A Prodrug Approach Involving In Situ Depot Formation to Achieve Localized and Sustained Action of Diclofenac After Joint Injection

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

A Prodrug Approach Involving In Situ Depot Formation to Achieve Localized and Sustained Action of Diclofenac After Joint Injection 1 ˚ ARDH, ˚ METTE THING,1 LI AG SUSAN LARSEN,1 RUNE RASMUSSEN,1 JAKOB PALLESEN,1 NINA MERTZ,1 JESPER KRISTENSEN,2 2 MARTIN HANSEN, JESPER ØSTERGAARD,1 CLAUS SELCH LARSEN1,3 1

Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark 3 DepXplora Aps, Gammelbyvej 17, Lejre DK-4320, Denmark 2

Received 15 August 2014; revised 15 September 2014; accepted 30 September 2014 Published online 29 October 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.24221 ABSTRACT: Long-acting nonsteroidal anti-inflammatory drug formulations for intra-articular injection might be effective in the management of joint pain and inflammation associated sports injuries and osteoarthritis. In this study, a prodrug-based delivery system was evaluated. The synthesized diclofenac ester prodrug, a weak base (pKa 7.52), has relatively high solubility at low pH (6.5 mg mL−1 at pH 4) and much lower solubility at physiological pH (4.5 ␮g mL−1 at pH 7.4) at 37◦ C. In biological media including 80% (v/v) human synovial fluid (SF), the prodrug was cleaved to diclofenac mediated by esterases. In situ precipitation of the prodrug was observed upon addition of a concentrated slightly acidic prodrug solution to phosphate buffer or SF at pH 7.4. The degree of supersaturation accompanying the precipitation process was more pronounced in SF than in phosphate buffer. In the rotating dialysis cell model, a slightly acidic prodrug solution was added to the donor cell containing 80% SF resulting in a continuous appearance of diclofenac in the acceptor phase for more than 43 h after an initial lag period of 8 h. Detectable amounts of prodrug were found in the rat joint up to 8 days after knee injection of C 2014 Wiley Periodicals, Inc. and the American Pharmacists Association J Pharm Sci 103:4021–4029, 2014 the acidic prodrug solution.  Keywords: drug delivery systems; joint injection; injectables; non-steroidal anti-inflammatory drugs; precipitation; prodrugs; suspensions; physicochemical properties

INTRODUCTION Generally, the rationale for intra-articular (IA) depot injectables is to maintain high therapeutic drug concentrations in the joint over extended time periods while keeping systemic drug levels low to minimize side effects.1 A significant dose reduction (by a factor of 3000 or more), as compared with intravenous administration, can be achieved by direct injection into the joint target area.2 Local joint injection of aqueous microcrystalline suspensions of anti-inflammatory corticosteroid esters, for example, methylprednisolone acetate (Depo-Medrol ), is common practice in the relief of pain- and inflammation-associated rheumatoid arthritis3,4 as well as osteoarthritis.5 Also, oral nonsteroidal anti-inflammatory drugs (NSAIDs) are used for these indications but their use may be prohibited because of the emergence of severe adverse outcomes even after short-term use.6 As this class of drugs possesses anti-inflammatory as well as antinociceptive properties,7,8 long-acting NSAID formulations for IA injection might afford optimal alleviation of joint associated pain and inflammation arising from minor arthroscopic procedures and sports injuries that do not require surgery. Also, as regards flare up pain episodes in osteoarthritis, IA NSAID depot-based therapy may be feasible and may possibly replace corticosteroid injections.9 R

Correspondence to: Claus Selch Larsen (Telephone: +45-35336466; E-mail: [email protected]) This article contains supplementary material available from the authors upon request or via the Internet at http://onlinelibrary.wiley.com/. Journal of Pharmaceutical Sciences, Vol. 103, 4021–4029 (2014)

 C 2014 Wiley Periodicals, Inc. and the American Pharmacists Association

Like most dissolved low-molecular-weight drugs, NSAIDs disappear relatively fast from the synovial space after IA injection.1 IA injections of aqueous solutions of NSAIDs may, at best, provide pain relief for up to 6 h after arthroscopic procedures.10 In comparison, the duration of the action of the long-acting steroid ester prodrugs is influenced by the intrinsic solubility of the ester derivatives as well as their enzymemediated cleavage rate in the synovial fluid (SF).11,12 In analogy to these steroid suspensions, depot NSAID injectables might be developed in the form of suspensions comprising poorly soluble NSAID ester prodrugs. Parenteral suspensions have attractive attributes, for example, a high drug load involving a minimum of pharmaceutical excipients.1 On the contrary, this formulation type may pose major challenges as regards the manufacturing process (terminal sterilization13 ) and the physical stability on storage (particle size distribution14 ). A means to circumvent such barriers to dosage form development is to use a liquid formulation (a preformulation), which after IA injection forms the depot suspension in situ upon contact with the SF. Previously, various approaches to provide in situ depot formation after other routes of administration have been investigated.15,16 Here, we introduce a prodrug approach17 to enable local and sustained action of diclofenac after joint injection. The potential utility of the prodrug approach taken is illustrated mainly by in vitro characterization but also preliminary in vivo data in the rat are presented, demonstrating substantially extended joint residence times of the diclofenac prodrug. A sketch of the contemplated prodrug strategy is presented in Figure 1. The role of the promoiety or immobility promoting unit (IPU) is to confer the synthesized diclofenac ester prodrug, a pH-dependent

Thing et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:4021–4029, 2014

4021

4022

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Figure 1. Schematic representation of the in situ precipitation formulation principle. After injection of a slightly acidic solution of the prodrug, the prodrug precipitates in the joint because of the lower solubility at pH 7.4. The prodrug in solution is either enzymatically cleaved to the active compound and the IPU or transported intact out of the joint.

solubility, that is, a suitable high solubility at slightly acidic pH and a low solubility in the (patho) physiological pH range close to 7.4. This can be accomplished by employing IPUs that in addition to an OH group comprise a weak base functionality with a pKa value of about 5–8. In this manner, it is possible to obtain a concentrated slightly acidic diclofenac prodrug solution (the preformulation), which after injection into the joint leads to in situ precipitation of the neutral form of the prodrug on contact with the SF (pH 7.4). To exert pharmacological activity, the ester prodrug has to be converted to active diclofenac by enzyme cleavage of the prodrug bond mediated by esterases and other hydrolases residing in the SF. This activation process only takes place for the small amount of dissolved prodrug that is in equilibrium with precipitated prodrug in the SF. The objective of the current work was to investigate in vitro the potential suitability of the above outlined prodrug concept through characterization of ionization and solubility properties, prodrug stability in aqueous solution, parent compound regeneration in biological media, in situ precipitation behavior, and release in biological media. A further objective was to investigate the synovial joint residence time of the prodrug after IA injection in the rat.

MATERIALS AND METHODS Materials The following chemicals were purchased from Sigma–Aldrich (Munich, Germany): diclofenac for synthesis, 2-(1-methyl1H-imidazol-2-yl)ethanol, 4-dimethylaminopyridine (DMAP), dichloromethane, dicyclohexylcarbodiimide (DCC), methanol, acetonitrile, Tween 80, bovine serum albumin (BSA) > 98%, and NH3 . Diclofenac > 98% was obtained from TCI (Tokyo, Japan). All other chemicals used were of analytical grade or highest grade possible. Demineralized water was used throughout the study. Diclofenac sodium, fluid for injection, 25 mg mL−1 (Novartis Healthcare A/S) was purchased from Nomeco (Copenhagen, Denmark). Outdated human plasma was obtained from Rigshospitalet (Copenhagen, Denmark). SF from arthritis patients was obtained from the Parker Institute, Frederiksberg Hospital (Frederiksberg, Denmark). SF samples from several patients with different diagnoses were pooled and frozen in Thing et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:4021–4029, 2014

Figure 2. Chemical structure of the diclofenac prodrug DPX-1-0011, 2-(1-methyl-1H-imidazol-2-yl)ethyl 2-(2-(2,6dichlorophenyl)amino)phenyl)acetate.

smaller portions for later use. Rat serum was obtained from untreated Sprague–Dawley rats immediately after execution. Synthesis of DPX-1-0011 [2-(1-Methyl-1H-Imidazol-2-yl)Ethyl 2-(2-(2,6 Dichlorophenyl)Amino)Phenyl)Acetate] Diclofenac-free acid (2.96 g, 10 mmol), 2-(1-methyl-1Himidazol-2-yl)ethanol18 (10 mmol) and DMAP (122 mg, 1 mmol) were dissolved in dichloromethane (25 mL) and cooled to 0◦ C in an ice bath. A solution of DCC (4.12 g, 20 mmol) in dichloromethane (25 mL) was added dropwise over the course of 30 min. After complete addition, the reaction was allowed to reach room temperature over 3 h. The reaction mixture was filtered and the filtrate evaporated to give a pale yellow oil, which was purified by flash chromatography (0%–10% 2 M methanolic NH3 in ethyl acetate) to give DPX-1-0011 (Fig. 2) in 67% yield as colorless crystals. See Supplementary Information for methods and results of the characterization of the prodrug. Solubility Studies Excess DPX-1-0011 was added to either 67 mM phosphate buffer (PBS) pH 7.4 or water to which 0.1 M hydrochloric acid was added to lower pH. Mixtures were rotated at 37 ± 0.5◦ C in an incubation hood and samples were withdrawn after approximately 24 and 48 h or later. Withdrawn samples were filtered through a 0.45-:m Millex –HV (Millipore, Billerica, MA) low protein-binding filter. After discarding the first 0.5 mL, filtrates R

DOI 10.1002/jps.24221

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

were diluted with PBS (pH 7.4 samples) or the HPLC mobile phase (the slightly acidic samples) prior to quantitative analysis by HPLC. Experiments were performed in triplicate. pKa Measurements pKa values of DPX-1-0011 and the corresponding IPU [2-(1methyl-1H-imidazol-2-yl)ethanol] were determined at 25 ± 1◦ C and 37 ± 1◦ C by potentiometric titration in the pH range 2.5–10 with a methanol content of 29.2%, 39.1%, and 49.5% (v/v) using a GLpKa-meter (Sirius Analytical, East Sussex, UK). Yasuda– Shedlovsky extrapolation was used to estimate pKa values at zero methanol concentration.19 Preparation of Preformed Suspension and Acidic Prodrug Solution The preformed suspension (10 mg mL−1 DPX-1-0011) was prepared by adding a solution of 0.1% (v/v) Tween 80 in 67 mM PBS to DPX-1-0011. The suspension was treated with ultrasound (Branson 2510 Ultrasonic cleaner; Fisher Scientific, Schwerte, Germany) for 10 min. The acidic prodrug solution was prepared by dissolving 10 mg prodrug per milliliter in a solution of 0.025 M HCl containing 8 mg sodium chloride per milliliter. This mixture was stirred at ambient temperature overnight and then filtered through a 0.2 :m filter. For the animal studies, a sterile acidic solution was used to dissolve the prodrug and the injection solution was filtered under aseptic conditions. Upon repetition, pH of the resulting prodrug solution was 3.0–3.5 at room temperature, n = 5. The preformed suspension and the acidic prodrug solution were always used within 48 h after preparation. Stability in Human SF and Plasma, Rat Serum, and Aqueous Solutions and Suspensions Solutions containing 80% (v/v) solutions of human plasma and SF as well as rat serum were made by mixing one part of a 67-mM PBS (pH 7.4) with four parts (by volume) of the respective biological matrices. Final reaction solutions were adjusted to pH 7.4 by addition of 0.1 M hydrochloric acid when needed. The stability of DPX-1-0011 at 37 ± 0.5◦ C was determined in PBS (10 mL), 80% SF and 80% human plasma (2.5 mL), and 80% rat serum (1.0 mL) to which 10, 20, or 50 :L of freshly prepared prodrug solution in methanol were added to give a final prodrug concentration of approximately 4.3 :g mL−1 . The stability was assessed by monitoring the decrease of intact prodrug concentration as a function of time. Stability of DPX-1-0011 at ambient temperature under acidic conditions was determined by splitting an acidic prodrug solution in three. In two of these portions, pH was slightly elevated by addition of 0.1 M NaOH to cover the pH range determined in the prepared acidic prodrug solutions (pH 3.0–3.5). Stability was evaluated by following the formation of diclofenac. Stability of DPX-1-0011 in the preformed suspensions (10 mg mL−1 ) at 37◦ C was evaluated by measuring the content of regenerated diclofenac as a function of time. Withdrawn samples were centrifuged (5 min, 14,000 × g) and aliquots of the supernatant were analyzed by HPLC. To samples taken from the biological media twice, the volume of acetonitrile was added. Mixtures were vortexed and after centrifugation (5 min, 14,000 × g) the supernatant was analyzed by HPLC. The 700 :L PBS samples were added to 100 :L PBS and mixed before analysis by HPLC. DOI 10.1002/jps.24221

4023

Determination of Particle Size Distribution The particle size distribution in the preformed prodrug suspensions was measured by laser diffraction with Malvern Mastersizer 2000 particle size analyzer using the Hydro 2000S wet sample dispersion unit (Malvern Instruments Ltd., Worcestershire, UK). Approximately 400 :L 10 mg mL−1 preformed DPX1-0011 suspension was added to 67 mM PBS pH 7.4 (120 mL) in the dispersion unit until the laser obscuration range was maintained between 2% and 5%. Suspension was stirred at 1015 rpm and a measurement time of 12 s was applied. The in situ suspensions formed upon addition of the acidic DPX-1-0011 solution to 67 mM PBS pH 7.4 or 80% (v/v) SF (2:1, v/v) were examined in a Leica DM750 light microscope (Ballerup, Denmark) 2 min and 24 h after mixing. The mixture was rotated at 37◦ C between sampling. For comparison reasons, microscope pictures of SF were also taken following addition of the acidic vehicle pH 3.5 in equivalent amounts to SF. In Situ Precipitation Studies An aliquot of 2.00 mL of the acidic prodrug solution was added to 4.00 mL preheated PBS, 80% SF, or 2% BSA in screwcapped test tubes, vortexed, and then rotated at 37 ± 0.5◦ C in an incubator hood. Samples were withdrawn at appropriate time intervals after addition of the prodrug. Samples were applied to a glass insert in an eppendorf tube and centrifuged (5 min, 14,000 × g). The supernatant of the PBS samples was diluted 1:4 or 1:5 (v/v) with PBS prior to analysis. In case of the biological samples, 200 :L of the supernatant was added to 400 :L acetonitrile, vortexed, and centrifuged (5 min, 14,000 × g). The supernatant was analyzed by HPLC, if needed after appropriate dilution. Release from the Rotating Dialysis Cell The release experiments in the rotating dialysis cell model were carried out at 37 ± 0.5◦ C as previously reported.20 The cumulated amount released, MA,t , of either diclofenac or DPX-1-0011 into the acceptor compartment was calculated according to:

MA,t = VS

n 

Ci−1 + VA Cn

(1)

i=1

where VS and VA are the volumes of the samples withdrawn from the acceptor phase and the volume of the acceptor phase at time t, respectively. Ci is the concentration in the sample i. Evaluation of Rat Joint Residence Time Twelve male Sprague–Dawley rats weighing 300–400 g (Taconic, Ejby, Denmark) were used in the study. Rats were housed at the animal facility at Department of Drug Design and Pharmacology, University of Copenhagen, Denmark. All rats were kept in a temperature-controlled room (20◦ C– 22◦ C) and subjected to regular 12 h light-dark cycles with free access to food and water. Animals were allowed to acclimatize for 7 days prior to use. The experimental protocol was preapproved by the Danish Animal Experimental Board (No. 2009/561-1622) and all procedures were carried out according to the Danish Animal Test Act. Thing et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:4021–4029, 2014

4024

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

For detailed information on injection procedure, sample preparation and analysis by liquid chromatography-tandem mass spectrometry (LC–MS/MS) see Supporting Information. Analysis Samples from the stability, solubility, precipitation, and release experiments were analyzed on an Elite LaChrom HPLC system (VWR International, Tokyo, Japan) employing a Merck-Hitachi L-2130 pump connected to a Merck Hitatchi L-2450 diode array detector and a VWR-Hitachi L-2200 autosampler. Reversedphase chromatography was performed using a Gemini C18 column (150 × 4.60 mm2 ; 5 :m particles; Phenomenex, Værløse, Denmark) equipped with a Gemini C18 precolumn (4 × 3.0 mm2 ; Phenomenex) heated at 30◦ C by a VWR Hitachi L-2300 column oven. The mobile phase consisted of methanol and 0.02 M acetate buffer pH 4.4 in a ratio (v/v) of 65:35. The injection volume was 20 :L, the flow rate was set at 1 mL min−1 , and the column effluent was monitored at 275 nm. Calibration curves were constructed by diluting stock solutions of the prodrug in acetonitrile. Further dilutions were performed either in PBS (for the release studies), mobile phase (for the acidic prodrug solutions), or a mixture of PBS-H2 O–acetonitrile (1:4:10, v/v/v, for samples containing the biological matrices). For all experiments, solvents used for the standards were similar to those of the samples. The concentration of the highest standard was 0.105 mg DPX-1-0011 mL−1 for the in situ suspensions diluted in mobile phase. The lowest standard used was 0.19 :g DPX1-0011 mL−1 diluted in PBS in the release studies. Limit of quantification (LOQ) was approximately 0.16 :g mL−1 for DPX1-0011 and 0.14 :g mL−1 for diclofenac as determined from the slope of the standard curves.

RESULTS AND DISCUSSION The purity of the synthesized imidazole ester diclofenac prodrug, DPX-1-0011 (Fig. 2), exceeded 97% as assessed by nuclear magnetic resonance (NMR). After incubation in 80% (v/v) human plasma, the prodrug was converted to parent diclofenac in a quantitative manner.

Figure 3. pH solubility profile for DPX-1-0011 at 37◦ C (n = 3). The full line is obtained by fitting the data to Eq. 2 using a pKa of 7.52.

where St is the total solubility at a given pH and S0 represents the intrinsic solubility of the free base form. S0 was estimated from Eq. 2 to 2.0 ± 0.05 :g mL−1 (mean ± SD). The favorable agreement between experimentally and fitted solubility data (Fig. 3) may suggest that diclofenac formed during the solubility experiments did not significantly affect the equilibrium solubilities. The obtained solubilities at pH 7.4 and 4.0 of 4.5 :g mL−1 and 6.5 mg mL−1 , respectively, show that an increase of pH from 4.0 to 7.4 results in a DPX-1-0011 solubility decrease amounting to a factor of 1400. Whenever possible, pH of injectables should be kept close to physiological pH in order to minimize irritation and eventual tissue damage. In case of low-dose products and the use of injection sites where rapid exchange of electrolytes between the tissue fluid and the blood takes place, pH of the injection solution might be less critical. To this end, intramuscular injection of acidic (pH about 3) morphine chloride solutions (20 mg mL−1 ) is common practice.22 Likewise, the SF is a vital tissue fluid, the small-molecule constituents of which are continually cleared and replenished in an ongoing equilibrium with the perfusing blood.23 Hence, IA injection of a slightly acidic preformulation solution of the diclofenac prodrug in doses of 10 mg or less is expected to be well tolerated. Stability Evaluation

Solubility properties ◦

The pKa values of DPX-1-0011 (7.66 at 25 C and 7.52 at 37◦ C) and the corresponding IPU, 2-(1-methyl-1H-imidazol-2yl)ethanol, (7.65 at 25◦ C and 7.43 at 37◦ C) were determined by potentiometric titration. Similar values have been reported for other imidazole derivatives with alkyl substituents in the 2-position.21 The data reveal a slight tendency toward a decrease in pKa with increasing temperature. It appears that the two methylene groups situated between the ester bond and the imidazole ring effectively shield for the electron-withdrawing properties of the ester function because the ionization constants of the prodrug and the IPU are almost identical. The aqueous solubility of DPX-1-0011 was determined as a function of pH at 37◦ C as shown in Figure 3. The solid symbols represent the experimentally determined equilibrium solubilities (RSD ≤ 3%). The full line drawn is obtained by fitting these values to 2 using a pKa value of 7.52,   St = S0 1 + 10(pKa−pH) Thing et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:4021–4029, 2014

(2)

In connection with the solubility experiments, rough estimates of the stability of DPX-1-0011 in acidic solution (pH 3.0–3.5) were obtained by following the formation of diclofenac as a function of time (using the initial rate approach) in the DPX-10011 reaction solutions by HPLC. At ambient temperature, the time for 5% degradation of the prodrug in the investigated pH range was estimated to be ≥12 days. In 67 mM PBS (pH 7.4) at 37◦ C, the degradation of DPX1-0011 in dilute solution followed apparent first-order kinetics. The pseudo-first-order rate constant, (khyd = 1.22·10−3 ± 1.8·10−5 min−1 , n = 6) was determined from the slope of linear plots of the logarithm of the concentration of intact prodrug against time from degradation reactions followed for at least two half-lives. Likewise, the degradation of DPX-1-0011 in 80% (v/v) human plasma (khyd = 3.3·10−2 ± 4.7·10−4 min−1 , n = 3) and SF (from arthritis patients, khyd = 1.3·10−2 ± 8.5·10−5 min−1 , n = 3) as well rat serum (khyd = 5.3·10−1 ± 2.6·10−2 min−1 , n = 3) obeyed pseudo-first-order kinetics. These observations strongly suggest that the DPX-1-0011 ester bond is subject to significant enzyme-mediated cleavage by esterases or DOI 10.1002/jps.24221

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

4025

creases with decreasing pH 3 Eq. (3). To this end, X-ray powder diffraction (XRPD) experiments revealed that crystals of the preformed and in situ-formed suspension have crystal packing identical to that of the solid free base form of DPX-1-0011. Preliminary Evaluation of the Particle Precipitation Process

Figure 4. Formation of diclofenac from an in situ formed (resulting pH 7.0) and a preformed suspension (resulting pH 7.4) at 37◦ C in 67 mM PBS containing 0.03% Tween 80 with an amount of prodrug added corresponding to 3.3 mg mL−1 . Bars represent standard deviation (n = 3). Full lines are linear regressions of the data.

other hydrolases residing in the three biological matrices with the fastest diclofenac regeneration occurring in rat serum. The cleavage rate of DPX-1-0011 in human plasma is about twice that in human SF. Similar cleavage rate ratios between the two biological matrices have been observed for naproxen glycolamide ester prodrugs.23 The hydrolytic stability of a preformed DPX-1-0011 suspension (3.3 mg mL−1 ) in PBS containing 0.03% Tween 80, added for wettability purposes, at 37◦ C was investigated by measurement of the concentration of regenerated diclofenac in the supernatant as a function of time (Fig. 4). During the experiments carried out in triplicate (7 days rotation), the rate of diclofenac formation was constant (in total less than 0.5% of the initial amount of DPX-1-0011 was cleaved). Hence, under these conditions, the degradation applied to pseudo-zero-order kinetics according to: k0 = khyd [DPX−1−0011]saturated

(3)

where k0 is the pseudo-zero-order rate constant and khyd represents the apparent first-order rate constant for hydrolysis of the diclofenac prodrug in a PBS solution. Zero-order kinetics will be maintained as long as the dissolution rate of solid prodrug is sufficiently fast to maintain a DPX-1-0011 concentration equal to the equilibrium (saturation) solubility. From the slopes of the rectilinear plots in Figure 4 (n = 3), a k0 value of 0.24 ± 0.01 :g mL−1 h−1 was calculated. By addition of 1.00 mL of an acidic prodrug solution (10 mg mL−1 ) to 2.00 mL 67 mM PBS with 0.05% Tween 80, a DPX1-0011 suspension was generated in situ. The stability of the latter suspension was investigated in a manner similar to that described for the preformed suspension and a zero-order rate constant of 0.27 ± 0.003 :g mL−1 h−1 was derived. Although comparable data were obtained (Fig. 3), direct comparison of the results of the two experiments is not possible as upon termination of the stability study of the in situ-formed suspension, pH was lowered to 7.0 because of the insufficient buffer capacity. As most ester prodrugs undergo specific base catalysis of hydrolysis at neutral and alkaline pH, the slightly enhanced rate of diclofenac formation from the in situ-formed suspension is most likely a reflection of the fact that prodrug solubility inDOI 10.1002/jps.24221

The overall in vivo performance of the prodrug precipitate formed in situ in the joint cavity is expected to be influenced by the character (amorphous or crystalline) and the particle size distribution of the precipitate as well as the rate of precipitation. Preliminary in situ precipitation characteristics were evaluated by microscopy from mixtures containing 2.00 mL of a slightly acidic solution of DPX-1-0011 (10 mg mL−1 ) and 4.00 mL PBS and 80% SF, respectively. Stoppered test tubes were rotated at low speed at 37 ± 0.5◦ C and microscope pictures were taken of withdrawn samples of the unclear mixtures as a function of time. In both media, instantaneous precipitation in the form of oil-like droplets was observed (Fig. 5). After 24 h, solid particles were formed in the two media. Whereas crystalline material was observed in PBS, determination of the state of the particulate matter formed in 80% SF on storage was not possible based on the data obtained. Efforts were not made to elucidate whether the structure in Figure 5e represents a large single crystal or an agglomerate of minor crystals. Despite the qualitative nature of the pictures taken, it appears that the mean particle size generated over time in the SF medium is smaller than that formed in PBS. Precipitation is a complex process comprising (1) nucleation from a supersaturated solution and (2) growth of the initially formed particles.24 It is well known that the presence of polymers in supersaturated solutions may delay nucleation25 and slow the crystal growth process because of mass transfer particle aggregation limitations.26 The difference in particle size between the two in situ-formed prodrug precipitates may therefore be ascribed to the fact that the SF is relatively rich in macromolecules including high-molecular-weight hyaluronic acid.1 As a rough estimate, the size of the DPX-1-0011 particles formed in situ in SF is in the lower micrometer range. Although the physical state of the precipitate formed in situ in SF is not determined, the terms in situ-formed suspension and in situ-formed precipitate will be used interchangeably in the following. By laser light scattering measurements, the particle size characteristics of the preformed DPX-1-0011 suspension in PBS used for the stability and release studies were: d(0.1), d(0.5), and a d(0.9) of 1.5, 109.5, and 195.7 :m, respectively. In an identical in situ precipitation setup, the initial degree of supersaturation of DPX-1-0011 was monitored by determination of the prodrug concentration in the supernatant as a function of time (Fig. 6). Inhibition of nucleation effected by macromolecules in SF may explain the pronounced initial supersaturation observed in the SF as compared with that determined in PBS. Similar high initial prodrug concentrations were observed upon addition of the acidic prodrug solution to PBS containing 2% BSA (w/v). As albumin accounts for approximately 60% of the total protein content in SF,27 the observations performed may imply that this protein is playing a significant role in the prodrug precipitation process. After about 2 h, the DPX-1-0011 concentration approaches a constant level in all three media with the apparent (kinetic) solubility of the prodrug in 80% SF far exceeding the value of 4.5 :g mL−1 obtained for the thermodynamic solubility of DPX-1-0011 in PBS Thing et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:4021–4029, 2014

4026

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Figure 5. Microscope pictures of precipitates formed upon addition of an acidic DPX-1-0011 solution to 80% SF and 67 mM PBS pH 7.4 (2:1, v/v). (a) SF added to a slightly acidic solution without DPX-1-0011 (placebo). (b) and (c) obtained in SF after mixing in 2 min and 24 h, respectively. (d) and (e) obtained in PBS after mixing in 2 min and 24 h, respectively. A magnification of 400× was used.

(Fig. 3). To this end, the kinetic solubility of the prodrug in PBS amounted to 20 :g mL−1 . Addition of BSA to PBS also led to an increase in the apparent solubility of the prodrug. Hence, the enhanced solubility of DPX-1-0011 in SF can most likely be ascribed to prodrug affinity to SF components. This increased solubility is less likely to afford major variation of IA pharmacokinetics because excess of precipitated prodrug in the joint Thing et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:4021–4029, 2014

will result in a fairly constant free fraction of dissolved prodrug that is the target for SF esterases. In Vitro Diclofenac Release in the Rotating Dialysis Cell Previously, the rotating dialysis cell model has been used to emulate drug transport processes taking place in a synovial environment.28 Using this methodology, a preliminary DOI 10.1002/jps.24221

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

4027

Figure 6. In situ precipitation of DPX-1-0011 in 80% (v/v) SF, 67 mM PBS pH 7.4 and 2% (w/v) BSA in PBS at 37◦ C. Bars represent standard deviation (n = 3). The solid line represents the solubility of DPX-1-0011 in 67 mM PBS at pH 7.4. The insert is focusing on the data points between 0 and 100 :g mL−1 .

in vitro–in vivo relationship has been established in the area of IA drug delivery of injectable drug solutions.2,29 In the present study, the appearance of regenerated diclofenac (and intact prodrug) in the acceptor phase was monitored as a function of time over a period of 55 h. Experiments were initiated by adding 2.0 mL of a slightly acidic DPX-1-0011 solution (10 mg mL−1 ) or 2.0 mL of a preformed DPX-1-0011 suspension (10 mg mL−1 ) to the donor cell containing 4.0 mL 80% SF. The resulting release profiles (Fig. 7) are characterized by an initial lag time of diclofenac appearance in the acceptor phase (Table 1) resulting from the high affinity of diclofenac for proteins. In the therapeutic concentration range, about 99.7% of the diclofenac is bound to plasma proteins.30 The shorter lag time observed for the in situ-formed suspension compared with the preformed suspension is most likely a reflection of the fact that much more prodrug in solution is available for enzymatic cleavage as a consequence of the observed high degree of supersaturation (Fig. 7). The pronounced supersaturation may thus constitute a means to achieve a desirable fast onset of diclofenac action. After 24 h (maybe earlier), the amount of diclofenac released per time unit into the acceptor phase becomes constant for both the preformed and in situ-formed suspension. Almost identical apparent pseudo-zero-order transport rate constants were Table 1.

Figure 7. Release profiles of DPX-1-0011 and diclofenac at 37◦ C from in situ-formed and preformed suspensions in SF in the rotating dialysis cell model. To 4 mL 80% (v/v) SF in the donor cell, 2 mL preformed suspension or acidic solution of DPX-1-0011 was added with a total dose of 20 mg DPX-1-0011 added. Bars represent standard deviations (n = 3) for the in situ-formed suspension. Only DPX-1-0011 concentrations above quantification level are shown.

Release Data Obtained from the Release Studies Performed in the Rotating Dialysis Cell Model

Formulation Cdiclofenac in donor cell after 55 h (:g mL−1 ) Total amount of diclofenac and DPX-1-0011 released after 55 h (%) Release rate of diclofenac (:mol h−1 )b Lag time for diclofenac appearance (h)c

Preformed Suspension 55.4 10.8 0.13 15

± ± ± ±

1.0 0.3a 0.003 0.11

In Situ-Formed Suspension 61.4 13.9 0.13 7.7

± ± ± ±

1.3 0.8 0.07 0.95

Mean ± SD, n = 3 for in situ-formed suspension and mean ± error, n = 2 for preformed suspension. a Assuming that 20 mg DPX-1-0011 was added to the donor cell. b On the basis of the appearance of diclofenac in the acceptor chamber from 23.5 to 55 h. c On the basis of the linear regression of the diclofenac appearance from 23.5 to 55 h. DOI 10.1002/jps.24221

Thing et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:4021–4029, 2014

4028

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

calculated from the slopes of the linear plots of cumulated amounts of diclofenac against time (Table 1). In this setup, a prerequisite for obeying zero-order kinetics is that the free diclofenac concentration in the donor compartment remains constant. This free diclofenac concentration (C) might be estimated from the relationship.28 J = PC

(4)

where J represents the steady-state flux (obtained by dividing the zero-order rate constant in Table 1 with the area of the dialysis cell membrane of 21.8 cm2 ) and P is the permeability coefficient (0.29 cm h−1 determined experimentally). Thus, the free diclofenac concentration in donor cell is in both experiments calculated to about 6 :g mL−1 (Table 1). At termination of the experiments after 55 h, the total diclofenac solution concentration in the donor compartment was determined to be approximately 60 :g mL−1 (Table 1). Thus, the obtained data suggest that under the present high total diclofenac concentration in the donor cell, diclofenac protein binding only amounts to about 90%. In comparison to the 60 :g mL−1 found in the donor compartment, the peak joint concentration of diclofenac in man amounts to about 300 ng mL−1 after an oral dose of 75 mg diclofenac.31 The NSAID is usually administered twice daily and the plasma half-life is about 1–2 h.32 Hence, peak joint concentrations of 300 ng mL−1 are only maintained for short periods of time following oral dosing. The relative steep increment of the initial prodrug release profile is afforded by the initial supersaturated prodrug concentrations in the donor compartment generating a sufficient concentration gradient to facilitate passive transport across the membrane. At later sampling times, the prodrug was still detected in the acceptor phase; however, concentrations were below LOQ. As regard release from the preformed suspension, the content of DPX-1-0011 in all samples was below LOQ. It should be noted that sink conditions were not maintained for DPX-1-0011 during the experiments. Results of the in vitro release experiments indicate that the prodrug-based formulation principle can provide a constant diclofenac concentration in an aqueous compartment emulating the knee joint cavity. Animals The local sustained drug release capabilities of different depot formulation types after joint injection have been investigated using various animal models including the rat.1 IA injection volumes in the range of 20–80 :L per 100 g have been used in rats.33,34 Currently, the selection of a feasible injection volume that does not result in major increase in the IA hydraulic pressure, and subsequently the trans-synovial absorption rate,35 is impeded because of the apparent lack of information on the SF volume of rat joints. As the slightly acidic prodrug solution after IA administration has to be neutralized in the joint space to accomplish precipitation of the free base form of the prodrug, the size of the injection volume should be kept as low as possible. In this preliminary study, 30 :L injection volumes of a 10 mg mL−1 solution corresponding to 0.75–1 mg kg−1 were used. Although subject to considerable variation in the amount of prodrug recovered from the joints of the sacrificed rats, the data clearly demonstrated the presence of DPX-1-0011 in the joints of the three animals sacrificed on day 2, 4, 6, and 8, respectively, after IA injection of the slightly acidic preformulaThing et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:4021–4029, 2014

tion solution. Pharmacokinetic analysis was prohibited as the amount of DPX-1-0011 extracted from the joints was low and variable resulting in analyte concentrations close to or slightly above LOQ of the LC–MS/MS methodology. On the contrary, this initial in vivo study shows that an immobile depot is formed that can be detected 8 days following injection of the prodrugbased preformulation into the joint of rats. After sacrifice, all injected rat knees were opened and visually inspected for signs of toxicity. No toxicity signs were observed. Additional experiments are needed for demonstration of efficacy.

CONCLUSIONS The present study has shown that a diclofenac prodrug-based technology is capable of forming an immobile particulate depot in situ following joint injection in the form of a liquid, lowviscous preformulation. The prodrug was detected in rat knees 8 days after injection of the slightly acidic preformulation. The long joint residence time is accomplished by prodrug precipitation in the joint cavity exploiting the difference in prodrug solubility at slightly acidic and neutral pH (about a factor of 1400). The observed difference between the in situ prodrug precipitation processes taking place in PBS and 80% SF is most likely a reflection of the presence of macromolecules in the biological matrix that may delay prodrug nucleation as well as particle growth. Diclofenac exposure to the joint is achieved as the diclofenac regeneration rate from the prodrug is much faster than the rate of prodrug transport out of the joint. Thus, results of the present study strongly suggest that the prodrugbased approach involving in situ precipitation of the injected prodrug in the joint constitutes a promising depot principle for diclofenac as well as for other NSAIDs containing a carboxylic acid group.

ACKNOWLEDGMENTS We thank Dr. Erling B. Jørgensen, Lundbeck Pharma A/S, for the determinations of the pKa values and associate professor Charlotte Gabel-Jensen for help with setting up the LC–MS/MS analyses. Mette A. Thing, Susan W. Larsen, Jesper L. Kristensen, Jesper Østergaard, and Claus Larsen are founders and shareholders in Dep-Xplora ApS seeking to commercialize in situ depot forming prodrug technologies for the management of joint conditions characterized by pain and inflammation.

REFERENCES 1. Larsen C, Østergaard J, Larsen SW, Jensen H, Jacobsen S, Lindegaard C, Andersen PH. 2008. Intra-articular depot formulation principles: Role in the management of postoperative pain and arthritic disorders. J Pharm Sci 97:4622–4654. 2. Frost AB, Larsen F, Østergaard J, Larsen SW, Lindegaard C, Hansen HR, Larsen C. 2010. On the search for in vitro in vivo correlations in the field of intra-articular drug delivery: Administration of sodium diatrizoate to the horse. Eur J Pharm Sci 41:10–15. 3. Caldwell JR. 1996. Intra-articular corticosteroids. Guide to selection and indications for use. Drugs 52:507–514. 4. Hunter JA, Blyth TH. 1999. A risk-benefit assessment of intraarticular corticosteroids in rheumatic disorders. Drug Saf 21:353–365. 5. Gerwin N, Hops C, Lucke A. 2006. Intraarticular drug delivery in osteoarthritis. Adv Drug Deliv Rev 58:226–242. DOI 10.1002/jps.24221

RESEARCH ARTICLE – Pharmaceutics, Drug Delivery and Pharmaceutical Technology

6. Forrest JB, Camu F, Greer IA, Kehlet H, Abdalla M, Bonnet F, Ebrahim S, Escolar G, Jage J, Pocock S, Velo G, Langman MJ, Bianchi PG, Samama MM, Heitlinger E. 2002. Ketorolac, diclofenac, and ketoprofen are equally safe for pain relief after major surgery. Br J Anaesth 88:227–233. 7. Anthony D, Jasinski DM. 2002. Postoperative pain management: Morphine versus ketorolac. J Perianesth Nurs 17:30–42. 8. Torres-Lopez JE, Lopez-Munoz FJ, Castaneda-Hernandez G, Flores-Murrieta FJ, Granados-Soto V. 1997. Pharmacokinetic– pharmacodynamic modeling of the antinociceptive effect of diclofenac in the rat. J Pharmacol Exp Ther 282:685–690. 9. Aly MN. 2008. Intra-articular drug delivery: A fast growing approach. Recent Pat Drug Deliv Formul 2:231–237. 10. Rømsing J, Møiniche S, Østergaard D, Dahl JB. 2000. Local infiltration with NSAIDs for postoperative analgesia: Evidence for a peripheral analgesic action. Acta Anaesthesiol Scand 44:672–683. 11. Derendorf H, Mollmann H, Gruner A, Haack D, Gyselby G. 1986. Pharmacokinetics and pharmacodynamics of glucocorticoid suspensions after intraarticular administration. Clin Pharmacol Ther 39:313– 317. 12. Derendorf H, Mollmann H, Voortman G, van den Ouweland FA, van de Putte LB, Gevers G, Dequeker J, van Vliet-Daskalopoulou E. 1990. Pharmacokinetics of rimexolone after intra-articular administration. J Clin Pharmacol 30:476–479. 13. Akers MJ, Fites AL, Robison RL. 1987. Formulation design and development of parenteral suspensions. J Parenter Sci Technol 41:88– 96. 14. Hiestand EN. 1964. Theory of coarse suspension formulation. J Pharm Sci 53:1–18. 15. He C, Kim SW, Lee DS. 2008. In situ gelling stimuli-sensitive block copolymer hydrogels for drug delivery. J Control Release 127:189–207. 16. Packhaeuser CB, Schnieders J, Oster CG, Kissel T. 2004. In situ forming parenteral drug delivery systems: An overview. Eur J Pharm Biopharm 58:445–455. 17. Larsen C, Larsen SW, Thing MA, Kristensen JL, Jensen H, Østergaard J. 2012. Prodrugs of non-steroid anti-inflammatory agents (NSAIDs). Patent WO 2013/064153 A1. 18. Lawson JK. 1953. 2-Vinylimidazole and 1-methyl-2-vinylimidazole. J Am Chem Soc 75:3398–3400. ´ ´ K, Box KJ, Avdeef A. 1997. Potentiometric pKa 19. Takacs-Nov ak determination of water-insoluble compounds: Validation study in methanol/water mixtures. Int J Pharm 151:235–248. 20. Larsen SW, Østergaard J, Friberg-Johansen H, Jessen MN, Larsen C. 2006. In vitro assessment of drug release rates from oil depot formulations intended for intra-articular administration. Eur J Pharm Sci 29:348–354.

DOI 10.1002/jps.24221

4029

21. Geibel J, Cannon J, Campbell D, Traylor TG. 1978. Model compounds for R-state and T-state hemoglobins. J Am Chem Soc 100:3575– 3585. 22. Dansk Lægemiddel Information. 2014. Morfin “DAK”. Accessed 10 July, 2014, at: http://pro.medicin.dk/Medicin/Praeparater/993. 23. Simkin PA, Bassett JE. 2011. Pathways of microvascular permeability in the synovium of normal and diseased human knees. J Rheumatol 38:2635–2642. 24. Arnold YE, Imanidis G, Kuentz MT. 2011. Advancing in-vitro drug precipitation testing: New process monitoring tools and a kinetic nucleation and growth model. J Pharm Pharmacol 63:333–341. 25. Xu S, Dai WG. 2013. Drug precipitation inhibitors in supersaturable formulations. Int J Pharm 453:36–43. 26. Dirksen JA, Ring TA. 1991. Fundamentals of crystallization - Kinetic effects on particle-size distributions and morphology. Chem Eng Sci 46:2389–2427. 27. Cajori FA, Pemberton R. 1928. The chemical composition of synovial fluid in cases of joint effusion. J Biol Chem 76:471–480. 28. Østergaard J, Larsen SW, Parshad, H, Larsen C. 2005. Bupivacaine salts of diflunisal and other aromatic hydroxycarboxylic acids: Aqueous solubility and release characteristics from solutions and suspensions using a rotating dialysis cell model. Eur J Pharm Sci 26:280–287. 29. Frost AB, Lindegaard C, Larsen F, Østergaard J, Larsen SW, Larsen C. 2011. Intra-articular injection of morphine to the horse: Establishment of an in vitro-in vivo relationship. Drug Dev Ind Pharm 37:1043–1048. 30. Riess W, Stierlin H, Degen P, Faigle JW, Gerardin A, Moppert J, Sallmann A, Schmid K, Schweizer A, Sulc M, Theobald W, Wagner J. 1978. Pharmacokinetics and metabolism of the anti-inflammatory agent Voltaren. Scand J Rheumatol Suppl 7:17–29. 31. Fowler PD, Shadforth MF, Crook PR, John VA. 1983. Plasma and synovial fluid concentrations of diclofenac sodium and its major hydroxylated metabolites during long term treatment of rheumatoid arthritis. Eur J Clin Pharm 25:389–394. 32. Willis JV, Kendall MJ, Flinn RM, Thornhill DP, Welling PG. 1979. The pharmacokinetics of diclofenac sodium following intravenous and oral administration. Eur J Clin Pharmacol 16:405–410. 33. Mountziaris PM, Sing DC, Mikos AG, Kramer PR. 2010. Intraarticular microparticles for drug delivery to the TMJ. J Dent Res 89:1039–1044. 34. Zhang Z, Bi X, Li H, Huang G. 2011. Enhanced targeting efficiency of PLGA microspheres loaded with lornoxicam for intra-articular administration. Drug Deliv 18:536–544. 35. McDonald JN, Levick JR. 1993. Effect of extravascular plasma protein on pressure-flow relations across synovium in anaesthetized rabbits. J Physiol 465:539–559.

Thing et al., JOURNAL OF PHARMACEUTICAL SCIENCES 103:4021–4029, 2014