Development of in situ forming implants for controlled delivery of punicalagin

Abstract

Due to efficient drainage of the joint, the development of intra-articular depots for long-lasting drug release is a difficult challenge. Moreover, a disease-modifying osteoarthritis drug (DMOAD) that can effectively manage osteoarthritis has yet to be identified. The current study was undertaken to explore the potential of injectable, in situ forming implants to create depots that support the sustained release of punicalagin, a promising DMOAD. In vitro experiments demonstrated punicalagin’s ability to suppress production of interleukin-1β and prostaglandin E2, confirming its chondroprotective properties. Regarding the entrapment of punicalagin, it was demonstrated by LC-MS/MS to be stable within PLGA in situ forming implants for several weeks and capable of inhibiting collagenase upon release. In vitro punicalagin release kinetics were tunable through variation of solvent, PLGA lactide:glycolide ratio, and polymer concentration, and an optimized formulation supported release for approximately 90 days. The injection force of this formulation steadily increased with plunger advancement and higher rates of advancement were associated with greater forces. Although the optimal formulation was highly cytotoxic to primary chondrocytes if cells were exposed immediately or shortly after implant formation, upwards of 70% survival was achieved when the implants were first allowed to undergo a 24-72 h period of phase inversion prior to cell exposure. This study demonstrates a PLGA-based in situ forming implant for the controlled release of punicalagin. With modification to address cytotoxicity, such an implant may be suitable as an intra-articular therapy for OA.

Graphical Abstract

1. Introduction

The global prevalence of osteoarthritis (OA) ranges from 2090 to 6128 cases per 100,000 population, with the highest burden of disease found in the United States.¹ The prevalence of disease is also rising; for example, it increased by 23.2% in the U.S. between 1990 and 2017.¹ The burden of OA is associated with pain, stiffness, decreased range of motion, and swelling which restricts activity and diminishes quality of life. In fact, OA is among the leading causes of years lived with disability,² and the medical cost in some developed countries may be as high as 2.5% of GDP.³ Therefore, there is strong interest in DMOADs that can be injected intra-articularly. The benefit of this localized delivery is that it maximizes drug activity at the target location, while minimizing exposure of other organs and the risk of unwanted side effects.⁴ Furthermore, orthopedists are trained to perform intra-articular injections for viscosupplementation, and the procedure is safe with low risk of infection.^5,6 However, because lymphatic drainage of intra-articularly injected drugs is so efficient, the drug dwell time is quite short. We propose that an in situ forming implant (ISI) may overcome this limitation. Such implants form through a process of controlled polymer transformation from a liquid phase to a solid phase. When a solvent/polymer/drug solution is exposed to water, the efflux and influx of solvent and water, respectively, cause the polymer concentration to increase until the polymer’s solubility limit is exceeded and it undergoes phase inversion, becoming a solid. Use of weaker solvents with low water miscibility allows for slow phase inversion and the formation of uniformly dense structures that exhibit zero-order drug release kinetics.⁷ In addition to their injectability, ISI”s are advantageous for their simple fabrication process and capability of sustaining drug release for up to one year.⁸

The current study is a first step in the development of an ISI for ultralong sustained delivery of punicalagin (PCG), a candidate DMOAD. PCG (MW 1084.71) is the major polyphenol present in pomegranate (Punica granatum L.), and it contributes to the anti-inflammatory properties of this fruit.⁹ Orally administered pomegranate fruit extract has been shown to lessen the severity of induced OA in rats and rabbits,^10,11 and daily intraperitoneal doses of PCG significantly reduced paw edema in an adjuvant-induced arthritis rat model.¹² We observed that semi-weekly, intra-articular injections of PCG seemed to result in less overall erosion of cartilage compared to a saline control in a monoiodoacetate-induced model of OA in rats.¹³

PCG is of particular interest because it targets cartilage degeneration and synovium inflammation. Our laboratory demonstrated that PCG can interact with collagenase in vitro and inhibit its enzymatic activity.¹³ It has been shown to exert a similar inhibitory effect on matrix metalloproteinase-13 mediated degradation of type II collagen, as well as on interleukin-1 beta-induced release of proteoglycan.¹² PCG can also prevent the degeneration of type II collagen by binding directly to it, which may block access of destructive enzymes to the fiber. For example, PCG was shown to bind non-covalently to collagen type II with high affinity via multiple hydrogen bonds (punicalagin has 17 hydroxyl groups) and π-π and electrostatic interactions.¹² In addition to inhibiting cartilage degeneration, PCG can also suppress synovium inflammation. For example, PCG suppressed lipopolysaccharide-stimulated production of the inflammatory mediators nitric oxide (NO), prostaglandin E2 (PGE2), and IL-6 cytokine by murine monocyte/macrophage-like cells in a dose-dependent manner.⁹

We have conducted preliminary testing to identify a suitable ISI for sustained delivery of PCG. Polycarbonate, poly(lactic acid), and poly(lactide-co-glycolide) (PLGA) were dissolved in various solvents and screened for combinations that could also dissolve punicalagin and release it gradually from an implant formed in situ. PLGA emerged as the polymer that offered the best opportunity for the tuning of PCG release kinetics. Systems composed of PLGA and PCG dissolved in N-methyl-2-pyrrolidone (NMP) and/or benzyl alcohol (BA) consistently formed solid or hydrogel implants. 2-Pyrrolidone (2P) also dissolved both PLGA and PCG, but it did not form well-defined implants. Although benzyl benzoate (BB) did not dissolve PCG, its inclusion in the system greatly slowed the rate of PCG release because BB is practically immiscible with water. Thus, the main objective of this study was to characterize PLGA ISI’s (originating from solutions containing NMP, BA, and BB) with respect to PCG stability and release kinetics. Demonstration of PCG chondroprotective properties, as well as evaluation of implant injectability and cytotoxicity, were also investigated.

2. Materials and Methods

Two types of LACTEL PLGA were from Evonik Corporation (Birmingham, AL, USA): 50:50 poly(DL-lactide-co-glycolide), inherent viscosity 0.95 - 1.20 dL/g and 75:25 poly(DL-lactide-co-glycolide), inherent viscosity 0.80 - 1.20 dL/g, both of which were ester-terminated. Acid-terminated poly(D,L lactide), inherent viscosity 0.16 - 0.24 dL/g, was from Sigma-Aldrich (St. Louis, MO, USA). NMP, BA, BB, and 2-methyltetrahydrofuran were also from Sigma-Aldrich. PCG (98%) from was from Chengdu Biopurify Phytochemicals Ltd. (Sichuan, China). Collagenase Type 2 was purchased from Worthington Biochemical (Lakewood, NJ, USA), and recombinant human ADAMTS-5 from R&D Systems (Minneapolis, MN). THP-1 human monocytes were from ATCC (Manassas, VA, USA). RPMI 1640 medium, fetal bovine serum (FBS), antibiotic-antimycotic solution, phorbol 12-myristate 13-acetate (PMA), lipopolysaccharides (LPS), dimethyl sulfoxide (DMSO), and Cell Counting Kit-8 (CCK-8) were from Sigma-Aldrich. ELISA kit for Interleukin-1β (IL-1β) was procured from Lifeome Biolabs, Inc. (Oceanside, CA, USA), and a Blyscan^™ Glycosaminoglycan Assay kit from Biocolor (Carrickfergus, United Kingdom). LC-MS solvents (water, acetonitrile) and LC-MS grade formic acid were from Fisher Scientific.

2.1. Chondroprotective Properties of PCG

In addition to the inhibition of released collagenases, PCG is capable of suppressing the biosynthesis of inflammatory mediators that play important roles in the pathogenesis of OA. Such chondroprotective properties were investigated using THP-1 monocytes that had been differentiated into macrophages. THP-1 is a human leukemia monocytic cell line. THP-1 monocytes were expanded by suspension culture in RPMI 1640 medium supplemented with 10% FBS and 1% antibiotic-antimycotic solution. They were differentiated into macrophages by adding PMA to a final concentration of 20 nM and transferring them to monolayer culture for 72 h.¹⁴ Inflammation was then induced by treatment with 1 μg/ml LPS. PCG was first dissolved at 5 mM in complete culture medium. One hour after LPS stimulation, cells were treated with PCG at final concentrations of 4, 20, 100, and 500 μM. Culture supernatant was collected 24 h later and IL-1β quantified using ELISA. Results are reported as amount of IL-1β produced per cell (pg/cell).

In addition to IL-1β, prostaglandin E2 (PGE2) in the culture supernatants was quantified using mass spectrometry. The same culture supernatant used for the IL-1β ELISA was fortified with a cocktail of deuterated internal standards (TxB2-d4, PGE2-d4, 15-HETE-d8, AA-d8, 100 pmol each, Cayman Chemicals). An equal volume of 3:2 hexane:isopropanol (v/v) + 0.1% acetic acid was mixed with the medium by vortex mixing for 1 min. Following centrifugation of the mixture (800 x g, 5 min), the top organic layer was collected and the medium extracted again with an equal volume of hexane. The subsequent hexane layer was combined with the first extract and the organics evaporated to dryness under nitrogen. The residues were then reconstituted in 100 μL of methanol and analyzed by liquid chromatography with tandem mass spectrometry (LC-MS/MS), as previously described.¹⁴ Results are reported as the amount of PGE2 produced per cell (pmol/cell).

2.2. Implant Formation and Measurement of Released PCG

Implant formulations are specified by the volume ratio of solvents and mass concentration of polymer and PCG. Each of approximately 0.5 ml was started by mixing solvents in the desired v/v ratio in a microcentrifuge tube and then adding PCG to a final concentration of 32 mg PCG per gram of calculated total implant weight. Once the PCG had dissolved, the appropriate amount of polymer was added based on the desired mass concentration (15-25% w/w) and the solvents’ densities. The polymer was allowed to dissolve overnight in a 70 °C water bath. The solvent/drug/polymer mixture was stirred to achieve a homogeneous solution. In vitro cumulative drug release was determined using methods similar to those described in other studies of ISI’s, wherein the polymer solution is injected into an aqueous release medium and incubated under sink conditions at 37°C; thereafter, samples of release medium are periodically withdrawn and replaced by fresh medium.^8,15-17 The particular method employed in this study was most similar to that of Camargo et al.¹⁷ A single implant was produced by injecting 125 μl of the solution into a 13×100 mm disposable glass culture tube containing 3 ml distilled water with 0.25% sodium dodecyl sulfate (SDS). Each batch yielded 3 implants. The tubes were kept at 37 °C under orbital shaking at 100 rpm in a Thermo Scientific MaxQ 4000 Incubated/Refrigerated Shaker (Marietta, OH, USA). Release of PCG was characterized by periodically measuring the absorbance of the bathing solution at 378 nm using a Bio-Tek uQuant Universal Microplate Spectrophotometer (Winooski, VT, USA). PCG concentration was determined by comparison to a standard curve produced from serial dilutions of a PCG/distilled water solution.

2.3. Effects of Implant Formulation on Kinetics of PCG Release

Several experiments were performed to gain a better understanding of how implant formulation affects the release of PCG (Table 1). The long-term goal is to find a formulation that sustains steady release of PCG for up to 100 days with minimal burst release. The first experiment was to determine if either NMP or BA solvents were suitable by themselves when holding the polymer type and its concentration constant. PLGA 75:25 was the default polymer type because degradation by hydrolysis generally takes longer the higher the lactide:glycolide ratio. Note that BB could not be tested as the only solvent because it did not dissolve PCG. The next experiment was to determine the general effect of solvent water miscibility by changing the NMP/BB ratio while holding polymer type and concentration constant. Whereas BB is practically immiscible with water, NMP is completely miscible. Because it does not dissolve PCG, BB did not make up more than 50% of the total solvent. The effect of polymer type was then examined while holding BB constant at the intermediate level of 30%. PLA was included because it had been shown to be advantageous compared to PLGA for sustained delivery of a relatively high molecular weight drug using in situ forming microparticle systems.¹⁸ Having demonstrated a slight advantage of PLGA 75:25 over the others, an additional experiment was carried out to determine the effect of polymer concentration, again holding NMP/BB constant at 70/30. A final experiment was performed as a first step towards optimization. BA, which has moderate solubility in water, was included to replace a majority of the highly water soluble NMP.

Table 1.

Experiments designed to determine the effect of in situ forming implant formulation on rate of PCG release.

Experiment	Polymer (Concentration)	Solvents (Volume Ratio)
Effect of single solvent	PLGA 75:25 (20%)	NMP
	PLGA 75:25 (20%)	BA
Effect of solvent water miscibility	PLGA 75:25 (20%)	NMP/BB (50/50)
	PLGA 75:25 (20%)	NMP/BB (70/30)
	PLGA 75:25 (20%)	NMP/BB (90/10)
Effect of polymer type	PLGA 50:50 (20%)	NMP/BB (70/30)
	PLGA 75:25 (20%)	NMP/BB (70/30)
	PLA (20%)	NMP/BB (70/30)
Effect of PLGA concentration	PLGA 75:25 (15%)	NMP/BB (70/30)
	PLGA 75:25 (20%)	NMP/BB (70/30)
	PLGA 75:25 (25%)	NMP/BB (70/30)
Optimization of PCG release	PLGA 75:25 (20%)	NMP/BA/BB (20/40/40)

2.4. PCG Stability within In Situ Forming Implants

The stability of PCG within a PLGA implant was indirectly examined by determining the ability of released PCG to inhibit collagenase-mediated cartilage degradation.^19,20 Pig stifle (knee) joints were obtained from a local meat processor and full-thickness, cylindrical disks of articular cartilage (5 mm diameter) were harvested from the distal femurs. The disks were freeze-dried, weighed, and placed in Dulbecco's Modified Eagle Medium (DMEM) containing 0.0025% type 2 collagenase with and without 100 μM PCG (n=12 disks per group). Samples were incubated at 37 °C under gentle orbital shaking. At 3, 6, 9, and 12 days of incubation, three disks from each group were rinsed in distilled water, freeze-dried and weighed again, while the rest were allowed to continue incubation after adding fresh collagenase. The same experiment was repeated using implants made with 20% PLGA (50:50), NMP/BB 90/10, and 0 or 32 mg PCG per gram total implant weight (n = 12 per group). The NMP/BB 90/10 formulation was chosen to achieve a relatively high rate of PCG release during the 12-day study period. Implants were formed in microcentrifuge tubes by pipetting 100 μl of the solution into distilled water. One hour later the water was replaced by the collagenase solution. Custom 3D-printed baskets were used to suspend the cartilage disks in the collagenase just above the implants. Disks were collected and weighed at the same 3-day intervals, with replacement of the collagenase solution each time. PCG concentration in the spent collagenase solution was determined by measuring absorbance at 373 nm and comparing to a standard curve generated from serial dilutions of PCG dissolved directly into DMEM. For both experiments, data are reported as residual mass of cartilage (final weight/initial weight). The overall effect of PCG was analyzed by two-way ANOVA with treatment and time as factors. Statistically significant effects of PCG at each time point were determined by independent t-test (equal variances not assumed).

PCG stability within ISI’s was directly examined by analyzing the release medium from two different time intervals. PCG-containing ISI’s were fabricated using the NMP/BB 70/30 solvent mixture and 20% PLGA as above, with the exception of using distilled water without SDS as the release medium. The release medium was completely removed and replaced with fresh on Days 3, 7, 17, and 21, thus allowing for comparison of the PCG released during two different intervals, Day 3 – 7 and Day 17 – 21. The PCG levels in water were determined by injecting 10 μL of the samples onto a Waters BEH C18 column (2.1 x 50 mm) and chromatography using a Waters UPLC that was interfaced with a Thermo Quantis triple-quadrupole mass spectrometer. The LC mobile phase solvents were 95:5 H2O/ACN+0.1% formic acid (mobile phase A) and 95:5 ACN/H2O+0.1% formic acid (mobile phase B). The gradient program used was as follows: 0-0.5 min (100% A, 0%B), 0.5-4.9 min (70%A, 30%B), 4.9-5.0 (70%A, 30% B), 5.0-5.1 min (100% A, 0%B), with column equilibration for 5 min. The column temperature was held at 30°C and the mobile phase flow rate was 0.2 mL/min. PCG was analyzed by negative electrospray ionization using the precursor-product transition 1083.1>600.8. An external calibration curve was prepared for PCG in water to enable its concentration in the experimental samples to be estimated.

2.5. Injectability

From Table 1, implants formulated to optimize PGC release were evaluated for their injectability. Injection forces were measured for implants formed using NMP/BB/BA (20/40/40) at 20% w/w PGLA (75:25). The solution was prepared as above; however, PCG was not added for the purpose of determining injectability. Hydrogel implants were formed by injecting 750 μl of the formulation into 10 ml of distilled water, followed by incubation at 37 °C under gentle orbital shaking for 24 hours (n=3). Each implant was pipetted into a 5 ml syringe following removal of the plunger. All air was aspirated from the syringe following the replacement of the plunger. The syringe was mounted upright in a Mach-1 V500C Mechanical Tester (Biomomentum, Inc., Laval, QC, Canada). A flat, smooth plate was lowered onto the plunger with a contact force of 0.098 N. The force to expel the implant through a 21G x 1.5 inch needle was then measured at a frequency of 100 Hz as the plunger was depressed at rates of 0.1, 0.3, and 0.5 mm/s Thirty seconds was the maximum time allowed for injection. Thus, the plunger was advanced a distance of 3 mm for each trial, which resulted in a total expelled implant volume of 400 μl. Friction of the plunger moving in an empty syringe, as well as the resistance of water, were tested for comparison.

2.6. Cytotoxicity

Cytotoxicity was examined using fresh porcine articular cartilage explants. As a promising candidate with respect to PCG release kinetics, implants composed of PLGA (75:25) dissolved in NMP/BA/BB (20/40/40) at 20% w/w were prepared as above in a 24-well tissue culture plate. PCG has previously been demonstrated to exhibit very low cytotoxicity.²¹ Nonetheless, the cytotoxicity experiment was carried out using ISIs with and without PCG. An implant of 70 μl was formed in each well. They were prepared 2, 24, 48, and 72 hours ahead of exposure to cartilage. The water/SDS was discarded and replaced with PBS every 24 hours and/or 1 hour prior to the introduction of explants. At the time of explant harvest, PBS was replaced with 1 ml of DMEM containing 10% FBS. Full-thickness disks of cartilage, 4 mm in diameter, were harvested from the medial condyle of the distal femur using a biopsy punch and scalpel. They were placed directly into the wells containing implants. Control disks were cultured in wells without implants. Explants were cultured for 24 hours in the presence of implants, at which time they were transferred to a new plate for viability assessment using the CCK-8 assay. Absorbance (450 nm – 650 reference) for each experimental sample was normalized to the average control absorbance, and results expressed as percent viability relative to control. Statistically significant differences with respect to control viability were determined using one-way ANOVA and Dunnett posthoc t-tests.

3. Results

3.1. Chondroprotective Properties of PCG

The anti-inflammatory nature of PCG was demonstrated by stimulating THP-1 macrophages with LPS. As shown in Figure 1A, PCG reduced LPS-stimulated IL-1β production by THP-1 macrophages in a concentration-dependent manner, with return to the unstimulated control level at 4 μM and suppression of practically all IL-1β synthesis at 20 μM and above. PCG’s effect on PGE2 synthesis was not as pronounced (Fig 1B), but it did suppress PGE2 levels. In the range of 4-100 μM, PCG decreased PGE2 production by about 35%. However, at 500 μM PCG, the LPS-stimulated production of PGE2 was attenuated by approximately 85%.

Figure 1.

Effect of PCG on production of IL-1β (A) and PGE2 (B) by human THP-1 macrophages (mean ± standard deviation, n=4). *p<0.05.

3.2. NMP or BA as Lone Solvent

When PCG and PLGA were dissolved in NMP solvent alone, phase inversion occurred very rapidly but the implants were immediately hard and crumbly. Those formed from PCG and PLGA dissolved in BA alone were initially soft hydrogels that contracted and stiffened over the course of many days (Fig. 2). PCG release profiles are shown in Figure 3. Both exhibited a burst biphasic drug release profile.²² In the case of NMP, a 2-day burst accounted for over 80% of the PCG loaded. The small amount remaining released quite slowly, and 95% had released within 20 days. For BA, approximately one-third of the PCG loaded was released in burst fashion. Thereafter the rate of release was practically zero. Furthermore, the remaining PCG was not recovered in the 20-day implants upon dissolving them in NMP.

Figure 2.

Images of in situ forming implants 3 days after formation. Each one was formed by injecting 125 μl of polymer solution (20% w/w PLGA) into 3 ml of distilled water with 0.25% SDS.

Figure 3.

PCG release from PLGA ISI’s (75:25 lactide:glycolide, 20% w/w) formed using NMP or BA as the only solvent.

3.3. Effect of solvent miscibility with water

BB is practically immiscible in water, and it was used to alter the overall solvent water miscibility by mixing it with NMP in various proportions. Physically, the implants were soft hydrogels that did not markedly change in size throughout the course of the experiment (Fig. 2). PCG release results are summarized in Figure 4. With the addition of BB to NMP, PCG displayed sigmoidal shaped release curves that indicate a delayed biphasic release profile characterized by an initial lag period followed by a rapid release in the power-law phase.²² The power-law phase refers to a period during which (Δc)~(Δt)ⁿ, where Δc is the cumulative amount of drug released and Δt is the elapsed time.²² There was a clear trend in increasing lag period with increasing BB content (lower water miscibility). Sustained PCG release for approximately 55 days was achieved using NMP:BB 50:50.

Figure 4.

Effect of NMP/BB ratio on PCG release from PLGA ISI’s (75:25 lactide:glycolide, 20% w/w).

3.4. Effect of polymer type

Polymer type had a profound effect on PCG release kinetics as shown in Figure 5. PLA implants were not well-defined and turned the water cloudy. The profile was burst biphasic, with approximately 60% of the loaded PCG detected in the water just two hours after implant formation, and the majority of what remained released within 24 hours. PLGA 50:50 and PLGA 75:25 were able to sustain PCG release for almost 60 days, but the release profiles were different. Whereas PLGA 75:25 produced a classic delayed biphasic profile, PCG release from PLGA 50:50 was roughly triphasic with burst, power-law, and accelerated release phases.²²

Figure 5.

Effect of polymer type on release of PCG from ISI’s made using NMP/BB 70/30 as solvents. Polymer concentration was 20%. 75:25 and 50:50 refer to lactide:glycolide ratio.

3.5. Effect of PLGA concentration

The influence of PLGA concentration on PCG release kinetics is shown in Figure 6. Implants created from 15% PLGA displayed a biphasic, perhaps triphasic, drug release profile, which included burst release of approximately one-third of the loaded PCG in two days. In contrast, PCG released from 20% and 25% PLGA implants in a delayed biphasic manner, and there was very little difference between the two higher concentrations.

Figure 6.

Effect of PLGA concentration on rate of PCG release from ISI’s made using NMP/BB 70/30 as solvents. PLGA lactide:glycolide ratio was 75:25.

3.6. Optimization of PCG release profile

The ideal PCG release profile would be monophasic, requiring several months to deliver the supply of drug. Based on the previous results, the first attempt to optimize the PCG release involved blending NMP, BA, and BB in a 20:40:40 ratio. Since little difference was observed between 20% and 25% PLGA, 20% was selected to facilitate dissolution and pipetting. The shape of the release profile (Fig. 7) does not closely resemble any of the typical shapes of in vitro cumulative release profiles that represent known types of drug release behavior from PLGA delivery systems. Release followed a pattern of short lag, burst, and then steady release between 14 and 80 days. Overall, PCG release was sustained for approximately 90 days in vitro.

Figure 7.

PCG release from ISI made using PLGA (75:25) at 20% w/w and NMP/BA/BB 20/40/40 as solvents.

3.7. PCG Stability within ISI’s

PCG significantly inhibited collagenase-mediated degradation of cartilage, whether it was added directly to the collagenase solution (Fig. 8A) or released from an ISI (Fig. 8B). The addition of 100 μM PCG limited cartilage mass loss to about 55% over 12 days of exposure to 0.0025% type 2 collagenase, whereas control cartilage disks lost approximately 95% of their mass in the same span of time. In the presence of an ISI containing PCG, the average cartilage disk exposed to collagenase for 12 days retained over 85% of its original mass. Control disks incubated in collagenase with ISI’s without PCG retained less than 10% of their original mass. The concentration of PCG released into the collagenase solution, as determined by UV/Vis spectrometry, was 158, 152, 82, and 96 μM on Days 3, 6, 9, and 12, respectively. These results demonstrate that PCG is stable within the PLGA ISI’s for at least 12 days.

Figure 8.

Inhibition of type 2 collagenase by PCG (mean ± standard deviation). A – PCG added directly to collagenase solution, B – PCG released from PLGA ISI’s. *p<0.05 compared to the control cartilage disks collected at the same time.

LC-MS/MS analysis confirmed the presence of punicalagin in the release medium (water) from an implant incubated under sink conditions for 21 days (Fig. 9). Punicalagin undergoes tautomerism; therefore, it exists as both α- and β-anomers in aqueous solution, which are detected as unique LC-MS/MS peaks. Total PCG concentration on Day 21, estimated by summing α and β peaks, was 195 μM, compared to 155 μM on Day 7. In both instances, the water had been completely removed and replaced with fresh water 4 days prior to collection. The PCG concentrations are consistent with a slightly accelerated rate of release during the later interval, which is demonstrated by the increased slope of the cumulative release curve in Figure 4. The chromatograms indicate a negligible difference in the PCG released during the second interval compared to the first, confirming its stability within the ISI over the 21-day evaluation period.

Figure 9.

LC-MS/MS chromatograms of medium (water) into which PCG had released from a PLGA ISI during two different time intervals, Day 3 – 7 and Day 17 – 21.

3.8. Injectability

The ideal ISI formulation would require similar injection forces to water or other commonly intra-articularly injected solutions, such as hyaluronic acid or corticosteroids. Additionally, the injection rate should allow an ISI to be fully injected within a reasonable time. Injection forces were found to be directly proportional to the injection rate for ISI, with a peak of 12.7 N at 0.1 mm/s and a peak force of 24.2 N reached during injection at 0.5 mm/s (Fig. 10). For comparison, a peak injection force for water at 0.5 mm/s injection rate was 1.23 N. These results demonstrate the ability to inject ISI through needles used during intra-articular injections.

Figure 10.

Injection forces for ISI made using PLGA (75:25) at 20% w/w and NMP/BA/BB 20/40/40 as solvents (mean ± standard deviation of three trials).

3.9. Cytotoxicity

Preliminary experiments demonstrated that implants were highly cytotoxic unless they were allowed to undergo a period of phase inversion to eliminate a substantial amount of solvent. Therefore, they were tested for cytotoxicity up to 72 hours after formation. As shown in Figure 11, a period maturation of 24 hours or more drastically decreases the implant’s cytotoxicity to primary chondrocytes. Cell survival was not affected by the release of PCG. The data suggest that a period of pre-incubation ≥24 hour to allow for removal of solvent did not render the implants completely safe, but significantly decreased viability could not be demonstrated given the degree of variability.

Figure 11.

Viability of primary porcine chondrocytes exposed to PLGA ISI’s (75:25, 20% w/w) made using NMP/BA/BB 20/40/40 as solvents (mean ± standard deviation). (+) PCG ISI’s contained 32 mg PCG per gram of total implant. *p<0.05 with respect to the non-exposed control group.

4. Discussion

Punicalagin is the major polyphenol present in pomegranate (Punica granatum L.) and contributes to its anti-inflammatory properties.⁹ PCG is of interest as an experimental DMOAD because there are multiple mechanisms by which it may be able to protect cartilage from degradation. Importantly, PCG has been shown to inhibit the production of pro-inflammatory cytokines, including TNF-α, IL-6, and IL-1β, by attenuating NF-κB/iNOS/COX-2/TNF-α and mitogen-activated protein kinases (MAPK) signaling pathways.^23,24 Suppression of NF-κB/iNOS/COX-2/TNF-α also lowers the production of PGE2, which, if left unchecked, enhances the expression of MMP-13 and ADAMTS-5 and accelerates extracellular matrix (ECM) degradation.²⁵ Furthermore, PCG can directly inhibit the secreted enzymes that degrade the cartilage ECM in OA. PCG forms complexes with collagenases in solution, including MMP-13, and inhibits their enzymatic activity.^12,13 Another way PCG can protect cartilage is by acting as a scavenger of reactive oxygen species. Punicalagin possesses multiple phenol groups that can participate in redox reactions by donating hydrogen atoms to oxidizing agents. Thus, it has potent electron-donating capacity, enabling it to efficiently reduce and detoxify free radicals, sparing vital macromolecules from attack. Finally, PCG binding to cartilage collagen may limit access to degradative enzymes. PCG binds to collagen type II with high affinity, which occurs through formation of multiple hydrogen bonds (PCG has 17 hydroxyl groups), in addition to π-π and electrostatic interactions.¹²

There is evidence that PCG is effective for the treatment of OA. Orally administered pomegranate fruit extract has been shown to lessen the severity of induced OA in rats and rabbits.^10,11 Daily intraperitoneal doses of PCG at 50 mg/kg significantly reduced paw edema in an adjuvant-induced arthritis rat model.¹² Our own study showed that semi-weekly injections of a solution containing 9.2 mM PCG seemed to result in less overall erosion of cartilage compared to a saline control in a rat model of osteoarthritis.¹³ As part of an investigation of PCG as a potential therapeutic compound for relieving rheumatoid arthritis, PCG was found to reduce the TNF-α induced expression of IL-1β, IL-6, and MMP-13 in fibroblast synoviocytes.²⁶ Such studies point to PCG as a promising DMOAD. Moreover, PCG safety is well established. For example, cytotoxicity is very low,²¹ and hematological and histopathological analysis of rats fed a 6% w/w PCG-containing diet for 37 days indicated no systemic toxicity.²⁷ Likewise, histology of major organs in our study turned up no evidence of systemic toxicity.¹³

The current study demonstrates additional chondroprotective properties of PCG. In a previous study, it was reported that PCG attenuated the LPS-induced production of IL-1β and PGE2 in RAW264.7 murine macrophages through disruption of NF-κB signaling pathway.²⁴ IL-1β is among the main proinflammatory cytokines implicated in OA.¹⁷ It triggers the secretion of additional inflammatory mediators and enzymes that degrade cartilage.¹⁶ Specifically, IL-1β activates chondrocytes, the cells responsible for maintaining cartilage, to increase the production of matrix metalloproteinases (MMPs), particularly MMP-1, MMP-3, and MMP-13. These enzymes target and degrade key components of the cartilage ECM, such as collagen fibers and proteoglycans. Furthermore, IL-1β stimulates the synthesis of prostaglandins and nitric oxide, which contribute to cartilage breakdown and inhibit the production of collagen and proteoglycans, essential for maintaining cartilage integrity. Eicosanoids, such as prostaglandins and leukotrienes, are lipid-based signaling molecules and mediators of inflammation. Substantially increased synthesis of eicosanoids is associated with progression from acute to chronic inflammation.⁵² Results presented herein show that PCG has suppressive effects on the LPS-induced production of IL-1β and PGE2 in human THP-1 macrophages. Thus, PCG’s persistence within a joint could reasonably be expected to decrease cartilage degeneration.

While oral administration of PCG is possible, it relies on transport through the systemic circulation to reach the intra-articular site of action. Localized delivery via intra-articular injection maximizes drug activity at the target location while minimizing exposure of other organs and the risk of unwanted side effects.⁴ Furthermore, orthopedists are trained to perform intra-articular injections for viscosupplementation, and the procedure is safe with low risk of infection.^5,6 Unfortunately, intra-articularly injected small molecule drugs are rapidly cleared from the synovial space via efficient lymphatic drainage. Therefore, it is highly advantageous that the drug be immobilized in an injectable depot formulation from which it undergoes sustained release. The current study explores the potential for PLGA-based ISI’s to act as such a depot. PLGA is an FDA-approved biopolymer, and PLGA ISI’s can deliver hydrophilic or hydrophobic drugs. They generally sustain release for much longer than conventional drug delivery systems. In situ forming implants have been administered by oral, ocular, rectal, vaginal, injection (intramuscular, subcutaneous) and intraperitoneal routes.²⁸ This study considers their prospective use for the intra-articular route of administration.

NMP and BA are solvents for both PLGA and PCG, but neither produced suitable implants by themselves as most PCG was released in burst fashion. Being highly miscible with water, NMP forms a drug depot very quickly, but it is highly porous and permeable. A high burst release is a consistent feature of ISI’s containing NMP and has been observed for polycaprolactone, polylactide, and poly(trimethylene carbonate), in addition to PLGA.¹⁵ With BA (very low water miscibility), because the rate of solvent exchange was much slower, a high proportion of PCG may have transferred into the aqueous medium before it became encapsulated. A mixture of NMP and BB, however, produced implants that could sustain PCG delivery for many weeks. It seems that PCG was encapsulated quickly but that the BB, which is practically immiscible with water, promoted the formation of a dense implant from which PCG diffused slowly. ISI’s containing BB alone could not be investigated because BB did not dissolve PCG.

Because almost all loaded PCG released in burst fashion, the particular PLA used in this study did not form a suitable drug depot. A difference in solubility may have been the culprit, but it was not tested. ISI’s made from PLGA 50:50 and PLGA 75:25 both sustained release of PCG for several weeks, but the patterns of release were different. As over half of the loaded PCG released from the PLGA 50:50 within two weeks, PLGA 75:25, from which approximately 12% of the loaded PCG released in that period, is the preferred candidate for further development.

With respect to PLGA concentration, burst release of PCG was mitigated by an increase from 15% to 20%, but a further increase to 25% had no additional effect. Polymer concentration is a well-known tuning parameter.²⁹ It has been shown to affect drug release through porosity.¹⁸ The higher viscosity of a more concentrated polymer solution usually leads to the formation of a more dense structure from which the drug releases more slowly.³⁰ The higher concentrations of PLGA (20%, 25%) formed highly viscous solutions, and it was difficult to work with concentrations above 25%. Based on results of system characterization studies, one attempt was made at optimization. ISI’s composed of PLGA (75:25) dissolved in NMP/BA/BB 20/40/40 at 20% w/w released PCG for approximately 90 days. However, about 50% of it released in under 10 days, and future studies are needed to identify a formulation that yields a monophasic-like release profile.

Stability of the PCG within the ISI is an important feature, as interaction with excipients can lead to drug degradation or inactivation.²⁹ Judging from the collagenase inhibition experiment, there was no loss of functionality of PCG released from PLGA ISI’s. ISI’s were highly chondroprotective of cartilage explants exposed to type 2 collagenase for 12 days. In fact, there was no evidence of collagenase-induced erosion from Day 6 to Day 12. Day 12 explants retained approximately 85% of their original mass compared to about 15% for controls. LC-MS/MS confirmed the release of undegraded PCG during Day 3-7 and Day 17-21 intervals. Furthermore, the concentration of PCG detected was consistent with the measured PCG release kinetics. This study did not investigate the possibility that the degradation of PLGA could be accelerated by PCG. PCG’s many hydroxyl groups could interact catalytically to increase exposure of PLGA’s ester linkages to water, thereby facilitating PLGA hydrolysis.²⁹

The aforementioned ISI capable of sustaining PCG release in vitro for up to 90 days was further investigated with respect to injectability. For all rates tested, the force required to depress the plunger steadily increased, and the faster the ejection rate, the higher was the rate of force increase. During the injection of 400 μl through a 21G needle, the average peak forces ranged from approximately 13 to 24 N. For reference, data from ten evaluators were used to establish the upper threshold force for smooth subcutaneous injection as 35 N, with injection feasible up to 45 N, albeit with difficulty.³¹ Extrapolating the data from the 0.5 mm/s tests, the force required to inject a 1 ml ISI solution in about 10 s would just reach 45 N, assuming the joint space did not offer substantial resistance.

Solvent toxicity is one of the key issues to be addressed in the development of ISI’s.³² It is especially important in the context of the intra-articular route of administration because the integrity of articular cartilage extracellular matrix depends on chondrocyte viability,³³ and there is no vascular source of cell replenishment. Chondrocyte death is a prominent feature of OA,³⁴ and it could lead to rapid cartilage degradation, as is observed in the monoiodoacetate-induced model of OA in rats.¹³ Due to concerns about chondrocyte survival, the current study addressed excipient cytotoxicity. Unfortunately, ISI’s composed of PLGA (75:25) dissolved in NMP/BA/BB were acutely cytotoxic to chondrocytes in cartilage explants. Similarly, solvents such as NMP that were used in the formulation of ISI’s for intramuscular injection were found to be cytotoxic to myocytes.³⁵ While an extended period of phase inversion and solvent removal prior to cartilage explant exposure did lead to greater chondrocyte survival, it was insufficient to prevent a 20% decrease in viability of ISI-exposed chondrocytes relative to controls. As expected from its excellent safety profile, PCG released from ISI’s had no effect on chondrocyte viability.^21,27 These data suggest that alternative approaches, such as in situ forming microparticles, may be needed. In such systems, dissolved PLGA constitutes an internal phase, which is emulsified into a biocompatible external oily phase such as sesame oil. When the emulsion is injected into an aqueous environment, solvent diffuses out of the droplets, and the PLGA and drug precipitate and form microparticles.³⁶ In situ forming microparticles are reported to have lower toxicity compared to ISIs.^35,37 Furthermore, they may be made even less cytotoxic by using sucrose acetate isobutyrate, which dissolves in a much smaller amount of organic solvent than PLGA, to replace a substantial amount of the PLGA.³⁷

The current study has several limitations, the main one being that it involved only in vitro experimentation. The in vivo performance of an in situ forming drug delivery system may differ considerably from the in vitro release profile. For example, a pharmacokinetic study of a rivastigmine-in situ forming microparticle system in rabbits showed that the plasma concentration of drug was maintained for approximately 24 hours, which did not match the month-long in vitro release profile.³⁷ Furthermore, safety of the PLGA ISI system was evaluated only with respect to chondrocyte survival. This study did not address the sensitivity of other cells found in a joint (e.g. synoviocytes and ligament fibroblasts), the potential for solvents released into the systemic circulation to affect other organs, or the effects of long-term exposure to the ISI. The kinetics of in vitro PCG release were determined only for selected polymer-solvent combinations using water as the release medium, and the system remains far from optimized. Lastly, injection forces were measured only for implants of a single formulation drawn through a 21G needle 24 hours after their formation. While this provides a preliminary understanding of the fluid properties of ISI, other formulations need to be examined before any conclusions can be made about their injectability. The high viscosity of formed implants resulted in an inability to draw them into the 5 ml syringe during testing, which limits their usability in a clinical setting. ISI volume was also lost during transfer from incubation containers to syringes due to limited accuracy when pipetting the medium. While further work needs to be done to understand whether current formulations of ISI can be reasonably administered, more important issues such as biocompatibility take precedence.

5. Conclusions

This study demonstrated that PCG suppressed production of IL-1β and PGE2 by human THP-1 macrophages and that PCG was stably incorporated into PLGA-based ISIs for controlled release. PCG release from ISIs is tunable, and at least one formulation sustained release for up to 90 days. The force needed to inject ISI’s of the same formulation is marginally acceptable, and cytotoxicity was greatly reduced when the ISI was allowed at least 24 h for removal of solvent before its presentation to cartilage explants. Overall, this study revealed that PCG-releasing ISI’s are worthy of additional investigation as an approach to localized, long-lasting release of an experimental DMOAD.