Shear-thinning viscous materials for subconjunctival injection of microparticles

Abstract

While drug-loaded microparticles (MPs) can serve as drug reservoirs for sustained drug release and therapeutic effects, needle clogging by MPs poses a challenge for ocular drug delivery via injection. Two polymers commonly used in ophthalmic procedures - hyaluronic acid (HA) and methylcellulose (MC) - have been tested for their applicability for ocular injections. HA and MC were physically blended with sunitinib malate (SUN)-loaded PLGA MPs for subconjunctival (SCT) injection into rat eyes. The HA and MC viscous solutions facilitated injection through fine-gauged needles due to their shear-thinning properties as shown by rheological characterizations. The diffusion barrier presented by HA and MC reduced burst drug release and extended overall release from MPs. The significant level of MP retention in the conjunctiva tissue post operation confirmed the minimal leakage of MPs following injection. The safety of HA and MC for ocular applications was demonstrated histologically.

INTRODUCTION

Subconjunctival (SCT) injection has been used to achieve localized, site-specific delivery of ophthalmic drugs to treat a variety of eye diseases (1–6). Compared to systemic administration where drugs are subject to poor ocular absorption and potential systemic side effects, SCT injection holds the potential to enhance drug availability in the eye while reducing systemic side effects (7, 8). Local drug delivery is most commonly achieved by the use of topical eye drops, but the rate of drug removal of medicines in eye drops is typically rapid due to blinking, tear film turnover, fast dilution by the lacrimal fluid, and drainage into the nasolacrimal duct (9, 10). Also, the requirement for frequent dosing of eye drops with many products results in poor patient compliance (11). For drugs with poor barrier-penetrating abilities, SCT injection may enhance drug penetration into not only the anterior segment of the eye but also the back of the eye, where eye drops normally fail to reach (12). On the other hand, drugs in immediate-release formulations (i.e., “free drug” formulations without controlled release mechanism) administered through SCT injection can also be rapidly cleared, especially water-soluble drugs (13, 14). Therefore, drug delivery systems have been applied for SCT injection to provide sustained drug release and further reduce administration frequency (15, 16).

Drug-loaded poly(lactic-co-glycolic acid) (PLGA) microparticles (MPs) are widely employed in drug delivery due to their biodegradability and biocompatibility (17). They have been extensively explored as polymeric carriers for a variety of medications (18–20). In comparison to nanoparticles (NPs), MPs could provide higher drug loading and more sustained drug release (21). Combined with SCT injection, drug-loaded MPs have been shown to effectively combat diseases in the front of the eye in rats (22, 23). Fine-gauge needles, commonly 26–31G, help to reduce potential SCT injection-related risks, such as subconjunctival hemorrhage and elevated intraocular pressure (24). However, an issue arises when highly concentrated suspensions of MPs cannot be injected subconjunctivally due to particle sedimentation and clogging within fine needles. A second challenge is that a significant amount of injected particles can leak out through the hole created by a needle. This leakage not only limits the bioavailability of ocular drugs at the injection site, but also renders difficult the precise control of dosage by ophthalmologists. Therefore, it is necessary to develop a strategy for facile SCT injection of MPs at high concentrations with minimal leakage.

In this paper, we describe the application of shear-thinning materials to the subconjunctival injection of highly concentrated MPs. The shear-thinning vehicles have a relatively high viscosity at low shear to reduce liquid backflow after injection. Previously, hyaluronic acid (HA) and methylcellulose (MC) were employed for a stem cell transplantation procedure to treat retinal degenerative diseases (25, 26). HA and MC-based products for ocular applications are also commercially available, such as Healon^® (Abbott Medical Optics Inc., Abbott Park, IL) and Murocel^® (Bausch & Lomb Inc., Rochester, NY) (27, 28). We tested whether viscous solutions of HA and MC could facilitate the SCT injection of drug-loaded MPs. Due to its promising applications in ocular disorders (29, 30), sunitinib malate (SUN) was chosen as a model drug for the PLGA MPs. The shear-thinning characteristics of HA and MC viscous solutions were investigated via rheology measurements. In vivo ocular retention and histological examinations were conducted to confirm the feasibility and safety of HA and MC as vehicles for SCT injection of MPs. The effects of HA and MC solutions on the drug release kinetics of SUN were demonstrated in vitro.

MATERIALS AND METHODS

Materials

Poly(lactic-co-glycolic acid, 50:50, Mw ~5.1 kDa, acid terminated) (PLGA) was purchased from Evonik ((Birmingham, AL), and sunitinib malate (SUN) from LC laboratories (Woburn, MA). Hyaluronic acid (MW 1.01–1.8 MDa) and methylcellulose (intrinsic viscosity 4000 mPa·s) were supplied by Lifecore Biomedical (Chaska, MN) and SE Tylose USA, Inc. (Totowa, NJ), respectively. Poly(vinyl alcohol; MW ~25 kDa, 88 mol% hydrolyzed) (PVA) and Fluoresbrite^™ carboxylate YG 10 micron polystyrene microspheres (PS) (Cat. no. 17136) were purchased from Polysciences, Inc. (Warrington, PA). Dimethyl sulfoxide (DMSO), dichloromethane (DCM), and phosphate buffer solution (PBS, 1X, pH 7.4) were all purchased from Sigma-Aldrich (St. Louis, MO).

Preparation of SUN-loaded Microparticles

PLGA MPs were prepared following a modified single emulsion solvent evaporation protocol (31). Briefly, 50 mg of SUN and 250 mg of PLGA were fully dissolved in 0.625 mL of DMSO and 2.5 mL of DCM, respectively. The mixture of the two solutions was emulsified (L4RT High Shear Mixer, Silverson, East Longmeadow, MA) at 4000 rpm for 2 min in 60 mL of 1% w/v PVA aqueous solution. The emulsion was placed on a magnetic stirrer at 700 rpm for 2 h, and was then placed in a vacuum chamber for 2 h under stirring to remove residual DCM. The suspension was filtered through a 40 μm filter to remove aggregates, after which SUN-encapsulated MPs were collected by centrifugation at 500 ×g for 10 min and washed with deionized water. The MPs were lyophilized for subsequent procedures and storage.

Microparticle Size and Morphology

Diameters of SUN-loaded PLGA MPs and nondegradable 10 μm polystyrene (PS) MPs (Polysciences, Cat. no. 17136) were characterized using a Multisizer 4e Coulter Counter (Beckman Coulter Inc., Brea, CA). Scanning electron microscopy (SEM) images of SUN-loaded PLGA MPs were obtained from LEO/Zeiss Field-Emission scanning electron microscope (Carl Zeiss AG, Oberkochen, Germany). The representative SEM image of the 10 μm polystyrene (PS) MPs were available online www.polysciences.com/default/polybead-microspheres-1000181m.

Preparation of HA and MC Solutions and Blends

HA and MC viscous solutions of concentration 2 % w/v were prepared in PBS. HA and MC powders were weighed and manually stirred in PBS for preliminary mixing. The mixtures were centrifuged at 2000 g for 5 min to remove entrapped air bubbles, after which the solutions were placed in a 4 °C cold room overnight to fully dissolve remaining HA and MC to obtain transparent viscous solutions. For the blends of MPs with HA and MC, lyophilized SUN-loaded MPs or SUN powders were carefully weighed and manually blended into the viscous solutions (2% w/v) according to the following compositions: 10% w/v MPs in HA, 20% w/v MPs in HA, 20% w/v free SUN in HA, 10% w/v MPs in MC, 20% w/v MPs in MC, 20% w/v free SUN in MC.

Rheological Characterization of HA and MC Solutions and Blends

Rheology studies of 2% w/v HA and MC were conducted on an ARES RFS III rheometer (Rheometric Scientific, Piscataway, NJ). The rheometer has a parallel cone-plate geometry with a cone diameter of 25 mm. Samples of 200 μL in volume were carefully pipetted onto the plate to avoid bubble formation. The gap between the cone and the plate was adjusted to 0.3 mm and the temperature to 37 °C for all measurements. Throughout experiments, the cone and the plate were covered by a humidity chamber to minimize water evaporation from the viscous HA and MC solutions. The elastic modulus (G’) and the viscous modulus (G”) were measured by dynamic frequency sweep (0.1–10 Hz). The viscosity (η) was profiled as a function of shear stress (0–100 Pa). HA and MC blends with 10% w/v and 20% w/v MPs were also rheologically characterized to assess the effect of MPs on the viscosity of pure HA and MC solutions.

Animals

All animal procedures followed the regulations of the Johns Hopkins Animal Care and Use Committee and the Association for Research in Vision and Ophthalmology (ARVO). Sprague Dawley rats (200–250 g, 6–8 weeks old, male) were purchased from Harlan (Indianapolis, IN). Anesthesia was performed on animals by intramuscular injection of a mixture of Ketamine (50 mg/kg) and Xylazine (5 mg/kg) prior to experimental procedures.

Subconjunctival Retention of PS Microparticles

Nondegradable fluorescent carboxylated polystyrene (PS) MPs with a hydrodynamic diameter of 10 μm were washed with DI water, and manually blended with HA 2% w/v and MC 2% w/v viscous solutions, or pure PBS. The final concentration of PS MPs in the solution was controlled at 1.3% w/v. After SCT injection of 30 μL blend through 26G needles, rats were sacrificed at various time points and their whole eyes were enucleated. The collected eyes were placed with the injection site up on an 18-well plate. Xenogen IVIS Spectrum optical imaging system (Caliper Life Sciences Inc., Hopkinton, MA) was used to image the whole eyes at an excitation wavelength of 500 nm and an emission wavelength of 430 nm. The total fluorescent counts in the conjunctival tissue were quantified using the Living Image 3.0 software (Caliper Lifesciences, Inc.). For subconjunctival retention data analysis, the baseline (100% retention) was assumed to be the total fluorescent counts immediately following SC injection.

Histological Examination

To investigate the safety of HA and MC for ocular administration, we injected 30 μL of pure HA 2% w/v, MC 2% w/v solutions or PBS into the conjunctiva tissue with 26G needles, and collected rat eyes at postoperative days (POD) 2, 7, 14, and 30. No MPs were used, in order not to complicate the histological toxicity effects of HA and MC viscous materials. The eyes with conjunctiva tissue were fixed with 10% formalin for 24 h before being embedded in paraffin. Axial sections of 5 μm thickness were sliced from the cornea to the optic nerve and stained with hematoxylin and eosin (H & E). To ensure that the representative histological sections from the injection area, we marked the injection sites with 6–0 suture and the cross-sections were through the marked positions. A pathologist qualitatively assessed the degree of inflammation.

Microparticle Drug Loading and Drug Release

The SUN-loaded PLGA MPs prepared using the solvent evaporation procedures described above were spherical in shape as observed under SEM (Fig. 1) and had diameters of around 13.8 μm (Table 1). The dimensions of the SUN-loaded PLGA MPs were comparable to those of the nondegradable PS MPs used in subconjunctival retention experiments (Table 1).

Fig. 1.

SEM image of SUN-loaded PLGA MPs.

Table I.

Physiochemical characterizations of SUN-loaded PLGA and PS MPs

MPs	Diameter (μm)	Drug loading (%)
SUN-loaded PLGA	13.8 ± 5.8	7.6
PS	10.1 ± 1.0	N/A

The drug loading of MPs was calculated using the mass of SUN in a known mass of MPs according to the formula below. 50 μL of MPs were lyophilized, carefully weighed, and dissolved in 1 mL of DMSO. The SUN concentration in the solution was measured by a SynergyMX BioTek UV-Vis spectrophotometer (BioTek Instruments, Inc., Winooski, VT) at a wavelength of 441 nm. The mass of SUN was then calculated based on a standard curve.

$$\text{drug\hspace{0.17em}loading\hspace{0.17em}}\left(\text{\%}\right)=\frac{\text{mass\hspace{0.17em}of\hspace{0.17em}drug}}{\text{mass\hspace{0.17em}of\hspace{0.17em}MPs}}\times 100$$ (1)

In vitro drug release was set up by adding 0.1 mL of 2% w/v HA or MC blends (weight of the HA over the volume of PBS) containing either 20% w/v SUN-loaded MPs (weight of the cargo over volume of viscous solution/blend) or 20% w/v SUN free drug to the bottom of a 1.5 mL Eppendorf tube. To measure SUN release from MPs without HA or MC, equivalent amounts of SUN-loaded MPs were suspended in 0.1 mL PBS and added to the tube. Afterward, the tube was filled with 1.1 mL PBS, without disturbing the viscous blends or the MPs. The drug release samples were incubated at 37 °C in a water bath. At predetermined time points, the supernatant was carefully collected without agitating the HA and MC blends or the MPs, and replaced with 1.0 mL of fresh PBS for further drug release.

Statistical Analysis

Drug release experiments were conducted in triplicate. In vivo subconjunctival retention experiments were done using three rat eyes for each experimental group. All values and error bars are reported as mean plus minus standard deviations.

RESULTS

Rheological Characterization of HA and MC Viscous Materials

Six samples, 2% w/v HA, 2% w/v HA with 10% w/v SUN-loaded MPs, 2% w/v HA with 20% w/v SUN-loaded MPs, 2% w/v MC, 2% w/v MC with 10% w/v SUN-loaded MPs, 2% w/v MC with 20% w/v SUN-loaded MPs were investigated for their rheological moduli and shear thinning profiles. As shown in Fig. 2, the elastic modulus is higher than the viscous modulus for both HA and MC samples during low-range dynamic frequency sweep measurements, indicating that they are viscoelastic solids (gel-like) before injection. However, the viscosities of HA and MC solutions decreased significantly with increasing shear stress (Fig. 3), provided that the given shear stress was greater than the yield stress, which is lower than 20 Pa, of HA and MC viscous solutions. The presence of MPs in concentrated quantities led to enhanced viscosities compared to HA or MC in solution alone. However, the shear-thinning properties of HA and MC were preserved in the presence of MPs, as is evident in the observation that viscosities still dropped significantly as higher shear stress was applied. We further confirmed the injectability of 2% w/v HA and MC solutions containing 20% w/v SUN-loaded MPs through 26G needles, as presented in Fig. S1.

Fig. 2.

Elastic (G’) and viscous (G”) moduli of (A) HA and (B) MC samples as a function of angular frequency. The vertical axis is shown in logarithmic scale to demonstrate the marked differences between the two moduli.

Fig. 3.

Shear-thinning profiles of (A) 2% w/v HA aqueous solution with or without MP suspended at various concentrations and (B) equivalent samples using MC as the carrier. The initial shear thickening behaviors reflect the yield stress of the solutions.

Subconjunctival Retention of PS Microparticles

Nondegradable fluorescent PS MPs of diameter 10 μm were used to investigate the effect of HA and MC on subconjunctival particle retention after SCT injection. The non-degradability of PS MPs served to eliminate fluorescent signal drop due to PLGA MPs degradation and/or release of SUN. The size of the PS particles was chosen to be similar to the SUN MPs used in the drug release study. Fig. 4 suggests that after a rapid initial fluorescent signal drop, likely due to MP leakage through the needle puncture created by SC injection, both HA and MC kept SC injected MPs within the conjunctiva tissue at a relatively constant level of approximately 80% fluorescent signal remaining for the 30 day duration of the study. While MPs suspended in PBS would serve as a control group for HA and MC, no meaningful retention data for MPs in PBS injection was obtained due to unsuccessful SCT injection caused by PS MP sedimentation and needle clogging issues. It should be noted that, due to PLGA degradation, the subconjunctival retention time of PLGA MPs blended with HA or MC may be shorter than the observed periods for PS MPs.

Fig. 4.

(A) Fluorescence imaging of rat eyes following SCT injection of non-degradable fluorescent PS MPs suspended in 2% w/v PBS solutions of either HA or MC. (B) Subconjunctival retention of PS MPs after SCT injection (N=3).

Ocular safety of HA and MC Viscous Solutions

To examine the in vivo toxicity of HA and MC, HA, MC, and PBS with no MPs suspended were injected with 26G needles into rat conjunctiva tissue in 30 μL aliquots, as shown in Fig. S2. Animals receiving the PBS control injection showed no inflammation in the corneal or conjunctival tissue at indicated time points (2, 7 and 30 days). HA and MC injection caused a mild increase in inflammation in the conjunctiva tissue at POD 2, however, thisappeared to be related to the injection of the viscous suspension and recovered quickly, as shown by normal histology observations at later time points in Fig. 5.

Fig. 5.

Histological assessment of the ocular safety of HA and MC at POD 2, POD 7, and POD 30 showing conjunctiva tissue (A, B, C, G, H, I, M, N, O) and cornea (D, E, F, J, K, L, P, Q, R). Magnification is 400×.

Drug Release Kinetics In Vitro

Fig. 6. shows SUN release over a one-month period. For both HA and MC, SUN free drug dispersed in HA and MC 2% w/v viscous solutions exhibited the fastest release with an obvious burst release within the first three days. Drug release from PLGA MPs showed a more sustained profile that lasted over three weeks, the burst release of which was reduced but still evident. For the purpose of extended and controlled release, whereby the daily dosage of drug released should remain roughly constant, the burst release is undesirable. The daily release may drop below the desired drug concentration after a short period of time.

Fig. 6.

In vitro drug release profiles of SUN free drug and SUN-loaded MPs in 2% w/v HA and MC. The drug release of SUN-loaded MPs in PBS was used as a control.

By contrast, when drug-loaded MPs were suspended in HA or MC viscous solution, burst release was significantly reduced, as shown by the nearly linear curves in both cases, and the overall release was prolonged to over a month. We speculate that the sustained and stable drug release observed here is due to a complex combination of PLGA degradation, SUN diffusion from PLGA MPs, and SUN diffusion from surrounding HA and MC.

DISCUSSION

HA and MC were assessed as two potential viscoelastic vehicles for ocular drug delivery of MPs. There exist other shear-thinning polymers that can be employed for ocular applications, such as carboxymethylcellulose (CMC) and hydroxypropylmethylcellulose (HPMC) (32, 33). However, the viscous solutions of these polymers can exhibit gel-like behaviors and cause the clogging of fine-gauge needles. Therefore, the shear-thinning nature of viscous polymer solutions represent a promising direction toward facile injection through fine needles, especially when particle sizes are in the order of micrometers.

Besides shear-thinning properties, the ease of ocular injection with the assistance of HA and MC can also be attributed to the slower sedimentation of particles within viscous solutions than within PBS. Decreasing the amount of MPs suspended in the media can ameliorate sedimentation and clogging issues. However, for ocular administration routes such as SCT and intravitreal injection, the space of the subconjunctival tissue and vitreous chamber is limited. Thus, the volume of MP suspension injected ought to be minimized to prevent damage to the eye and reduce injection-related pain. Additionally, the higher concentrations of MPs suspended in viscous media may be required to achieve therapeutic levels of released drug over time from small volume injections. While injectability of MPs in PBS can be enhanced by shaking the syringe immediately prior to operation, MPs dispersed in PBS typically resettle within seconds (rate of settling depends on MP size), which makes PBS as a vehicle impractical in clinical settings. Therefore, the higher viscosities of HA and MC at low shear (e.g., zero shear inside the syringe barrel prior to injection) in comparison to PBS present an advantage in reducing sedimentation of MPs, further facilitating ocular injection.

With regard to subconjunctival retention of MPs after SC injection, the approximately 20% initial drop presumably resulted from the leakage of particles via the backflow of liquid medium following needle retraction. This leakage occurs since the hole punched by the needle cannot heal immediately post-operation. Previous studies have shown that viscous ophthalmic solutions decrease the rate of fluid removal from the administration site (15, 34, 35). In accordance to our rheological characterizations, approximately 80% of particles stayed in the conjunctiva tissue likely because the removal of shear stress resulted in the recovery of HA and MC to their original gel-like states, which trapped the particles in highly viscous polymer meshwork. Although the viscosities of HA and MC alone can adequately account for the retention of MPs over an extended period of time, the physical adhesion and attachment of HA or MC to the conjunctiva tissue may also play a role in enhancing and prolonging subconjunctival retention.

Due to the high molecular weights and chemical stability of HA and MC, viscous solutions might not be able to be rapidly cleared from the injection site. However, starting from PO 14 d, we could not observe elevated areas in the conjunctiva tissue due to HA or MC injection. The possible routes for clearance of HA and MC from conjunctiva tissues include enzymatic degradation of these polymers, systemic absorption by conjunctival vasculature, and transsceleral diffusion (36, 37). The absence of granulation tissue from histological examinations at PO 30 d suggests that the HA and MC viscous fluid vehicles studied here are likely to be well-tolerated following SCT injection.

CONCLUSION

Shear-thinning viscous HA and MC solutions were found to possess advantageous rheological properties to maintain MP suspension within syringes (viscoelastic solids at low shear) and to facilitate SCT injection of concentrated suspensions of SUN-loaded PLGA MPs (due to their strong shear-thinning behavior that dramatically reduced suspension viscosity as they passed through the syringe needle). MPs suspended in HA and MC solutions were tested for in vivo subconjunctival retention and toxicity as well as in vitro drug release. HA and MC viscous solutions remarkably improved the injectability of MPs, exhibited high retention of MPs post-injection, induced minimal injection-related inflammation that resolved within a week, reduced burst drug release in vitro and extended overall release.