Effect of Particle Size and Viscosity of Suspensions on Topical Ocular Bioavailability of Budesonide, a Corticosteroid

Abstract

Purpose:

To determine the effect of particle size and viscosity of suspensions on topical ocular bioavailability of budesonide, a corticosteroid drug.

Methods:

Budesonide microparticle and nanoparticle (MP and NP) suspensions were prepared with or without homogenization and microfluidization. Using different grades of hydroxyl propyl methyl cellulose, low viscosity NP (NP-LV) and low and high viscosity MP (MP-LV and MP-HV) were prepared. Suspensions were characterized for particle size, viscosity, and osmolality. Budesonide suspensions were administered topically to rabbits and aqueous humor was collected and analyzed for budesonide. Budesonide C_max, t_max, and the area under the concentration time curve (AUC _(0–6h)) values were determined. The geometric mean ratio of AUC and bioequivalence was evaluated using a bootstrap method.

Results:

The particle sizes for NP and MP were ∼700 and 2,000 nm. The viscosities for low and HV formulations were ∼5 and 50 cP. The geometric mean budesonide C_max values for the suspensions NP-LV, MP-LV, and MP-HV were 0.22, 0.22, and 0.31 μg/g, t_max values were 0.67, 0.60 and 0.53 h, and AUC_0–6h values were 0.72, 0.53, and 0.95 μg h/g, respectively. Bootstrap analysis indicated that the 90% confidence intervals of the geometric mean ratio of AUC_0–6h values were 1.00–1.74 (MP-HV vs. NP-LV), 0.57–0.96 (MP-LV vs. NP-LV), and 0.45–0.70 (MP-LV vs. MP-HV).

Conclusions:

The 3 budesonide suspensions assessed in this study were not bioequivalent. Results suggested that an increase in viscosity improves the bioavailability of budesonide from the microsuspension formulation.

Introduction

According to Title 21 Code of Federal Regulation (CFR) §314.94 (a)(9)(iv),^1,2 a generic drug product intended for ophthalmic use must contain the same inactive ingredients in the same concentration as the reference listed drug (RLD). However, an applicant may seek approval of a generic drug product that differs from the RLD in preservative, buffer, substance to adjust tonicity, or thickening agent provided that the applicant identifies and characterizes the differences and provides information demonstrating that the differences do not affect the efficacy or safety of the proposed drug product. In some cases, bioequivalence (BE) may be considered self-evident. For instance, for ophthalmic solutions that are qualitatively (Q1) and quantitatively (Q2) the same as the RLD, BE is considered to be self-evident and a waiver for in vivo BE studies can be requested.³ Unlike solutions, suspensions require an additional step of “dissolution or release” before the drug is absorbed from the dosage form. Pharmaceutically equivalent ophthalmic suspensions can have varying physicochemical properties either because of the differences introduced during manufacturing or the formulation design, which may affect the precorneal residence time, drug release, and ocular bioavailability.

Literature suggests that differences in the physicochemical properties of ophthalmic suspensions can result in differences in ocular drug bioavailability.⁴ An investigation of the relationship between physicochemical properties and their effect on ocular bioavailability is crucial to formulate suspensions that are bioequivalent. However, in vivo BE studies for ophthalmic suspensions can be cumbersome, often requiring hundreds of patients in a clinical study. In this study, we hypothesized that key physicochemical properties such as particle size and viscosity can affect the bioavailability of budesonide, a model corticosteroid capable of exerting pharmacological activity in ocular cells,⁵ since these properties can influence drug release and precorneal residence time.^6–10 To address this hypothesis, we manufactured budesonide (corticosteroid) suspensions with similar compositions and different physicochemical properties (i.e., size and viscosity), and assessed the in vivo drug release in a rabbit model.

Methods

Materials

Micronized budesonide, E.P. (Cat. No. B1595) was purchased from Spectrum Chemicals Mfg. Corp. (New Brunswick, NJ). Polysorbate 80 (Tween 80-LQ-MH™) was obtained as a gift from Croda, Inc. (Edison, NJ). Different grades of the viscoelastic polymer Hypromellose, also known as hydroxyl propyl methylcellulose (HPMC grades K100 or K15M), were obtained as a gift from Colorcon (Harleysville, PA) to adjust viscosity. Sodium hydroxide (NaOH), ethylene diamine tetra acetic acid (EDTA), monosodium dihydrogen phosphate (NaH₂PO₄), disodium hydrogen phosphate (Na₂HPO₄), sodium chloride (NaCl), acetonitrile (ACN), and formic acid were all purchased from Sigma-Aldrich (St. Louis, MO).

Manufacture of the budesonide suspensions (0.1% w/w) with different particle sizes and viscosities

The purpose of this study was to determine the effect of particle size and viscosity on the bioavailability of budesonide suspensions. The following budesonide (0.1% w/w) suspensions differing in particle size and/or viscosity were prepared: (1) nanoparticle-low viscosity (NP-LV)—nanosuspension with a smaller particle size and a lower viscosity target of 5 cP, (2) microparticle-low viscosity (MP-LV)—microsuspension with a larger particle size and a lower viscosity target of 5 cP, and (3) microparticle-high viscosity (MP-HV)—microsuspension with a larger particle size and a higher viscosity target of 50 cP. The typical components and composition of 0.1% w/w budesonide suspensions used are shown in Table 1. All the formulation components were the same, except for the incorporation of either K100 (LV) or K15M (HV) HPMC polymer to obtain the viscosity differences.

Table 1.

Typical Composition of a 0.1% w/w Budesonide Suspension

Ingredient	Amounts for 100 mL (mg)
Budesonide	100
Polysorbate 80	100
EDTA	20
NaCl	400
Na2HPO4	500
NaH2PO4	250
HPMC (K100 or K15M)	500
NaOH	q.s.a to adjust to pH 6.2
Water	q.s.a

^a q.s. or quantum sufficit is as much as is sufficient.

EDTA, ethylene diamine tetra acetic acid; HPMC, hydroxyl propyl methylcellulose; NaCl, sodium chloride; Na₂HPO₄, disodium hydrogen phosphate; NaH₂PO₄, monosodium dihydrogen phosphate; NaOH, sodium hydroxide.

A suspension master batch without HPMC was first prepared and then further processed to obtain MP-LV, MP-HV, or NP-LV. The typical procedure for preparing various budesonide suspension formulations was as follows; for the master batch, 100 mg of polysorbate 80 was weighed into a 100 mL beaker and 50 mL of double-distilled water was added to prepare a solution. All the salts (NaH₂PO₄, Na₂HPO₄, NaCl, and EDTA at weights listed in Table 1) were weighed, added to the above solution, and dissolved. To this solution, 100 mg of micronized budesonide was added and stirred at room temperature for 20 min at 300 rpm using an Isotemp^® magnetic stir plate (Fisher Scientific, Pittsburgh, PA). To prepare MP-LV, one 15 mL aliquot of master batch was added and mixed with an equal volume of 1% w/v K100 grade HPMC in double-distilled water. To prepare MP-HV, to 15 mL aliquot of master batch, an equal volume of 1% w/v K15M grade HPMC of equal volume was added and mixed. To prepare NP-LV, a 10 mL aliquot of the master batch was first homogenized at around 25,000 rpm for 10 min using the Omni International Homogenizer (Kennesaw, GA). The contents were then subjected to the second particle size reduction step using the Microfluidics LV1 high pressure fluidizer (Westwood, MA). The microfluidizer was operated at about 28,000 psi and the contents were cycled through the unit 3 times. The resulting suspension was mixed with an equal volume of 1% w/v K100 grade HPMC. All final suspensions were adjusted to pH 6.2 using NaOH.

Measurement of the physicochemical properties, particle size, viscosity, and osmolarity

Dynamic light scattering is typically used for the characterization of particles dispersed in a liquid. In this study, particle size characterization for the budesonide suspensions was performed as follows: one milliliter of suspension was carefully transferred to a disposable cuvette in such a manner as to prevent air entrapment (i.e., bubbles) and particle size was then measured using a Malvern Zetasizer Nano-ZS (Westborough, MA) with the following settings: (1) material was selected as “silica,” (2) dispersant was selected as “water,” and (3) temperature was set at 25°C for the test.

Viscosity of the suspensions was measured using a Brookfield Viscometer Model DV-I Prime (Middleborough, MA) with a cone and plate. Five hundred microliters of each suspension was used for the measurement and the viscosity was obtained at 25°C.

Osmolarity was determined using an Advanced Instruments Osmometer Model 3250, (Norwood, MA).

Effect of physicochemical properties on in vivo delivery in the rabbit model

The animal study was conducted in accordance with the Guidelines of the Association for Research in Vision and Ophthalmology (ARVO) Statement on the Use of Animals in Ophthalmic and Vision Research and the guidelines of the Animal Care Committee of the University of Colorado Anschutz Medical Campus. For the suspension studies, New Zealand White (NZW) male rabbits in a weight range of 4 to 6 pounds were obtained from Harlan Laboratories (Frederick, MD). The rabbits were placed into the restrainer and allowed to acclimatize for about 5 min. Then a 30 μL topical eye drop of budesonide suspension (different in particle size and viscosity) was placed in the lower cul-de-sac of both eyes (n = 4 rabbits or n = 8 eyes per time point) using a variable Gibson 10–100 μL positive displacement pipette (Middleton, WI). To minimize the administered dose runoff, the eyelids were gently closed just after application for 3 s. Five minutes before the specified euthanasia time points, 0.117, 0.5, 1, 2, 6, and 24 h, each rabbit was transferred from the cage and placed in an anesthesia chamber supplied with isoflurane and oxygen. Once the rabbit was anesthetized, it was euthanized using 1 mL of sodium pentobarbitone (150 mg/kg) through the marginal ear vein. Immediately, eyes were proptosed and clamped. Each eye was washed with phosphate-buffered saline (pH 7.4) to remove any residual suspension from the eye surface. Aqueous humor (ranging from 100 to 150 μL) was aspirated using a tuberculin syringe (the needle angle was ∼180° to the eye) and transferred to a tarred sample vial. Immediately, the eye was snap frozen using a dry ice and isopentane bath and stored at −80°C until the analysis was performed.

During this study, topical administration of the budesonide suspensions and euthanasia were performed with and without the use of a randomization protocol. In the absence of randomization protocol (for time points 0.117, 1, and 24 h), all rabbits needed for each time point (3 suspensions × 4 animals each = 12 rabbits) were used at the same time, and for each rabbit, a topical dose was administered first to the left eye and then to the right. To eliminate any unintentional bias, a randomization protocol was used for the following time points (0.5, 2, and 6 h). The randomization studies were performed on 2 consecutive days.

Tissue sample processing for LC-MS/MS analysis

The aqueous humor was transferred to tubes and mixed with an internal standard solution. Samples were vortexed for 15 min on a multiple vortex mixer (VWR LabShop, Batavia, IL). About 1.25 mL of ethyl acetate was added to the tubes and they were again vortexed this time for 30 min. The supernatant was pipetted out and transferred to clean glass tubes and evaporated under a nitrogen stream (Multivap; Organomotion, Berlin, MA) at 40°C. The residue was then reconstituted with ACN: water (25:75 v/v) and subjected to LC-MS/MS analysis.

LC-MS/MS analysis of budesonide using the QTrap^® 4500

Budesonide samples generated from the in vivo study were analyzed using a Sciex QTrap 4500 mass spectrometer (AB Sciex, Farmingham, MA) with the AB Sciex Analyst software (version 1.6.1), coupled with a Shimadzu HPLC (Columbia, MD). Chromatographic separation of budesonide (analyte, time_R = 1.29 min) and triamcinolone acetonide (internal standard, time_R = 0.89 min) was achieved using an Agilent Zorbax Extend-C18 [4.6 × 50 mm, 5 μm] column (Agilent Technologies, Santa Clara, CA) and a linear gradient of 5 mM ammonium formate, pH 3.5 (A), and 0.1% formic acid in ACN (B). The column temperature was set to 40°C, the flow rate was set to 1 mL/min, and the gradient ran from 55% to 95% B with a total run time of 4.0 min. The mass spectrometer was operated in the ESI⁺ mode, and typical mass spectrometer settings were as follows: curtain gas, 30 psi; collision gas, medium; ion spray voltage, 5,500 V; source temperature, 650°C; ion source gas 1, 50 psi; and ion source gas 2, 60 psi. Quantitation was performed in multiple reaction monitoring mode (MRM) using the m/z 435.1 → m/z 213.2 transition for triamcinolone acetonide (collision energy, 37 V; declustering potential, 78 V), and the m/z 431.0 → m/z 147.1 transition for budesonide (collision energy, 34 V and declustering potential, 58 V).

Statistical significance and P value

The significance of aqueous humor budesonide concentration differences across the 3 suspensions was determined using analysis of variance, specifically, aov, HSD.test, and TukeyHSD [agricolae] using the RStudio program (v1.1.463) and R (v3.5.2).

Bootstrapping using R Program

A bootstrap approach, programmed in R, used aqueous humor concentrations at each time point, randomly chosen, with replacement to generate multiple concentration versus time data sets. This process was repeated 10,000 times. C_max (peak concentration) and t_max (time-to-peak concentration) values were calculated from each data set by inspection and the area under the aqueous humor concentration versus time curve (AUC) values was calculated by the linear trapezoidal rule. AUC ratios were calculated from the AUC values.

Results

Budesonide suspensions with different particle sizes and viscosities

Three 0.1% w/v budesonide suspension formulations with different particle sizes and viscosities were prepared: (1) NP-LV; (2) MP-LV; and (3) MP-HV), and used for the in vivo NZW rabbit study. Particle size, viscosity, and osmolality of each formulation are shown in Table 2.

Table 2.

Representative Particle Size, Viscosity, and Osmolarity of Budesonide Suspensions

Suspension	Size (nm)	PDI	Viscosity (cP)	Osmolarity (mOsm/L)
NP-LV	707	0.24	4.89	284
MP-LV	1954	0.21	4.92	285
MP-HV	1980	0.12	53.2	279

NP-LV, nanoparticle-low viscosity; MP-LV, microparticle-low viscosity; MP-HV, microparticle-high viscosity.

Effect of physicochemical properties on in vivo delivery in the rabbit model

Drug delivery assessment in rabbit aqueous humor indicated detectable budesonide concentrations in all samples except some of the samples collected at the 24-h time point. At 24 h, for the NP-LV, budesonide levels were not detectable in 6 out of 8 eyes; for MP-LV, low levels were detected; in the case of MP-HV, budesonide was at the highest levels for this study but not detectable in 1 out of 8 eyes. Two data points were missing due to procedural errors. A third data point was removed as an outlier (10 standard deviations from the mean). The concentration versus time data are presented in Fig. 1.

FIG. 1.

Concentration-time profile of budesonide in aqueous humor after administration of budesonide suspension (0.1% w/v) with variable size and viscosity. Data presented as mean ± SD n = 7 or 8 eyes. NP-LV, nanoparticle-low viscosity; MP-LV, microparticle-low viscosity; MP-HV, microparticle-high viscosity; SD, standard deviation.

Tukey HSD test indicated that budesonide concentrations were statistically different at 0.117 h (MP-HV was higher than MP-LV, but NP-LV was not different), 1 h (again MP-HV was higher than MP-LV, but NP-LV was not different), and 24 h (NP-LV lower than both MP-LV and MP-HV), but this was not the case for any of the other time points collected (Table 3).

Table 3.

Statistical Analysis of Budesonide Concentrations by Sample Time

Time (hr)	P	NP-LV	MP-LV	MP-HV
		Mean and MR designation
0.117	0.012	0.125 (ab)	0.064 (b)	0.191 (a)
0.5	0.151	0.214 (a)	0.214 (a)	0.310 (a)
1	0.036	0.205 (ab)	0.173 (b)	0.262 (a)
2	0.112	0.147 (a)	0.078 (a)	0.174 (a)
6	0.402	0.038 (a)	0.049 (a)	0.068 (a)
24	0.001	0.0005 (b)	0.006 (a)	0.0091 (a)

P values (<0.05) were significant for time points 0.117, 1, and 24 h. Formulations with the same Multiple Range (MR) designation are not significantly different. That is, all “a” formulations are not different from each other, all “b” formulations are not different from each other, and “ab” indicates not different from “a” or “b.” The units for aqueous humor concentration are μg/g.

Bootstrap analysis

Bootstrap analyses in Table 4 indicated that the 90% confidence intervals of the ratio of arithmetic mean AUC _(0–6h) values were 0.97–1.71 (MP-HV vs. NP-LV), 0.55–0.95 (MP-LV vs. NP-LV), and 0.44–0.70 (MP-LV vs. MP-HV), respectively. The 90% confidence intervals of the geometric mean ratio of AUC_0–6h values were 1.00–1.74 (MP-HV vs. NP-LV), 0.57–0.96 (MP-LV vs. NP-LV), and 0.45–0.70 (MP-LV vs. MP-HV) (Table 5). Thus, the 3 formulations were not bioequivalent.

Table 4.

Bootstrap Analysis of Rabbit Aqueous Humor Time-Course of Drug Concentrations and Determination of AUC Ratio, Arithmetic Mean

Suspension	Mean	SD	90% confidence interval
NP-LV
Cmax	0.222	0.019	0.191–0.253
tmax	0.720	0.341	0.160–1.280
AUC (0–6h)	0.723	0.094	0.569–0.878
AUC (0–24h)	1.075	0.146	0.0835–1.315
MP-LV
Cmax	0.223	0.046	0.146–0.299
tmax	0.625	0.216	0.269–0.981
AUC (0–6h)	0.535	0.048	0.456–0.613
AUC (0–24h)	1.029	0.167	0.754–1.303
MP-HV
Cmax	0.310	0.023	0.2733–0.348
tmax	0.532	0.126	0.325–0.739
AUC (0–6h)	0.952	0.097	0.792–1.112
AUC (0–24h)	1.642	0.209	1.299–1.986
AUC (0–6h) ratio
MP-HV/NP-LV	1.34	0.22	0.97–1.71
MP-LV/NP-LV	0.75	0.12	0.55–0.95
MP-LV/MP-HV	0.57	0.078	0.44–0.70

The units for C_max, t_max, and AUC are μg/g, hr, and μg·hr/g, respectively.

AUC, area under the concentration time curve; SD, standard deviation.

Table 5.

Bootstrap Analysis of Rabbit Aqueous Humor Time-Course of Drug Concentrations and Determination of AUC Ratio, Geometric Mean

Suspension	Mean	SD factor	90% confidence interval
NP-LV
Cmax	0.221	1.09	0.193–0.253
tmax	0.667	1.50	0.344–1.300
AUC (0–6h)	0.717	1.14	0.578–0.889
AUC (0–24h)	1.064	1.15	0.849–1.333
MP-LV
Cmax	0.218	1.22	0.157–0.304
tmax	0.596	1.35	0.363–0.977
AUC (0–6h)	0.532	1.09	0.459–0.616
AUC (0–24h)	1.013	1.17	0.779–1.318
MP-HV
Cmax	0.310	1.08	0.274–0.350
tmax	0.532	1.19	0.393–1.121
AUC (0–6h)	0.947	1.11	0.800–1.121
AUC (0–24h)	1.630	1.14	1.323–2.010
AUC (0–6 h) ratio
MP-HV/NP-LV	1.32	1.18	1.00–1.74
MP-LV/NP-LV	0.74	1.17	0.57–0.96
MP-LV/MP-HV	0.56	1.15	0.45–0.70

The units for C_max, t_max, and AUC are μg/g, hr, and μg·hr/g, respectively. SD factor is the geometric standard deviation factor.

Discussion

Budesonide is a corticosteroid approved for the treatment of Crohn's disease, irritable bowel syndrome, ulcerative colitis, chronic obstructive pulmonary disease, asthma, allergic rhinitis, and nasal polyps. Similar to other corticosteroids such as dexamethasone, budesonide is likely of therapeutic value in treating inflammatory conditions of the eye surface or intraocular tissues. Since budesonide can reduce vascular endothelial growth factor secretion from retinal pigment epithelial cells,⁵ it is also of potential value in treating eye diseases such as diabetic macular edema and wet age-related macular degeneration. In this study, we prepared budesonide microsuspensions and nanosuspensions and showed that 3 similar suspensions are not bioequivalent based on rabbit topical eye drop dosing and aqueous humor analysis. The suspension viscosity and particle size are critical factors in that order, contributing to this result.

Nanosuspensions as well as microsuspensions are of potential value for dosing corticosteroids as topical eye drops. The low solubility of corticosteroids makes them suitable for suspension dosage forms. Water solubility of budesonide at room temperature is about 20 μg/mL.¹¹ Based on surface area, nanosuspensions are expected to dissolve more rapidly compared to microsuspensions. However, microsuspensions may be retained for longer periods on the eye surface. Inclusion of a viscosity enhancing agent can further influence drug delivery since such an agent, while prolonging drug retention,¹² on the surface may also reduce diffusion¹³ rates for the dissolved solutes. Thus, particle size and formulation viscosity can influence drug dissolution, diffusion, and retention, thereby influencing topical ocular drug delivery in a complex manner. In this study, while keeping the ophthalmic suspensions of budesonide similar by including same inactive ingredients at the same composition, we varied the grades of a viscosity enhancer (HPMC, LV K100, vs. HV K15M) and the size of budesonide particles in the suspension using different manufacturing methods (without and with homogenization and microfluidization). Viscosities were close to 5 (LV) or 50 cP (HV) and the particle sizes were close to 700 nm (NP) and 2 μm (MP). These formulations were used to assess the influence of budesonide suspension physicochemical properties on their BE in a rabbit model.

Although, human pharmacokinetic data are ideal for determining BE of eye drops, such studies require a large number of patients since normal human eyes cannot be serially sampled for ocular fluids, without compromising the function of the eye. While noninvasive, nondestructive methods would be ideal, such methods are not available or routine in human eyes for ophthalmic drugs. For the purpose of this study, similar to most other ocular development and pharmacokinetic efforts, we used a rabbit model. Specifically, NZW rabbits were used in this study. This strain is most commonly used in preclinical development of ocular drug products. We used a drop volume of 30 μL,^14,15 which is comparable to a typical eye drop volume for marketed products.¹⁶ Aqueous humor was sampled in this study because this sample is easy to obtain in a reproducible manner and it reflects drug entry beyond ocular surface barriers. Also, the drug present in the aqueous humor is readily accessible to the tissues of the anterior segment of the eye. Drug concentrations in aqueous humor are most relevant for inflammatory conditions of the anterior segment of the eye.

Rabbit aqueous humor time-course studies indicated significant differences between the various budesonide formulations. Specifically, the aqueous humor concentrations tended to be the highest for the most viscous formulation MP-HV relative to a comparable formulation MP-LV, with a lower viscosity as well as the smaller particle size and LV NP-LV formulation (Fig. 1). The most viscous microparticle (MP) formulation exhibited significantly higher aqueous humor concentrations at 0.117 and 1 h relative to the MP preparation with LV and at 24 h relative to the NP preparation with LV (Table 3). The C_max and AUC _(0–6h) were the highest for the more viscous formulation (Tables 4 and 5). When BE was assessed using geometric mean AUC ratios of budesonide in rabbit aqueous humor, the 3 formulations tested in this study were not bioequivalent, since the 90% confidence interval fell outside the required 80% to 125% range.¹⁷ In summary, in vivo studies in rabbits indicated that an increase in viscosity by incorporating a different grade of HPMC from K100 to K15M can increase drug delivery from an MP budesonide suspension. However, it should be noted that the eye drop solution drainage is faster in humans than rabbits, with significant retardation of drainage in humans at 0.9% HPMC.¹⁸ Since human studies are cumbersome, this study is a first step in understanding the factors that influence ophthalmic budesonide suspension BE.

Conclusions

The 3 topical ophthalmic budesonide suspensions assessed in this study were not bioequivalent. While there was an enhanced drug delivery trend for the most viscous formulation MP-HV, there was no delivery advantage for the NP suspension when compared to the MP suspension. Thus, it appears that for the suspension formulations assessed in this study, the principal contributor to delivery differences appears to be viscosity and potential in vivo ocular surface drug retention differences as opposed to particle size differences or any dissolution difference. Future studies should focus on developing approaches to distinguish in vivo drug particle retention on the eye surface based on various physicochemical properties of ophthalmic formulations.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported by the US FDA grant U01FD004719.