed into a cryogenic medium, facilitating rapid freezing of the protein formulation into its native state while preserving its structure. Subsequently, residual moisture is eliminated through the lyophilization process [14]. A notable advantage of SFD is the formation of highly porous spherical particles, which could enhance the powder flow properties [17]. These characteristics of SFD could be beneficial for drying high concentration protein formulations.

The term “High Concentration Protein Formulation” (HCPF) lacks a specific definition or regulatory guidelines from the Food and Drug Administration (FDA). It is noteworthy that HCPFs extend beyond the mere quantitative measurement of protein content, encompassing various attributes such as protein solubility, phase separation, and colloidal properties. Notably, a substantial portion of commercially available HCPFs comprises antibody products, typically falling within the range of 100–200 mg/ml [18]. Protein concentration can alter the aggregation kinetics and nucleation, thereby impacting stability [19]. Some of the research studies suggested that aggregation in protein could be reduced at higher protein concentrations [20]. This happens due to the increased resistance of protein towards denaturation and unfolding caused during processing (e.g., during freezing) [21, 22]. This was further known to be due to the reduction in protein-ice-water surface interaction [23]. While other study suggested that high protein concentration could cause an elevation in protein aggregation during lyophilization [24].

Solid-state high concentration protein formulations pose significant processing challenges such as high viscosity (15–25 cP) and longer reconstitution time [25–29]. The abundance of protein limits the inclusion of protective excipients, necessitating the use of multifunctional excipients. Amino acids such as histidine and arginine are commonly employed in FDA approved HCPFs due to their roles as buffering agents, viscosity reducers and protein stabilizers [18, 30]. Histidine has been shown to shield hydrophobic regions of the protein, whereas arginine increases the required free energy for initiation of protein aggregation [31, 32]. Additionally, histidine and arginine have been demonstrated to reduce the viscosity of HCPFs in some instances. [33–35]. Various combinations of buffers are utilized in FDA-approved HCPFs [18]. Optimum pH is crucial for protein stability, as alterations can lead to partial or complete unfolding of the protein [36, 37]. Selection of buffers is a crucial step as some buffers have been observed to crystallize during the freeze-drying process, potentially affecting overall stability by causing phase separation [38, 39]. Additionally, some buffers can interact with stabilizing excipients such as sugars present in the formulation [40, 41]. Moreover, changes in buffer composition can influence changes in particle formation, thereby affecting surface area and overall stability [42].

The aim of this study is to investigate the impact of amino acid buffers, potassium phosphate buffer, and their combinations on the stability of HCPFs. The potassium phosphate buffer was chosen over the sodium phosphate buffer, as sodium phosphate buffer tends to crystallize during freezing, while potassium phosphate buffer provides more stability for the protein upon lyophilization [37, 43, 44]. Commonly used amino acid buffers such as histidine and arginine, with chloride as the counter-ion were used. The rationale for using chloride counter-ion was that it has shown better compatibility with these amino acid buffers thereby enhancing protein stability [45]. Bovine serum albumin served as the model protein, while stabilizers like trehalose and mannitol were incorporated to protect the protein in the solid state and mitigate stresses during freezing and drying steps. Detailed formulation information is provided in Table 1. The formulations were prepared at a pH of 6.8 as the isoelectric point of BSA is in the range of 5.1–5.5 [46]. These formulations underwent drying via spray freeze drying (SFD) or lyophilization, and various physical characterizations including particle size and morphology, BET surface area, true density, and crystallinity were assessed. Furthermore, 90 days accelerated stability was conducted at 40°C, with monomer loss analyzed using size exclusion chromatography (SEC). Changes in stability were investigated by assessing alterations in pH, secondary structure of the protein, and residual moisture content of the dried products.

Table 1:

Formulation design.

#	Trehalose (mg/ml)	Mannitol (mg/ml)	BSA (mg/ml)	Potassium phosphate buffer pH 6.8 (mM)	L-Histidine monohydrochloride monohydrate buffer pH 6.8 (mM)	L-Arginine monohydrochloride monohydrate buffer pH 6.8 (mM)
1	15	5	80	-	-	-
2				50	-	-
3				-	50	-
4				-	-	50
5				25	25	-
6				25	-	25
7				-	25	25

2. Material and methods

2.1. Material:

Bovine serum albumin (BSA≥98%), trehalose dihydrate, D-mannitol, potassium phosphate monobasic monohydrate, potassium phosphate dibasic trihydrate, L-histidine monohydrochloride monohydrate, and L-arginine monohydrochloride were obtained from Millipore Sigma (St. Louis, MO, USA). 30 mL Slide-A-Lyzer^™ Dialysis Cassettes, 10K MWCO were obtained from Thermo Fischer Scientific (Waltham, MA, USA) for performing dialysis. The formulation details are provided in Table 1. Each buffer was prepared by dissolving the specific buffer salt (concentrations for each buffer salt mentioned in Table 1) in de-ionized water and then the pH was adjusted to 6.8 using hydrochloric acid or potassium hydroxide. The buffers were filtered using a 0.2-micron filter and later BSA was dialyzed using a specific buffer used in each formulation, while the control formulation was dialyzed using de-ionized water. For dialysis, 25 mL of 160 mg/mL of BSA stock was prepared in specific buffer and at concentrations mentioned in Table 1 (de-ionized water in case of the control formulation). The dialysis cassettes were loaded and suspended in the specific buffer (4L) present in the formulations as per Table 1, over a period of 24 hours at 5°C. After 24 hours the dialyzed protein in buffer/de-ionized water was diluted to 80 mg/mL using mixture of trehalose and mannitol in specific buffer/deionized water to achieve final target concentrations mentioned in Table 1. All the formulations were freeze-dried in 20R type I borosilicate glass vials; 20 mm two-legged chlorobutyl stoppers; 20 mm aluminum crimps were used to ensure the seal. For the accelerated stability study, 2R type I borosilicate glass vials, 13 mm two-legged chlorobutyl stoppers and 13 mm aluminum crimps were used.

2.2. Drying processes for Spray freeze-dried (SFD) and Lyophilized (LYO) formulations

Spray freezing was executed using an ultrasonic nozzle (Buchi, New Castle, DE). The liquid formulation was fed to the nozzle at a flow rate of 2 ml/min and sprayed into a borosilicate glass beaker filled with cryogenic medium (liquid nitrogen). Following the evaporation of the cryogenic medium, the frozen particulates were subsequently transferred to 20 ml lyophilization glass vials, ensuring that the fill volume for all formulations remained below 50%. These vials were subsequently processed using a laboratory-scale lyophilizer (Revo^®, MillRock Technology, Kingston, NY). The lyophilization parameters were adapted from the previously published drying cycles [47]. The shelf temperature while loading the formulations was maintained at −30°C, the shelf was sealed under vacuum and was equilibrated at −30°C for 1h. The primary drying started by dropping the pressure to 100mTorr while gradually increasing the temperature to −20°C for 1h. The primary drying was conducted at −20°C for the next 40h. While maintaining the pressure at 100mTorr, the secondary drying was initiated by ramping the temperature to 25°C at the rate of 0.15°C/min. The system was held at 25°C for the next 10h, then the dried samples were stoppered inside the lyophilizer and crimped for future characterization. Both SFD and LYO formulations were dried using the same lyophilization parameters.

2.3. Karl Fischer (KF) Moisture Content Analysis

The determination of residual moisture content was conducted through coulometric titration using the 917 KF Coulometer (Metrohm, Riverview, FL, USA). For each formulation, three replicates of 2–5 mg powder were individually weighed. Subsequently, 1 ml of anhydrous methanol (DriSolv^®, Millipore Sigma St. Louis, MO, USA) was added to each sample. The resulting mixture was then introduced into the Coulomat reagent (Honeywell Research Chemicals, Seelze, Germany) via the septum on the titration cell. The titration process was allowed to proceed until reaching the endpoint, at which point the moisture content in parts per million (ppm) was determined. This value was recorded, along with the corresponding sample weight and the weight of the injected solvent to calculate the moisture content in %w/w.

2.4. pH Study

The pH of all formulations was assessed using an InLab Micro pH electrode (Mettler Toledo, Columbus, OH, USA). pH values for each formulation were measured at three distinct stages: pre-freeze-drying, post-freeze-drying, and upon completion of the 90-day accelerated stability study. In the case of dried samples, the pH analysis was conducted by reconstituting the samples with an appropriate amount of solvent.

Powder X-ray diffraction (PXRD)

Both spray freeze-dried and lyophilized solids underwent an examination for the presence of crystallinity using the Rigaku SmartLab X-ray diffractometer (The Woodlands, TX, USA). This instrument is equipped with a highly sensitive D/tex ultra-detector and utilizes a Cu-Kα radiation source. The analysis involved solids was performed with a step size of 0.02° and a scanning rate of 5°/min, covering a 2θ range from 4 to 40°. The voltage for this X-ray diffraction analysis was set at 40 kV with a current of 44 mA.

2.5. Solid-state Fourier Transform Infrared Spectroscopy (ssFTIR)

Spectral analysis of the dried formulations was conducted using the Nicolet Nexus spectrometer (Thermo Scientific, Waltham, MA, USA) operating in attenuated total reflectance (ATR) mode with Smart iTR functionality. The absorbance mode was employed with a spectral resolution of 4 cm⁻¹, and data acquisition spanned the range of 800 to 4000 cm⁻¹. The obtained spectra were subjected to second-derivative calculation using OPUS 6.5 analysis software (Bruker, Billerica, MA, USA).

2.6. Particle Morphology and Particle Size

Particle morphology analysis was conducted using Scanning Electron Microscopy (NOVA nanoSEM, FEI Company, Hillsboro, OR, USA). Acceleration voltage of 5.0 kV was employed to examine the morphology of the formulations. All samples were affixed onto a double-sided carbon tape coated with platinum film, achieved at 40 mA for 120 seconds utilizing a sputter coater (208 HR, Cressington Sputter Coater, Watford, UK). Furthermore, for the spray freeze-dried formulations, particle size analysis was performed using the ImageJ software platform (National Institute of Health, Rockville, Maryland, USA) [48]. The particle size distribution was determined by analyzing the percentile of particle diameter at 10%, 50%, and 90% (referred to as D₁₀, D_50, and D₉₀).

2.7. Accelerated Stability Study using Size Exclusion Chromatography (SEC)

Aggregation of proteins represents a form of physical instability often linked with partial unfolding or colloidal instability, leading to the formation of high molecular weight species. These aggregates can be quantified using Size Exclusion Chromatography (SEC). For the stability study, samples were sealed and subjected to storage at 40°C for a duration of 90 days, with analyses conducted at 0, 30, 60, and 90-day intervals. The determination of the monomer to aggregate ratio of Bovine Serum Albumin (BSA) was carried out. Five replicates of each dried formulation were utilized for this analysis. Before analysis, the dried formulations were reconstituted using deionized water to achieve a final protein concentration of 1 mg/ml. The reconstituted samples were then analyzed utilizing a High-Performance Liquid Chromatography (HPLC) system, specifically the Agilent 1260 Infinity series (Santa Clara, CA), equipped with a TSK gel G3000SWXL column (Tosoh Bioscience LLC, King of Prussia, PA, USA). The mobile phase used was 100mM sodium phosphate buffer (pH 6.8) at the flow rate of 1 ml/min. The UV absorption wavelength for BSA to obtain chromatograms was set at 280nm. Subsequently, the area under the curve (AUC) for both the aggregate and monomer peaks was determined. The total monomer loss over the 0–90 day period was also assessed.

2.8. Surface Area and True Density

The surface area of the spray freeze-dried and lyophilized formulations was determined utilizing the 6-point BET method. Prior to analysis, the samples underwent a 24-hour degassing period at room temperature under nitrogen gas. Subsequently, analysis was conducted using nitrogen adsorption with a Tristar II 3020 instrument (Micromeritics, Norcross, GA, USA). Furthermore, the true density of the dried formulations was assessed using an AccuPyc II 1340 pycnometer (Micromeritics, Norcross, GA, USA).

2.9. Reconstitution time

The reconstitution process for both spray freeze-dried and lyophilized formulations involved weighing the appropriate quantity (≈50mg) of dried formulation necessary to prepare a 500μL liquid formulation, accounting for the specified solid content outlined in the formulation chart. Subsequently, 500μL of sterile water for injection was added to the measured solids, followed by gently swirling the vials once. The time taken for the formation of a clear solution was then recorded.

2.10. Inverse Gas Chromatography (iGC)

Inverse Gas Chromatography Surface Energy Analyzer (iGC-SEA) system (Surface Measurement Systems Ltd., Allentown, PA, USA) was used to analyze the specific surface energy (γ^sp) and total surface energy (γ_t) for both SFD and lyophilized formulations. iGC glass columns were used with inner diameter of 2mm and 4mm for SFD and lyophilized (LYO) formulations, respectively. Approximately 20mg for SFD and 200mg for LYO dried formulations were filled into the iGC column and both the ends were plugged with silanized glass wool. The column was conditioned using helium gas at 40°C for 4h before injection. The Relative Humidity (RH) was set at 0% with a gas flow rate of 10 sccm. Methane was injected before and after each measurement to calculate dead volume. All measurements were performed at column temperature of 25°C and a gas flow rate of 10 sccm. The hydrophilicity index (wettability) was calculated using Eq. [1].

$$Hydrophilicityindex=\frac{\mathrm{\gamma }\mathrm{s}\mathrm{p}}{\mathrm{\gamma }\mathrm{t}}$$ [1]

2.11. Statistical analysis

One-way Anova statistical analysis was performed using GraphPad Prism software version 9 (GraphPad, La Jolla, CA, USA) where p-values were obtained. (**** stands for p<0.0001, *** for p<0.001, ** for p<0.01 and * for p<0.05)

3. Results

3.1. Moisture content

The moisture content of both lyophilized (LYO) and spray freeze-dried (SFD) formulations was assessed at 0-day and 90-day intervals during storage at 40°C. At 0-day, the moisture content of LYO formulations was consistently below 3% (w/w), while SFD formulations exhibited levels below 4% (w/w). Over the 90-day storage period, a notable rise in moisture content was detected.

3.2. Particle Morphology and Particle Size

No collapse was observed in any of the lyophilized formulations. The observed microscopic structures of these cakes did not have a defined particulate structure. They showed a large flaky cake-like structure, while the SFD formulation has a spherical particulate structure. The particle size for SFD formulations was analyzed and observed to be 30–37 microns (Table 2).

Table 2:

Particle sizes of the SFD formulations (n=3).

Formulations	D10 (μm)	D50(μm)	D90(μm)	Span
No Buffer control (NB)	20.5 ±0.6	33.6 ±1.8	57.8 ±3.8	1.1 ±0.1
Phosphate Buffer 50mM (PB)	21.5 ±0.7	33.0 ±0.8	58.1 ±1.5	1.1 ±0.1
Histidine HCL Buffer 50mM (H)	19.8 ±0.5	32.9 ±0.6	57.3 ±2.3	1.1 ±0.1
Arginine HCL Buffer 50mM (A)	15.9 ±3.9	32.0 ±5.3	56.9 ±4.4	1.2 ±0.2
His+Arg Buffer 25mM each (HA)	16.1 ±3.4	29.6 ±3.9	47.5 ±5.1	1.1 ±0.1
Phos+His Buffer 25mM each (PH)	26.1 ±1.4	35.7 ±3.3	60.7 ±2.3	1.0 ±0.2
Phos+Arg Buffer 25mM each (PA)	24.2 ±3.3	37.2 ±1.7	54.3 ±3.1	0.9 ±0.2

3.3. Surface area and true density measurements

Figure 3A describes the surface area measurements utilizing the 6-point BET method, revealed that SFD formulations possess surface areas at least 20 times greater than the lyophilized cakes of the corresponding formulations. Specifically, in SFD formulations, those containing a combination of phosphate and histidine, as well as formulations containing histidine alone, exhibited significantly higher surface areas compared to other SFD formulations. Furthermore, the true density of the SFD formulations was notably higher (12–39%) than that of the lyophilized counterparts (Figure 3B).

Figure 3:

Surface area [A] and true density [B] measurements.

3.4. pH Study

The pH of liquid formulation before and after drying at 0-day and 90-day time points was analyzed for both SFD and LYO formulations. There was no significant difference (p>0.9698) in pH observed after drying and at 90-day interval. Table 3 shows the measured pH values.

Table 3:

pH measurements for SFD and LYO formulations

Formulations	Pre-LYO	Post-LYO	90 days LYO	Pre-SFD	Post-SFD	90 days SFD
No buffer control	6.90	6.86	6.89	7.00	6.92	6.97
Phosphate buffer 50mM	6.80	6.80	6.80	6.80	6.78	6.87
Histidine HCL buffer 50mM	6.80	6.50	6.58	6.80	6.55	6.48
Arginine HCL buffer 50mM	6.80	6.88	6.87	6.80	6.83	6.86
His+Arg buffer 25mM each	6.80	6.59	6.55	6.80	6.70	6.55
Phos+His buffer 25mM each	6.80	6.90	6.98	6.80	6.70	6.73
Phos+Arg buffer 25mM each	6.80	6.85	6.95	6.80	6.82	6.86

3.5. Powder X-ray Diffraction (PXRD)

Figure 4 describes the powder X-ray diffractograms. Analysis was conducted both immediately post-drying and following a storage period of 90 days at 40°C. Despite these conditions, the stability study revealed the absence of apparent crystalline peaks. Notably, all spray freeze-dried (SFD) and lyophilized (LYO) formulations maintained their amorphous state throughout the storage period according to PXRD.

Figure 4:

X-ray diffractograms for LYO [A] and SFD [B] formulations upon storage at 40°C (Y-axis adjusted using arbitrary units (A.U.))

3.6. Secondary structure analysis

ssFTIR analysis was conducted at 0-day and 90-day to determine changes in the secondary structure of protein (e.g. peaks at ~1656 cm⁻¹ indicate the potential existence of α-helix structures within the protein) [49]. The Amide I region was analyzed using second derivative calculations. There were no alterations observed in the α-helix structures within the protein in both LYO and SFD formulations.

3.7. Accelerated Stability Study

In this investigation, the protein aggregation was assessed during storage at 40°C for 90 days. The percentage of monomer loss was determined from the 0–90 day time span. For the lyophilized formulations (Figure 6A), a combination buffer of phosphate and arginine exhibited a significant increase in monomer loss compared to formulations with either histidine or arginine buffer alone. However, the lyophilized control formulation showed the highest level of aggregation, with no significant difference compared to the formulation containing phosphate buffer alone (p > 0.05). In the case of SFD formulation (Figure 6B), the combination buffer of phosphate and arginine resulted in the highest level of monomer loss compared to any other buffer containing SFD formulation. Notably, the SFD formulation with arginine buffer exhibited the highest protection against aggregation. Consistent with the observations in lyophilized formulations, histidine, and arginine buffers in SFD formulations demonstrated significantly better physical stability compared to the combination buffer of phosphate and arginine (p < 0.001). The formulations containing combination buffer of ‘histidine + arginine’ and ‘phosphate + histidine’ in both LYO (p<0.05) and SFD (p<0.0001) showed significantly lower % monomer loss compared to the ‘phosphate + arginine’ combination buffer formulation (Figure 6C & 6D). The initial (T₀) monomer content for SFD formulation ranges from 90–95%, this loss in monomer was not due to the formulation processing but was subjected to the differences in BSA stock received from the vendor.

Figure 6:

Percent monomer content for the lyophilized [A] and spray freeze-dried [B] formulations; percent monomer loss in 90 days at accelerated stability (40°C) for the lyophilized [C] and spray freeze-dried [D] formulations (**** stands for p<0.0001, ** for p<0.01 and * for p<0.05).

3.8. Reconstitution time study

The reconstitution time for lyophilized cakes was approximately 2.5 times faster than that for the spray freeze-dried counterparts. Among the SFD formulations, the formulation containing a combination buffer of phosphate and arginine exhibited the longest reconstitution time.

3.9. Hydrophilicity Index (wettability) using iGC-SEA

The hydrophilicity index for both LYO and SFD formulation was analyzed using iGC-SEA. The results in Table 5 demonstrate that the lyophilized cakes have higher surface hydrophilicity (wettability) compared to their spray freeze-dried counterparts.

Table 5:

Surface hydrophilicity index (wettability) for SFD and LYO formulations.

Formulations	Hydrophilicity Index
	LYO	SFD
No buffer control	0.13	0.06
Phosphate buffer 50mM	0.14	0.10
Histidine HCL buffer 50mM	0.12	0.02
Arginine HCL buffer 50mM	0.13	0.10
His+Arg buffer 25mM each	0.15	0.13
Phos+His buffer 25mM each	0.12	0.09
Phos+Arg buffer 25mM each	0.16	0.05

4. Discussion

This study consists of seven distinct formulations, comprising three different buffers, three combinations of buffer systems, and a control formulation containing no buffer. High concentration protein formulations were subjected to drying via either lyophilization or spray freeze drying. The investigation aims to elucidate the impact of altering buffer systems at the same pH on the accelerated stability of these high concentration protein formulations. To comprehend the underlying factors contributing to changes in physical stability, various physical characterizations were conducted, including analysis of the protein’s secondary structure, solid crystallinity, and monitoring of pH changes.

Change in moisture content of the dried formulation can directly impact the long-term stability of the protein present in the formulation [50, 51]. Moisture content analysis was performed before and after the accelerated stability study. Although some changes in moisture content were observed, these changes did not correspond to the changes in the physical stability of the dried formulations. The increased moisture content observed at 90-day time point could have been due to moisture transfer during the sample processing of dried product for the stability study. Following lyophilization, the lyophilized cakes were examined for any signs of visual collapse, considering previous studies indicated that the use of arginine hydrochloride buffer may lead to cake collapse or shrinkage [45, 52]. However, no collapse was observed in any of the lyophilized formulations, likely due to the high solid content present in the formulations [22, 53].

Particle size and surface area can significantly influence the stability of proteins within particles, as smaller particles tend to have larger surface areas, exposing more proteins to the surface and potentially compromising stability [54]. A study by Mutukuri et al. suggested that the use of different buffer systems could alter the particle formation during the drying step, consequently affecting the exposure of proteins to the surface [42]. Additionally, changes in buffer composition could influence ice crystal formation during the freezing step of lyophilization, thereby impacting the structure and surface area of the lyophilized cake [42, 55].

To investigate whether the observed changes in stability were attributed to variations in particle size, particle size analysis was performed for the SFD formulations, revealing a particle size (D50) close to 35 microns (Table 2). No significant differences in BET surface area were detected among the lyophilized formulations; however, for SFD formulations, those containing histidine, and a combination of phosphate and histidine buffer displayed a higher surface area. Nevertheless, this discrepancy did not align with the observed changes in physical stability data. In order to comprehend the material packing during the drying process, the true density of both SFD and lyophilized formulations was measured using a helium pycnometer. Since SFD particles typically have a highly porous interior, they typically exhibit higher bulk density [56, 57]. On the other hand, true density measurements exclude pores while assessing the density of dried formulations revealing the actual packing density [58]. Our findings indicate that despite the SFD formulations exhibiting a higher surface area, they demonstrated a higher true density compared to lyophilized cakes, implying denser material packing during the drying process.

Proteins can undergo unfolding in response to pH changes, with BSA exhibiting optimal stability near pH 7. A reduction in alpha helix domains is observed when the pH deviates from this range of 4–9 [36]. Hence, it was essential to investigate whether changes in physical stability could be correlated with alterations in pH. However, our data indicated no significant changes in pH for any of the SFD or lyophilized formulations (Table 3). This suggests that the observed stability changes within the formulations were not attributable to fluctuations in pH. In this study the pH was analyzed by reconstituting the dried product, but several studies have shown that microenvironment pH of protein formulations can change in freezing and drying step during lyophilization [59–61]. The changes observed in aggregation rate of formulations with different buffer systems could be attributed to a shift in pH during freezing step or in solid-state formulation before reconstitution. Self-buffering effect of proteins is typically observed at high concentrations [62, 63], in our study, no pH change was observed in no-buffer control formulations; however, this resulted in the highest monomer loss for both SFD and LYO formulations.

To evaluate the presence of crystallinity and its impact on the physical stability of the protein, powder X-ray diffraction (PXRD) analysis was conducted. Previous research has indicated that crystallinity can lead to phase separation of stabilizing excipients, consequently adversely affecting the protein [38, 64]. Studies have demonstrated that histidine and arginine, whether used as stabilizing excipients or buffers, can crystallize during freezing or drying processes [65, 66]. Our formulations included mannitol, typically added as a bulking agent to provide structural integrity for freeze-dried solids. Studies have shown that mannitol introduces some degree of crystallinity to formulations, which can shorten the reconstitution time of high concentration protein formulations [28, 67]. However, it is worth noting that several studies have indicated that upon drying, crystallized mannitol can affect the physical stability of the protein as it causes phase separation and disruption of protein-excipient interactions [38, 68, 69]. The XRD data revealed no presence of crystallinity after 90 days of storage for both SFD and lyophilized formulations. This suggests that none of the components within the formulation underwent significant crystallization that could potentially impact the physical stability of BSA.

Bovine serum albumin under stable conditions majorly consists of α-helix domains (53–66%), which are detected using ssFTIR and observed in Amide I region at wavenumber ~1656 cm⁻¹, upon instability the percentage of α-helix domain drops [70–72]. The reduction in α-helix domains also increases the formation of random coils, which can be detected at wavenumber ~1645 cm⁻¹ [73]. A study conducted with histidine and phosphate buffer revealed that the utilization of histidine buffer aided in maintaining the α-helix domains, whereas the use of phosphate buffer resulted in a significant reduction in α-helix domains compared to the native protein [65]. Our secondary structural analysis suggested no change in the α-helix domains (SFD and LYO), this suggests that the change in monomer content due to the presence of different buffer systems in the BSA formulations may not be attributed to the change in number of α-helix domains.

Our study highlighted that using histidine or arginine individually offers superior protection against aggregation compared to the combination of phosphate and arginine buffer system. Histidine and arginine buffers are commonly employed in high concentration protein formulations due to their multifunctionality as buffers, protein protective agents and lowering viscosity [18, 74–76]. Several molecular mechanisms for arginine have been proposed such as ‘cluster formation’, ‘neutral crowder’ and ‘preferential interaction’ through research studies, while histidine is proposed to have weak interactions towards the hydrophobic patches of the protein thus stabilizing it [30, 31, 77]. However, we need to understand that these proposed mechanisms have been studied in solution-state where the protein and excipients have more molecular mobility compared to the solid-state formulations. Further research in this area is needed to understand molecular behavior of amino acids in stabilization of biologics.

The combination of amino acid buffers with inorganic buffer salts has been explored in some approved high concentration protein formulations, potentially to enhance the buffer capacity of the amino acid buffers [18, 63, 78]. Our stability data indicates that the combination of phosphate and arginine buffers exhibits higher monomer loss (increased aggregation), yet the reason for the incompatibility of this combination remains unknown. A published study employing a combination of histidine and arginine buffer has shown an elevated aggregation rate upon using phosphate counterion, suggesting potential chemical interactions between arginine and phosphate affecting protein stability [45]. Another finding from our stability analysis of LYO formulations indicates that potassium phosphate buffer-containing formulations displayed a high aggregation rate, similar to the no-buffer control. This phenomenon could be attributed to minor crystallization or phase separation of buffer salt during the ice formation step of lyophilization [43, 79]. Conversely, this wasn’t observed in the SFD phosphate buffer formulation, as they undergo rapid freezing, allowing less time for phase separation in a particulate structure [80].

One of the major challenges for high concentration protein formulations is longer reconstitution times. While lyophilized cakes lack particulate structure, resulting in reduced surface availability for interaction with water, spray freeze-drying facilitates the production of highly porous particles, which typically aids in reducing reconstitution times by enhancing surface interaction with water [81–83]. Surprisingly, spray freeze-dried (SFD) high-concentration BSA formulations had notably longer reconstitution times compared to their lyophilized counterparts. Several potential reasons for this observation were considered: the true density analysis revealed that SFD formulations have denser material packing, making interaction with reconstitution fluid more difficult. Additionally, the formation of particles during SFD may trap air inside the tiny pores upon addition of reconstitution fluid, creating an air-water interface and thereby prolonging the reconstitution process. The surface hydrophilicity index (wettability) data demonstrated that the wettability reduced during the particle formation of SFD formulation by exposure of more hydrophobic domains towards the surface of the particle, which is a self-assembling behavior during atomization or particle formation [84, 85]. The lyophilization parameters used in the study were not specifically optimized for the SFD formulations, several studies have shown that annealing step can increase the ice crystal size, this in turn could be helpful to improve the reconstitution time of the SFD formulations [86, 87].

5. Conclusion

This study emphasizes that the utilization of different buffer systems and combinations thereof can significantly impact the long-term physical stability of high concentration protein formulations. Results indicate that certain combinations of buffer systems, such as phosphate and arginine, may result in higher protein aggregation, while amino acid buffers like histidine and arginine can offer protection against aggregation in the high concentration protein formulations. However, physical characterizations including particle size, surface area, true density, and residual moisture content were not found to correlated with the physical instability observed with the phosphate and arginine buffer. Secondary structural analysis, X-ray diffraction (XRD), and pH data did not have direct correlation with BSA physical instability either, suggesting that there could be some other underlying reason for the observed incompatibility in the combination buffer of phosphate and arginine. Further investigation is warranted to elucidate the interactions between phosphate and arginine and their overall effect on protein stability.

Figure 1:

Moisture content of SFD and LYO formulations at 0 and 90-day timepoint (n=3).

Figure 2:

Particle morphology of lyophilized cakes [A] and spray freeze-dried [B] BSA formulations.

Figure 5:

2^nd Derivative FTIR spectrums for LYO [A] and SFD [B] formulations. Solid line represents 0-days accelerated stability time point; Dashed line represents 90-days accelerated stability time point. Y-axis adjusted using arbitrary units.

Table 4:

Reconstitution time measurements for SFD and LYO formulations (n=3).

Formulations	Time (mins)
	Lyo	SFD
No buffer control	3.25 ±0.13	8.56 ±0.07
Phosphate buffer 50mM	2.50 ±0.28	7.38 ±0.12
Histidine HCL buffer 50mM	2.44 ±0.01	8.12 ±0.49
Arginine HCL buffer 50mM	3.23 ±0.16	8.56 ±0.14
His+Arg buffer 25mM each	2.58 ±0.49	6.45 ±0.14
Phos+His buffer 25mM each	2.49 ±0.05	8.45 ±0.11
Phos+Arg buffer 25mM each	3.32 ±0.24	14.20 ±0.12