INTRODUCTION

Peptides have become increasingly important as therapeutic products. Currently there are more than 60 approved peptide drugs, and an additional 130 peptide candidates are in clinical development (1). As is the case with protein therapeutics, peptide drugs are susceptible to both chemical and physical degradation, and stabilization of these products is challenging (2–5). One approach to achieving sufficient stability of a biological product is to develop a freeze-dried formulation (6,7). There is an extensive literature describing the capacities of various excipients to stabilize proteins during the freeze-drying process and long-term storage in dried formulations, as well as the mechanisms for such stabilization (6–8). For peptide drugs there is a more limited literature on stabilization in freeze-dried formulations (9–12). Therefore, the goal of the current study was to gain further insights into how the lessons learned from protein therapeutics apply to development of stable freeze-dried formulations of peptide drugs.

Typically, an essential stabilizer in a freeze-dried protein formulation is a nonreducing disaccharide such as sucrose or trehalose (6,7). These sugars can inhibit protein unfolding during the freezing and drying steps of freeze-drying, as well as provide a glassy matrix that is important for long-term storage stability of the dried product (6,7). For drugs that are formulated at acid pH, sucrose has the disadvantage of being susceptible to acid-catalyzed hydrolysis forming reducing sugars glucose and fructose,(13) which can chemically degrade proteins or peptides via the Malliard reaction (13). For example, the Malliard reaction has been reported when glucagon was formulated with the reducing disaccharide lactose (14). Trehalose is more resistant to acidic hydrolysis than sucrose and is more suitable for formulations at low pH (15,16).

The inclusion of a polymer such as hydroxylethyl starch (HES) to a lyophilized protein formulation may improve long-term storage stability of the protein in formulations that also contain sucrose or trehalose (17). The increased stability is observed because HES can form glassy matrix with very high Tg (i.e. >200°C). However, HES alone usually fails to confer stability to dried proteins because it does not inhibit protein unfolding during freeze-drying, and the rate of degradation during storage is greatly increased for unfolded proteins (6,7,17).

The non-ionic surfactants polysorbate 20 and 80 have been used extensively as excipients in freeze-dried and aqueous solution formulations of proteins due to their ability to reduce protein aggregation (18–20). This protective effect has been attributed to several different mechanisms including competing with protein molecules for interfaces, increasing thermodynamic stability of the native state through binding to the protein, fostering refolding and reducing the concentration of protein molecules in a stagnant boundary during rehydration of dried formulations (20). Cyclodextrins also have been shown to reduce protein and peptide aggregation by competing for the air-water interface (21) and by binding to the hydrophobic residues of proteins or peptides such as glucagon (14,22).

The potential stabilizing effects of each of these classes of protein stabilizers for small peptides in freeze-dried formulations are not well understood. Among therapeutic peptides, insulin has been most studied in dried formulations. Although insulin is different from most other peptides in that it has relatively well-defined tertiary and quaternary structures, it is still instructive to consider results for it in the broader context of peptide stabilization. Previous work has shown that insulin in freeze-dried formulations exhibited many of the same pathways of degradation (i.e. aggregation and deamidation) as in aqueous solution (9,10). In one study, the degradation rate was dependent on the pH of aqueous solution prior to freeze-drying as well as the water content of the freeze-dried cake (9,10). Covalent dimerization of human insulin was substantially decreased by incorporation into a glassy matrix of trehalose, presumably by inhibiting structural perturbation of the peptide, reducing molecular mobility in the dried formulation and physically separating the insulin molecules (10). In another study, insulin freeze-dried with trehalose exhibited a substantially lower local dynamics (β-relaxation) and lower degradation rates (i.e. deamidation and dimerization) than insulin freeze-dried with dextran (12). The authors speculated β-relaxation of insulin was reduced because of hydrogen bonding with trehalose, and thus chemical degradation was inhibited. On the other hand, dextran cannot form hydrogen bonds with insulin as readily due to steric hindrance, and therefore reduction of β-relaxation was not present in the dextran formulation.

The study of stability in freeze-dried formulation with other peptides is rather limited. For example, in one study, the level of mannitol freeze-dried formulations of atrial natriuretic peptide (ANP) affected the amount of multimers formed during storage. The authors speculated that in the less stable formulations, which had higher mannitol levels, mannitol crystalized and increased the water moisture content in the amorphous phase (11).

Glucagon (Fig. 1), the focus of the current study, is a polypeptide hormone composed of 29 amino acid residues that is currently used for the emergency treatment of insulin-induced hypoglycemia (23). Glucagon is known for its propensity to degrade both chemically (i.e., hydrolysis and oxidation)(24–26) and physically (i.e., aggregation)(27,28); including formation of aggregates during freeze-drying and rehydration in the absence of stabilizing excipients (29). Therefore, it is an excellent model peptide to evaluate the effects of different buffers, the surfactant polysorbate 20, trehalose, HES and β-CD on stability during freeze-drying and storage in dried formulations.

Fig. 1
figure 1

Primary structure of glucagon. The three Asp and Met residues are shown in bold.

Full size image

MATERIALS AND METHODS

Materials

All chemicals were of reagent grade or higher quality. Glucagon was purchased from the American Peptide Company (Sunnyvale, CA). β-CD was purchased from Cyclodextrin Technologies Development, Inc. (High Springs, FL). Trehalose dihydrate was purchased from J.T.Baker (Philipsburg, NJ). HES (Viastarch, MW 200KDa) was purchased from Fresenius (Graz, Austria). Phosphoric acid, citric acid, glycine, 2M hydrogen chloride, sodium chloride, sodium hydroxide, acetonitrile, hydrogen peroxide, potassium bromide and polysorbate 20 were purchased from Fisher Scientific (Hampton, NH). Nylon membrane filters (0.45 μm) were purchased from Whatman (Maidstone, England). Polyethylstyrene (PES) membrane filters (0.2 μm) were purchased from Millipore (Billerica, MA). 3-ml lyophilization vials (borosilicate glasses, type 8412-B) and gray butyl stoppers (Product # 10123524) were purchased from West Pharmaceutical (Lionville, PA). Other reagents and chemicals were purchased from Sigma-Aldrich (Milwaukee, WI).

Acidic Degradation of Glucagon in Aqueous Solution

Glucagon (0.5 mg/ml) was dissolved in 5 mM sodium phosphate buffer (pH = 3.0) and incubated for 5 days at 60°C. Aliquots (0.5 mL) of the solution were removed at various times and centrifuged at 14,500 rpm for 10 min. The supernatant was analyzed by reversed-phase high-performance liquid chromatography (RP-HPLC) as described below. Samples were also analyzed by electrospray ionization-mass spectrometry (ESI-MS; see below).

Hydrogen Peroxide-Induced Oxidation of Glucagon in Aqueous Solution

We followed the USP method (USP 30 Official Monograph for Glucagon, 2007) for forced oxidation of glucagon. Glucagon (1 mg/ml) was oxidized by incubation in 0.6% hydrogen peroxide (H2O2), 20% acetonitrile and 80% water with 10 mM HCl at 4°C for 210 min. Aliquots were removed at various times and analyzed by RP-HPLC using the method described below. The oxidation products were also analyzed by ESI-MS as well as tandem MS (MS-MS; see below).

ESI-MS and MS-MS Analysis of Glucagon and its Degradation Products

Glucagon and its degradation products were analyzed with an electrospray-triple quadrupole-time-of-flight mass spectrometer (ESI-qTOF-MS) from Applied Biosystems (PE SCIEX/ABI API QSTAR Pulsar i Hybrid LC/MS/MSESL). Mass Spectra were acquired by scanning a mass-to-charge ratios (m/z) range from 100 to 2000. Eluates (1 mg/ml glucagon, 50 μl) were injected into the mass spectrometer at a flow rate of 5 μl/min. Spray voltage was set at 4500 V, and capillary temperature was set at ambient temperature. For MS-MS experiments on glucagon and oxidized glucagon, various m/z components were selected and fragmented with suitable collision energy to have reasonable amount of peaks following fragmentation.

Sample Preparation for Freeze-Drying

Glucagon (10 mg/mL) was dissolved in three different buffers (glycine hydrochloride, sodium phosphate, and sodium citrate buffers, 5 mM, pH 3.0). The solution was then mixed in a 1:1 (v/v) ratio with various excipient solutions (prepared at twice the desired concentration using corresponding buffer) to obtain a final glucagon concentration of 5 mg/mL and the final desired excipient concentration of 0.01% for polysorbate 20 and 10 mg/mL for the carbohydrates. The solution was then filtered through 0.2 μm Millipore PES membrane. Sample preparation was conducted in a 4°C cold room. The glucagon concentration and the purity were determined by RP-HPLC (see below).

Freeze-Drying

The formulations were pipetted (0.3 mL) into 3-ml lyophilization vials (13-mm ID) and freeze-dried in a FTS Durastop freeze-drier (Stoneridge, NY). For freezing, samples were cooled to −40°C at 2.5°C/min and maintained at this temperature for 2 h. Then the shelf temperature was increased to -5°C at 2°C/min and held for 2 h as an annealing step (29). The temperature was then decreased to −30°C at 1.5°C/min and the chamber pressure was reduced to 8 Pascal. These conditions were maintained during 24 h for primary drying. Then the shelf temperature was increased to 40°C at 0.5°C/min and held at 40°C for 10 h for secondary drying. Then, the vials were stoppered under vacuum using gray butyl stoppers. None of the formulations showed any visual evidence of cake collapse following freeze-drying.

Storage Studies

Following freeze-drying, the sample vials were incubated at 60°C and analyzed after 2 weeks. Triplicate sample vials for each formulation were incubated. Control sample vials (equivalent to time zero of incubation) were stored at −80°C until analysis. Following storage, formulations were first rehydrated to 5 mg/mL with water and then diluted to 1 mg/mL glucagon with the corresponding buffer. An aliquot of the rehydrated sample (200 μL) was analyzed for turbidity (see below). Then, the solution was centrifuged (14,500 rpm × 10 min) to removed insoluble material. The supernatant was diluted 1:1 with the corresponding buffer for RP-HPLC and size exclusion (SE)-HPLC analysis (see below).

Differential Scanning Calorimetry (DSC)(30)

The glass transition temperature (Tg) of dried formulations was determined using a Perkin-Elmer Pyris-1 DSC (Norwalk, CT). Prior to sample analysis, the instrument was calibrated for melting temperature and heat of fusion with indium (onset of melting: 156.6°C; heat of fusion: 28.45 J/g). In a dry box, the sample vials were opened and about 1 mg samples of dry powder were placed in hermetically sealed aluminum pans. DSC thermograms were collected from 25°C up to 250°C (depending on the compositions of formulations) at a heating rate of 100°C/min. Baselines were determined using an empty pan, and all thermograms were baseline corrected. For measurement of Tg, the samples were first heated to above Tg to remove thermal history, and then cooled back to 25°C and rescanned at 100°C/min. The thermogram obtained during the second scan was used to measure Tg which was determined as the midpoint of the transition.

Infrared (IR) Spectroscopy

Infrared spectra of glucagon in aqueous solution and in freeze-dried formulations were collected at room temperature using a Bomem Prota spectrometer (Quebec, Canada) purged with dry nitrogen (31,32). Spectra of glucagon in aqueous solution were obtained at a peptide concentration of 20 mg/ml (5 mM sodium phosphate buffer, pH 3.0) in a 6 μm pathlength cell with CaF2 windows. For each sample, a 32-scan interferogram was collected in the single-beam mode with a resolution of 4 cm−1. Peptide spectra were processed to substract absorbance from water vapor and the buffer spectrum as previously described (31,32). For IR spectra of freeze-dried samples, a mass of dried formulation containing about 1 mg of glucagon was mixed with 500 mg potassium bromide powder. The mixture was ground gently and then annealed into a pellet using a hydraulic press (30). The spectra were transformed to second derivatives using Bomem Grams/32 software. The final protein spectra were smoothed with a seven-point Savitsky-Golay function. For comparison of spectra they were area normalized in the amide I region (1600–1700 cm−1) and overlaid (30).

Optical Density at 405 nm (OD405)

The presence of aggregates in rehydrated formulations (1 mg/mL glucagon) was assessed by measuring the OD405 using a Molecular Devices microplate reader (Sunnyvale, CA). The freeze-dried glucagon formulations were first rehydrated with water to 5 mg/mL glucagon and then diluted with corresponding buffer to 1 mg/mL peptide concentration. Prior to analysis, the plates were shaken for 10 s to mix the solutions in the wells.

Reversed Phase-High Performance Liquid Chromatography (RP-HPLC)

Chemical degradation of glucagon formulations was quantified by RP-HPLC, using an Agilent 1100 HPLC system (Agilent Technologies, Inc., Waldbronn, Karlsruhe, Germany) with a Thermo Biobasic C8 column (5 μm, 250 × 4.0 mm ID, Waltham, MA) and a Restek VWD G1314A detector (Bellefonte, PA). The mobile phase was 73% phosphate-cysteine buffer (pH 2.6) with 27% acetonitrile (USP 30 Official Monograph for Glucagon, pp 2230–2231, 2007). Before use, mobile phase was filtered with 0.45 μm Nylon membrane filters and degassed. The flow rate for the analysis was 1 ml/min, and elution was monitored at 214 nm. The temperature of the column was maintained at 37°C. The peak areas of glucagon and its chemical degradation products were used to determine the chemical degradation of glucagon occurring during freeze-drying and storage.

Size Exclusion-HPLC (SE-HPLC)

Aggregation of glucagon in different formulations was determined with SE-HPLC, using an Agilent 1100 HPLC equipped with a TSK G2000SWxl gel filtration column (5 μm, 300 × 7.8 mm ID) and a Restek DAD G1315A detector. The mobile phase was 3.2 mM HCl, 100 mM sodium chloride, pH 2.5. Before use, the mobile phase was filtered and degassed. The flow rate for the analysis was 1 ml/min and elution was monitored at 280 nm. The percent of monomer remaining in formulations was calculated based on the monomer area of the formulation samples compared to that for liquid control sample. Soluble aggregates were at minimal levels (<0.5%) in all of the samples (data not shown).

RESULTS AND DISCUSSION

Accelerated Acid Degradation of Glucagon in Aqueous Solution

Glucagon has an isoelectric point of about 7 and has a high solubility at pH values less than 3 or greater than 9. The recommended pH range for solutions of the peptide is between 2.5 and 3.0 (US pharmacopeia 24, 1999). However, as in the case with many peptide and protein drugs, glucagon is susceptible to acid-catalyzed degradations (i.e. Asp cleavage) (24–26). To develop an assay to characterize acid-catalyzed hydrolysis of glucagon, the peptide was first incubated in 5 mM sodium phosphate buffer (pH = 3.0) at 60°C and characterized as a function of incubation time. The commercial glucagon we used had some impurities prior to incubation (about 8–9%, see Fig. 3a and its enlarged version Fig. 3b). During incubation, glucagon underwent significant further chemical degradation, with 50% loss of the parent peptide after incubation for 5 days (Figs. 2 and 3d). As expected in the acidic conditions, the main degradation products were due to Asp cleavage, as confirmed by ESI-MS analysis (Table I). Glucagon has three Asp residues and, therefore, six main fragments were observed in ESI-MS.

Fig. 2
figure 2

Degradation of glucagon in 5 mM sodium phosphate buffer pH 3.0 at 60°C. The % remaining glucagon was determined from chromatograms obtained by RP-HPLC. Error bars indicating SD are smaller than the symbols.

Full size image
Fig. 3
figure 3

RP-HPLC chromatograms of glucagon and its chemical degradation products. (a) Control sample; (b) expanded scale of a; (c) oxidized glucagon by hydrogen peroxide; (d) degraded in 5 mM sodium sodium phosphate buffer, pH 3.0, at 60°C for 5 days; (e) following incubation of dry powder (freeze-dried in sodium phosphate buffer) at 60°C for 2 weeks (scale was expanded for improved visual resolution of peaks).

Full size image
Table I ESI-MS Analysis of Glucagon Hydrolysis Products After Incubating at 5 mM Sodium Phosphate Buffer, pH 3.0 at 60°C for 2 days
Full size table

In RP-HPLC (Fig. 3d), four main peaks (retention time (tR) around 7, 14, 19 and 29 min, respectively) and several smaller peaks (i.e. tR around 33, 37, 40, and 42 min) were observed. The main peaks were probably due to Asp cleavage whereas the smaller peaks could be due to Asp cleavage and/or deamidation (25,26). Glucagon has three Gln and one Asn residue. Deamidated residues were not specifically identified in the current study.

Accelerated Oxidation of Glucagon in Aqueous Solution

Although oxidation of Met does not impact the binding affinity of glucagon to its receptor (33), the potency the oxidized peptide in stimulating glucose production in isolated rat adipocytes and hepatocytes is decreased (34). Therefore, oxidation of the peptide should be monitored and minimized. To develop an assay for oxidized glucagon, the peptide was incubated in 0.6% (v/v) hydrogen peroxide at room temperature (Fig. 4). Samples were removed as a function of time and analyzed by RP-HPLC. Two new peaks with similar tR and almost identical areas were observed (tR around 8–9 min, Fig. 3c). Almost all of the parent peptide had decomposed after 210 min of incubation (Fig. 4). The oxidation products had molecular weights of [Glucagon + 16] (calculated mass = 3498.6; observed: [M + 2 H]2+ = 1750.2, [M + 3 H]3+ = 1167.1, [M + 4 H]4+ = 875.6, [M + 5 H]5+ = 700.7), suggesting one residue of glucagon was oxidized. MS/MS analysis (Table II) of the oxidation products showed the Met residue at position 27 was oxidized. The two peaks in the RP-HPLC chromatograms were probably due to two diastereomers formed with oxidation of the Met residue. Even though glucagon also has several other amino acids susceptible to oxidation (e.g. His, Trp, Tyr, and Phe), no degradation of these residues by hydrogen peroxide was observed.

Fig. 4
figure 4

Oxidative degradation of glucagon with 0.6% H2O2 at 4°C. The % remaining glucagon was determined from chromatograms obtained by RP-HPLC.

Full size image
Table II MS-MS Analysis of Glucagon and Oxidized Glucagon
Full size table

Effects of Buffers and Excipients on Glucagon’s Stability in Freeze-Dried Formulations

To determine formulation effects on stability of freeze-dried glucagon formulations, we evaluated effects of three buffers (glycine hydrochloride, sodium phosphate and sodium citrate buffers, 5 mM, pH 3.0), polysorbate 20 and three carbohydrate excipients, trehalose, HES and β-CD. An accelerated degradation condition (storage at 60°C for 2 weeks, followed by rehydration) was used to evaluate the stability of the peptide in the freeze-dried formulations. Important physical properties (peptide secondary structure and Tg) were also characterized for each freeze-dried formulation.

The Tg of the freeze-dried formulations was determined by DSC (Table III). When there was no excipient, the Tg of dry powders was between 122°C and 140°C, depending on the buffer used in formulation (Table III). As expected, the presence of polysorbate 20 (0.01% w/v) did not have much effect on Tg.

Table III Composition and Tg (°C) of Freeze-Dried Glucagon Formulations
Full size table

For formulations with glass forming carbohydrates, the Tg was highest for HES and β-CD, followed by the mixture of trehalose and HES, and then trehalose alone (Table III). These values were consistent with the Tg of the corresponding pure carbohydrates and their mass ratios in the formulations. The buffer also had a substantial effect on Tg, especially in the cases where HES and β-CD were used as excipients (Table III). The cause(s) of the buffer effect was not investigated further.

Infrared spectroscopic analysis of glucagon formulations was used to evaluate the effects of freeze-drying and storage on glucagon’s secondary structure, by comparing spectra for the peptide in the dried formulations to that for the native peptide in aqueous solution. The main secondary structure of native glucagon is a mixture of α-helix and random coil (35,36). Correspondingly, the infrared spectrum for native glucagon in aqueous solution has main absorbance around 1657 cm−1 (Fig. 5). The peak is broad, presumably due to high flexibility of backbone conformation.

Fig. 5
figure 5

Area-normalized second-derivative infrared spectra of native glucagon in aqueous solution and the peptide denatured in boiling water for 20 min followed by freeze-drying.

Full size image

For comparison of the native peptide spectrum to that of fully denatured and aggregated glucagon, the peptide solution was incubated in a boiling water bath for 20 min and then freeze-dried. It has been reported that upon fibrillation, the main secondary structure of glucagon is converted into intermolecular β-sheet (37,38). In the IR spectrum of boiled and freeze-dried glucagon the peak intensity around 1657 cm−1 was greatly decreased and a new peak around 1630 cm−1 was observed (Fig. 5), suggesting intermolecular β-sheet was formed between glucagon molecules during heating-induced aggregation.

For analysis of the freeze-dried formulations, first the infrared spectra of glucagon formulations freeze-dried in glycine hydrochloride buffer were studied (Figs. 6a and 7a). The dominant spectral feature affected by formulation was the intensity of the band at 1657 cm−1. Therefore this parameter was used to compare results for the different formulations (Fig. 7). Compared to the spectrum of the native, aqueous peptide, the band intensity was only slightly decreased in the spectrum for the peptide freeze-dried in buffer alone. There was also no increase of absorbance around 1630 cm−1—a strong band in the spectrum of denatured/aggregated glucagon—suggesting no significant amount of intermolecular β-sheet was formed during freeze-drying of glucagon in glycine buffer.

Fig. 6
figure 6

Area-normalized second-derivative infrared spectra of glucagon when freeze-dried in glycine hydrochloride buffer in the absence of additives (None) or in the presence of polysorbate 20 (PS 20), trehalose, HES, 1/1 mixture of trehalose and HES (Tre-HES), and β-CD. The infrared spectrum of the native glucagon in aqueous solution was also shown for comparison. (a) After freeze-drying; (b) after incubation at 60°C for 2 weeks.

Full size image
Fig. 7
figure 7

Intensity of α-helix peak of glucagon in formulations immediately after freeze-drying and after incubation at 60°C for 2 weeks. (a) After freeze-drying; (b) after incubation at 60°C for 2 weeks; Square: glycine hydrochloride buffer; Triangle up: sodium phosphate buffer; Triangle down: sodium citrate buffer.

Full size image

The presence of polysorbate 20, HES or β-CD alone in the formulation did not result in substantial difference in the peak intensity at 1657 cm-1 compared to that for the peptide freeze-dried in buffer alone (Figs. 6a and 7a). In contrast, freeze-drying in formulations containing trehalose resulted in a substantial increase in band intensity due to narrowing of the band. This result is presumably because of an increase in the structural compaction of the peptide and a more narrow distribution of conformational substates than that present in aqueous solution. The mixture of trehalose and HES also preserved the α-helix structure of glucagon, but was less effective compared with the formulation in which trehalose was the only excipient.

Trehalose effectively preserved the secondary structure of glucagon probably by hydrogen bonding to the peptide in place of water during drying, as has been observed with many proteins (6,7). HES or β-CD probably could not form effective hydrogen-bonds with dried glucagon and, therefore, did not have a protective effect on the secondary structure of glucagon.

The effects of incubation (60°C for 2 weeks) of the freeze-dried formulations prepared in glycine buffer on glucagon’s secondary structure were also evaluated (Figs. 6b and 7b). Interestingly, the α-helix peak intensity for most formulations was increased following incubation, suggesting that the secondary structure of glucagon was more ordered. The cause of this effect is not known.

The effects of excipients in sodium phosphate buffer were similar to those seen in glycine hydrochloride buffer (Fig. 7), with the greatest intensity of the 1657 cm−1 band noted in freeze-dried formulations containing trehalose. Also, the peak intensities in the glucagon spectra were increased by incubation of the freeze-dried formulations at 60°C for 2 weeks.

With sodium citrate buffer the effects of excipients also were similar to those seen in glycine buffer (Fig. 7a), with the greatest intensity of the 1657 cm−1 band noted in freeze-dried formulations containing trehalose (Fig. 7a). However, unlike the results noted in formulations in glycine and sodium phosphate buffers in which the presence of β-CD did not change the secondary structure compared to the formulations without any excipient, there was increased band intensity (at 1657 cm−1) in spectrum of glucagon in the β-CD formulation compared to that for the peptide in buffer alone (Fig. 7a). Similar to the results seen when glycine hydrochloride or sodium phosphate was used as buffers, after incubation at 60°C for 2 weeks, the band intensity in the spectra for glucgaon in all formulations was slightly increased (Fig. 7b).

Turbidity in reconstituted glucagon formulations was evaluated by measuring OD405 (Fig. 8). Formulations prepared in glycine or sodium phosphate buffer had very low turbidity after reconstitution (Fig. 8a, b), except for those containing β-CD. As discussed below, the turbidity of formulations with β-CD did not correlate to a loss of soluble glucagon as determined by SE-HPLC. Although β-CD has relatively low solubility in water (18.5 mg/ml at room temperature(39)), the concentration of this excipient after rehydration and subsequent dilution for turbidity analysis, 2 mg/ml, is well below the typical solubility. It is not known what alterations in β-CD may have occurred during freeze-drying or storage of freeze-dried formulations at 60°C, which could account for the putative reduced solubility of the excipient.

Fig. 8
figure 8

Turbidity (OD405) of glucagon formulations in different buffers following lyophilization and incubation at 60°C for 2 weeks. Dry formulations were reconstituted in corresponding buffers before analysis. A:glycine hydrochloride buffer; B: sodium phosphate buffer; and C: sodium citrate buffer. The values are the mean ± SD from three independent sample replicates.

Full size image

In formulations in sodium citrate buffer, substantial turbidity was observed after storage at 60 °C for samples without excipient or with polysorbate 20 (Fig. 8c). Also, the formulation with β-CD in sodium citrate buffer had much lower turbidity than those with this excipient prepared in glycine or sodium phosphate buffer. The other formulations studied had low turbidity when prepared in sodium citrate buffer.

To gain insight into peptide aggregation in formulations—immediately after freeze-drying and after 2 weeks of storage at 60°C—rehydrated samples were assessed by SE-HPLC. (Fig. 9). Soluble aggregate were at minimal levels (<0.5%) in all samples (data not shown), and the loss of monomeric glucagon was due to formation of insoluble aggregates (Fig. 9). The buffer used in the formulations had significant effects on monomer loss due to freeze-drying and to incubation at 60°C for 2 weeks, especially when no excipient was present. After incubation, the formulation with glycine buffer exhibited lowest degree of peptide aggregation (7%), followed by sodium phosphate buffer (11%), and then by sodium citrate buffer (22%). In formulations prepared in all three buffers, the presence of 0.01% polysorbate 20 or carbohydrate excipients reduced aggregation occuring during freeze-drying and incubation. Formulations containing carbohydrates in glycine buffer had the lowest level of insoluble aggregates. In contrast, for formulations containing sodium phosphate buffer the carbohydrates were minimally effective at inhibiting peptide aggregation during incubation at 60°C.

Fig. 9
figure 9

Loss of monomeric glucagon following freeze-drying and incubation at 60°C for 2 weeks as determined by SE-HPLC. The loss was determined as percent of monomeric glucagon relative to that present in solution prior to freeze-drying. A: glycine hydrochloride buffer; B: sodium phosphate buffer; and C: sodium citrate buffer. The values are the mean ± SD from three independent sample replicates.

Full size image

Chemical degradation of glucagon in the different formulations was determined by RP-HPLC (Fig. 10). In all of the formulations there was minimal (<1%) chemical degradation during freeze-drying and rehydration. However, after incubation at 60°C for 2 weeks, glucagon formulations without any excipient showed a significant amount (ca. 11%) of chemical degradation in all three buffers (Fig. 10). These degradation products were mostly due to hydrolysis as indicated by RP-HPLC (Fig. 3e). The samples have elution profiles similar to those for the degradation products obtained when the peptide was incubated in acidic solutions (Fig. 3d, e). The presence of 0.01% polysorbate 20 slightly decreased chemical degradation during incubation. Inclusion of carbohydrates, on the other hand, significantly decreased chemical degradation, except for the formulation in sodium citrate buffer containing β-CD. Formulations with carbohydrates in glycine hydrochloride buffer had lower chemical degradation (<4%) than those in the other two buffers.

Fig. 10
figure 10

Chemical degradation (%) of glucagon in different formulations following freeze-drying and incubation at 60°C for 2 weeks as determined by RP-HPLC. A: glycine hydrochloride buffer; B: sodium phosphate buffer; and C: sodium citrate buffer. The values are the mean ± SD from three independent sample replicates.

Full size image

Interestingly, in the freeze-dried glucagon formulations, oxidation was a very minor pathway of chemical degradation (<1%, with tR around 8–9 min, Fig. 3e), even after incubation at 60°C for 2 weeks. This is probably because the vials with the freeze-dried formulations were sealed under vacuum after freeze-drying cycle so that oxygen levels were reduced. In an earlier study (14), incubation of dried glucagon formulation at 45°C in open tubes exposed to 75% relative humidity resulted in a high degree of oxidation of the peptide. This degradation was noted in freeze-dried formulations prepared with either γ-CD or lactose and can be attributed both to the exposure to atmospheric oxygen and the uptake of water by the formulations. As in the case with formulations of protein therapeutics, with freeze-dried peptide drug formulations, it is important to reduce exposure to atmospheric oxygen and uptake of moisture into the dried formulation.

RP-HPLC was also used to monitor loss of soluble glucagon due to formation of insoluble aggregates (Fig. 11). By comparing the difference in the total area under peaks in the chromatogram of both unprocessed and centrifuged samples, we estimated the amount of insoluble aggregates formed during freeze-drying and incubation. The results observed were similar to those obtained with SE-HPLC (Fig. 9).

Fig. 11
figure 11

Loss of soluble glucagon following freeze-drying and incubation at 60°C for 2 weeks as determined by RP-HPLC. The loss was determined based on the total area for peaks of the chromatograms before treament and in those for the supernatant of samples centrifuged after treatment. A: glycine hydrochloride buffer; B: sodium phosphate buffer; and C: sodium citrate buffer. The values are the mean ± SD from three independent sample replicates.

Full size image

Combined Effects of Polysorbate 20 and Carbohydrates on Glucagon’s Stability in Formulations Prepared in Glycine Buffer

Based on the results reported above, we tested the hypothesis that the addition of polysorbate 20 to formulations with trehalose, HES or trehalose/HES in glycine hydrochloride buffer would provide greater stability than noted without the surfactant. β-CD was not included in this round of experiments since formulations containing this excipient had high turbidity after freeze-drying and rehydration (Fig. 8).

As expected, the Tg values for the freeze-dried formulations prepared with polysorbate 20 were similar to those for the respective formulations without surfactant (Tables III and IV). Also, infrared spectra for the peptide were similar in the freeze-dried formulations prepared without and with polysorbate 20, both immediately after drying and after 2 weeks incubation at 60°C (data not shown). Furthermore, turbidity was not detectable in any of the formulations after reconstitution, either immediately after freeze-drying or after incubation of the freeze-dried formulations (data not shown).

Table IV Composition and Tg (°C) of Optimized Freeze-Dried Glucagon Formulations
Full size table

These formulations also had low levels of aggregated glucagon, as determined by SE-HPLC (Fig. 12a). Less than 3% of monomer loss was observed after incubation at 60°C for 2 weeks. The addition of polysorbate 20 to each of the three carbohydrate formulations slightly improved the chemical stability of glucagon, as determined by RP-HPLC after reconstitution (Fig. 12b). Less than 2% degradation was observed for all three formulations, even after 2 weeks on incubation at 60°C.

Fig. 12
figure 12

Physical and chemical degradation of optimized glucagon formulations following freeze-drying and incubation at 60°C for 2 weeks. A: physical degradation as determined by SE-HPLC. B: chemical degradation as determined by RP-HPLC. The values are the mean ± SD from three independent sample replicates.

Full size image

CONCLUSIONS

As in the case with proteins, freeze-drying glucagon in carbohydrate formulations that formed glassy matrices was important for maintaining chemical and physical stability during processing and storage at elevated temperature. Also, the formulations with trehalose provided superior protection of glucagon secondary structure during freeze-drying than did those with HES alone or β-CD, which is similar to results with proteins. However, formulations with any of the three carbohydrates provided similar stabilization against aggregation and chemical degradation during incubation of the freeze-dried glucagon formulations. Therefore, even though maintaining secondary structure in the dried solid state is an important requirement for long-term storage stability of dried proteins, it appears to be less important for small peptides such as glucagon, which usually do not have well-organized higher-order structures.

Also, as in the case with protein formulations the presence of polysorbate 20 reduced glucagon aggregation during freeze-drying and storage. This effect was observed in formulations with buffer alone as well as with those containing carbohydrate excipients. Interestingly, chemical degradation during storage was also reduced when polysorbate 20 was included in the formulations (Fig. 10). The mechanisms for this effect is not known but may be due, at least in part, to the reduction peptide aggregation and a resulting decrease in the rate of chemical degradation.

We found that formulation lessons learned from studies of freeze-dried formulations of proteins can be applied successfully, to some degree, to development of stable formulations of glucagon. This conclusion may apply to other peptides as well. But as in the case with proteins, each peptide will have its unique physicochemical properties and impact of degradation on therapeutic product safety and efficacy. Thus, in each case, formulation development can follow general guidelines but will require empirical screening to obtain optimal peptide stability.