Long-Acting Diclofenac Ester Prodrugs for Joint Injection: Kinetics, Mechanism of Degradation, and In Vitro Release From Prodrug Suspension

Journal of Pharmaceutical Sciences xxx (2016) 1-9

Contents lists available at ScienceDirect

Journal of Pharmaceutical Sciences journal homepage: www.jpharmsci.org

Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Long-Acting Diclofenac Ester Prodrugs for Joint Injection: Kinetics, Mechanism of Degradation, and In Vitro Release From Prodrug Suspension Nina Mertz 1, Susan Weng Larsen 1, Jesper Kristensen 2, Jesper Østergaard 1, Claus Larsen 1, 3, * 1 2 3

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 Dep-Xplora ApS, Gammelbyvej 17, Lejre DK-4320, Denmark

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 March 2016 Revised 26 May 2016 Accepted 10 June 2016

A prodrug approach for local and sustained diclofenac action after injection into joints based on ester prodrugs having a pH-dependent solubility is presented. Inherent ester prodrug properties influencing the duration of action include their pH-dependent solubility and charge state, as well as susceptibility to undergo esterase facilitated hydrolysis. In this study, physicochemical properties and pH rate profiles of 3 diclofenac ester prodrugs differing with respect to the spacer carbon chain length between the drug and the imidazole-based promoiety were determined and a rate equation for prodrug degradation in aqueous solution in the pH range 1-10 was derived. In the pH range 6-10, the prodrugs were subject to parallel degradation to yield diclofenac and an indolinone derivative. The prodrug degradation was found to be about 6-fold faster in 80% (vol/vol) human plasma as compared to 80% (vol/vol) human synovial fluid with 2-(1-methyl-1H-imidazol-2-yl)ethyl 2-(2-(2,6 dichlorophenyl)amino)phenylacetate being the poorest substrate toward enzymatic cleavage. The conversion and release of parent diclofenac from prodrug suspensions in vitro were studied using the rotating dialysis model. The results suggest that it is possible to alter and control dissolution and reconversion behavior of the diclofenac prodrugs, thus making the prodrug approach feasible for local and sustained diclofenac action after joint injection. © 2016 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

Keywords: chemical stability drug delivery systems injectables joint injections non-steroidal anti-inflammatory drugs pH rate profile prodrug

Introduction Sustained and localized drug action after joint injection may constitute an optimal treatment option in pain relief after minor arthroscopic surgery, sports injuries, and flare up episodes in osteoarthritis.1 Advantages of low-dose intra-articular depot injectables include the achievement of sustained high drug concentrations within the joint combined with minimum systemic drug exposure, and various formulation strategies to achieve sustained drug action upon joint injection have been investigated.2-4

Conflicts of interest: 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. This article contains supplementary material available from the authors by request or via the Internet at http://dx.doi.org/10.1016/j.xphs.2016.06.013. * Correspondence to: Claus Larsen (Telephone: þ45-35336466). E-mail address: [email protected] (C. Larsen).

Currently, the only depot type for joint injection is aqueous microcrystalline corticosteroid ester injectables, for example, methylprednisolone acetate (Depo-Medrol®), which is a common practice in the relief of pain and inflammation associated arthritis conditions.5-7 Thus, intra-articular administration of long-acting injectables of non-steroidal anti-inflammatory drugs (NSAIDs) appears promising for the relief of joint associated pain because this class of drugs possesses anti-inflammatory as well as antinociceptive properties.8,9 Despite simplicity, aqueous suspensionbased injectables may suffer from manufacturing challenges and poor physical stability upon storage.10 Recently, a novel prodrug approach for the accomplishment of local and sustained diclofenac action after injection into joints was reported.4 Diclofenac is among the most potent and used nonsteroidal anti-inflammatory drugs. This approach involved the employment of pro-moieties that besides having an OH group (for diclofenac ester bond formation) comprised a weak base functionality with a pKa value of about 6-8.11 Hence, this prodrug type experienced pH-dependent solubility, that is, a desired low

http://dx.doi.org/10.1016/j.xphs.2016.06.013 0022-3549/© 2016 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

2

N. Mertz et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-9

solubility at physiological pH 7.4 in combination with a significantly higher solubility in slightly acidic solution. It was demonstrated that after injection of a concentrated slightly acidic prodrug solution (the preformulation) into human synovial fluid (SF) pH 7.4, in situ precipitation of the neutral prodrug species in this biological matrix occurred.4 Only the small amount of dissolved prodrug, in equilibrium with the precipitate, was subject to SF enzymemediated cleavage. Accordingly, it is to be expected that both onset and duration of diclofenac action can be modified by variation of inherent ester prodrug properties including their pHdependent solubility and charge, as well as their susceptibility to undergo esterase facilitated hydrolysis. From a formulation point of view, the pH-dependent stability of the prodrugs in aqueous solution will dictate whether the diclofenac ester prodrug preformulation can be marketed in the form of an aqueous injectable or needs to be manufactured as a dry powder for reconstitution prior to use. In this study, 3 diclofenac ester prodrugs differing with respect to the spacer carbon chain length (Fig. 1) were synthesized and evaluated in vitro. Thus, the objectives of this study were (i) to determine the effect of the carbon chain length on the pKa value and aqueous solubility of the prodrugs; (ii) to investigate the kinetics and mechanism of degradation of the 3 prodrugs in aqueous solution in the pH range 1-10, as well as in 80% (vol/vol) SF and 80% (vol/vol) human plasma at 37 C, and (iii) to characterize in vitro release of diclofenac from prodrug suspensions using the rotating dialysis cell model. Experimental Materials 2-(1-methyl-1H-imidazol-2-yl)ethanol, 4-dimethylaminopyri dine, dichloromethane, dicyclohexylcarbodiimide, methanol, and acetonitrile were purchased from Sigma-Aldrich (Munich, Germany). Diclofenac >98% was obtained from TCI (Tokyo, Japan). All other chemicals were of analytical grade or highest quality possible. Demineralized water was used throughout the study. Outdated human plasma was obtained from Rigshospitalet (Copenhagen, Denmark). SF samples from 10 arthritis patients were obtained from the Parker Institute, Frederiksberg Hospital (Copenhagen, Denmark) and were pooled and frozen in smaller portions for later use. Visking dialysis tubing size 27/3200 , 21.5 mm with a cutoff at 12,000-14,000 Da (VWR International, West Chester, PA) was employed for the in vitro release studies.

Calorimeter (DSC) 7 (Perkin Elmer, Norwalk, CT) controlled by Pyris software (version 7.0). NMR spectra were recorded on a Bruker Avance III spectrometer (Bruker, MA) and processed using MestreNova software. Tetramethylsilane was used as internal reference for 1 H-NMR spectra recorded in CDCl3. Solvent residual peaks3 were used as internal reference for all 13C-NMR spectra and for 1H-NMR spectra recorded in DMSO-d6, CD3OD, and C6D6. Thin layer chromatography was performed on Merck aluminum sheets pre-coated with silica gel F254. Compounds were visualized by ultraviolet or by heating after dipping in Cemol (6.25 g NH4Mo2O7 and 2.5 g Ce(SO4)2 in 250 mL 10% H2SO4), anisaldehyde (9.2 mL p-anisaldehyde, 3.75 mL acetic acid and 12.5 mL H2SO4 in 338 mL ethanol), or ninhydrin (1.5 g ninhydrin, 100 mL n-butanol, 3.0 mL acetic acid). Synthesis of (1-Methyl-1H-Imidazol-2-yl)Methyl 2-(2-(2,6 Dichlorophenyl)Amino)Phenyl)Acetate 2-(chloromethyl)-1-methyl-1H-imidazole hydrochloride11 (0.55 g, 3.3 mmol), diclofenac free acid (0.89 g, 3.0 mmol), and Cs2CO3 (2.25 g, 6.9 mmol) were suspended in N,N-dimethylformamide (15 mL) and stirred overnight under inert atmosphere (Scheme S2 in Supplementary Information). The reaction mixture was poured into water (200 mL) and extracted with CH2Cl2 (4  50 mL). The combined organic extracts were washed with dilute aqueous NH3 (3  100 mL) and water (100 mL) and then dried over Na2SO4, filtered and concentrated under reduced pressure. The residue was purified by flash chromatography (0%-2% 2 M methanolic NH3 in ethyl acetate) to give a colorless solid which was recrystallized from methyl tert-butyl ether/ heptane to afford the title compound (920 mg, 79% yield) as colorless crystals (see Scheme S1 in Supplementary Information). 1 H NMR (free base, 600 MHz, CDCl3) d 7.26 (d, J ¼ 8.1 Hz, 1H), 7.14 (dd, J ¼ 7.5, 1.1 Hz, 1H), 7.05 (td, J ¼ 7.9, 1.5 Hz, 1H), 6.95 (d, J ¼ 1.0 Hz, 1H), 6.91 (t, J ¼ 8.1 Hz, 1H), 6.88 (t, J ¼ 7.4 Hz, 1H), 6.80 (d, J ¼ 0.9 Hz, 1H), 6.67 (br s, 1H), 6.47 (d, J ¼ 8.0 Hz, 1H), 5.17 (s, 2H), 3.79 (s, 2H), 3.47 (s, 3H). 13C NMR (free base, 151 MHz, CDCl3) d 172.0, 142.8 (2C), 142.2, 137.9, 131.0, 129.6, 129.0 (2C), 128.5, 128.3, 124.2, 124.2, 122.6, 122.3, 118.6, 58.7, 38.4, 33.0. Mp (DSC) 123.3 C. Synthesis of 2-(1-Methyl-1H-Imidazol-2-yl)Ethyl 2-(2-(2,6 Dichlorophenyl)Amino)PhenylAcetate 2-(1-methyl-1H-imidazol-2-yl)ethyl 2-(2-(2,6 dichlorophenyl) amino)phenylacetate (DP-2) was prepared from 2-(1-methyl-1Himidazol-2-yl)ethanol11 in 67% yield as colorless crystals as previously described (see Scheme S2 in Supplementary Information).4

Synthesis of Diclofenac Esters Ethyl ether (Et2O) was dried over sodium wire. All other dry solvents were dried over 3 Å molecular sieves. The chemical structures of the synthesized prodrugs were confirmed by nuclear magnetic resonance (NMR) analyses and their melting points were determined as the onset of melting using a Perkin Elmer Differential Scanning

Synthesis of 3-(1-Methyl-1H-Imidazol-2-yl)propyl 2-(2-(2,6 Dichlorophenyl)Amino)PhenylAcetate Diclofenac free acid (2.96 g, 10 mmol), 3-(1-methyl-1H-imidazol2-yl)propan-1-ol11 (10 mmol), and 2-(1-methyl-1H-imidazol-2-yl) ethanol, 4-dimethylaminopyridine (122 mg, 1 mmol) were

Figure 1. Chemical structures of the diclofenac ester prodrugs.

N. Mertz et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-9

dissolved in CH2Cl2 (25 mL) and cooled to 0 C in an ice bath. A solution of dicyclohexylcarbodiimide (4.12 g, 20 mmol) in CH2Cl2 (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 was evaporated to give a pale yellow oil. The oil was purified by flash chromatography performed on silica gel 60 (35-70 mm) using 0%-10% 2 M methanolic NH3 in ethyl acetate to give 3-(1-methyl1H-imidazol-2-yl)propyl 2-(2-(2,6 dichlorophenyl)amino)phenylacetate (DP-3) (Scheme S1 in Supplementary Information) in 69% yield as a colorless oil. The free base was dissolved in Et2O (100 mL) and added dropwise to a stirred solution of HNO3 (5 mmol in Et2O [50 mL]). The resulting suspension was stored overnight at 18 C and then filtered to give the nitrate salt as white crystals (see Scheme S2 in Supplementary Information). 1 H NMR (nitrate salt, 600 MHz, DMSO-d6) d 14.03 (br s, 1H), 7.60 (d, J ¼ 2.0 Hz, 1H), 7.58 (d, J ¼ 2.0 Hz, 1H), 7.53 (d, J ¼ 8.1 Hz, 2H), 7.21 (t, J ¼ 8.1 Hz, 1H), 7.19 (dd, J ¼ 7.5, 1.4 Hz, 1H), 7.07 (dt, 7.8, 1.5 Hz, 1H), 7.05 (br s, 1H), 6.86 (td, J ¼ 7.4, 1.1 Hz, 1H), 6.26 (d, J ¼ 8.0 Hz, 1H), 4.14 (t, J ¼ 6.2 Hz, 2H), 3.78 (s, 2H), 3.72 (s, 3H), 3.01 (t, J ¼ 7.6 Hz, 2H), 2.09 e 2.01 (m, 2H). 13C NMR (nitrate salt, 151 MHz, DMSOd6) d 171.4, 146.5, 142.8 (2C), 137.0, 131.0, 130.7, 129.2 (2C), 127.81, 126.0, 123.3, 123.0, 120.6, 118.1, 115.8, 63.3, 37.0, 33.9, 25.1, 20.9. Mp (DSC) 116.1 C.

Kinetic Studies The hydrolytic stability of the prodrugs in the pH range 1-10 was investigated in aqueous buffer solutions at 37 ± 0.5 C. The following buffers were used: at pH 2: hydrochloric acid solutions; at pH 3 and pH 6.5-7.4: phosphate (pKa 2.2 and 7.212); at pH 4-5.8: acetate (pKa 4.7613); pH 8-9.5: borate (pKa 9.112); and at pH 10: carbonate (pKa 10.2412). A constant ionic strength of 0.3 M was

3

maintained for all buffers by adding an appropriate amount of potassium chloride. The ionic strength I was calculated from the following equation:

I¼

1X c $z2 2 i i i

(1)

where ci is the concentration and zi is the charge of the ith ion of the electrolyte, respectively. For the slowly proceeding degradation reactions at 2  pH  5.8, the initial rate method14 was used to determine the pseudo-firstrate constants (kobs) from prodrug solutions of about 1  104 M. In strongly acidic solution below pH 2, diclofenac undergoes an intra-molecular cyclization reaction to form an indolinone derivative.15 Thus, at low pH the degradation pattern of the diclofenac ester prodrugs involves formation of diclofenac and the indolinone derivative in a consecutive manner (Fig. 2). Accordingly, the pseudo-first-order rate constant for the degradation of prodrug to diclofenac (kC1) at pH 1-1.5 was determined from the general expression:

½Diclofenact ¼

 kC1 ½prodrug0  kC1 t $ e  ekC2 t kC2  kC1

(2)

where [prodrug]0 refers to the initial prodrug concentration, [diclofenac]t is the diclofenac concentration at time t, and kC2 refers to the pH-dependent pseudo-first-order rate constant for the formation of the indolinone derivative after incubation of diclofenac in acidic environment. For the faster hydrolysis reactions above pH 5.8, the kobs values were determined by monitoring the decrease in the concentration of intact prodrug as a function of time employing initial prodrug concentrations of about 1  105 M. The pH values of test solutions were measured at 37 C on initiation and at termination of the experiments using

Figure 2. Reaction schemes for formation of diclofenac and an indolinone (1-(2,6-dichlorophenyl)-1,3-dihydro-2H-indol-2-one) from degradation of the DP-2 prodrug at pH below 2 (a) and at 5.8  pH  10 (b) and 37 C.

4

N. Mertz et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-9

a 744 pH meter (Metrohm, Riverview, FL). The pH of the hydrochloric acid solutions were calculated from Equation 3

pH ¼ logðgHþ Þ  log

h

Hþ

i

(3)

where gHþ represents the hydrogen ion activity coefficient at 37 C and I ¼ 0.3 M (0.744).16 For each pH value in the pH range 3-10, first-order rate constants at 3-4 buffer concentrations (range 0.04-0.1 M) were determined. The reaction solutions were applied to high performance liquid chromatography (HPLC) vials and kept in an autosampler equipped with a Merck Peltier Sample Cooler (VWR International, Tokyo, Japan) capable of maintaining constant autosampler temperature of 37 ± 0.5 C throughout the experiments. Graph fitting was performed using GraphPad Prism version 6 for Windows (Graph Pad Software, San Diego, CA). Stability Studies in Human Synovial Fluid and Plasma Solutions of 80% (vol/vol) human plasma and SF were 4:1 (vol/vol) mixtures of the respective biological matrix and 290 mM HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) buffer (pH 7.4). The degradation of the 3 prodrugs at 37 C was determined by spiking 20 mL of freshly prepared prodrug stock solution in methanol into 4.0 mL 80% (vol/vol) preheated human plasma or SF to give final concentrations of approximately 1  105 M. The reaction solutions were kept at constant temperature using a ProBlot™ oven equipped with a rotisserie (Labnet, Edison, NJ). Aliquots of 200 mL were taken and added to twice the volume of acetonitrile. The mixtures were vortexed and after centrifugation (5 min, 6708 g) in an Eppendorf MiniSpin® plus centrifuge (VWR International, Albertslund, Denmark) the supernatant was analyzed by HPLC. The stability was assessed by monitoring the decrease in intact prodrug as a function of time. Analysis A previously validated HPLC method was used with minor modifications.4 Samples from the degradation and release studies were analyzed using an LaChrom HPLC system (VWR International, Tokyo, Japan) consisting of a Merck-Hitachi L7100 pump connected to a Merck-Hitachi L-7400 diode array detector and a Merck-Hitachi L-7200 autosampler equipped with a Merck Peltier Sample Cooler for L-7200. Reversedphased chromatography was performed using a SymmetryShield™ C18 column (150  4.60 mm2, 5 mm particles; Waters, Hedehusene, Denmark) equipped with a Gemini C18 precolumn (4  3.0 mm2; Phenomenex, Allerød, Denmark). A constant column temperature was maintained at 30 C by a MerckHitachi L-7300 column oven. The mobile phase consisted of methanol and 20 mM acetate buffer (pH 4.94) in the ratio 67.5:32.5 (vol/vol). The injection volume was 20 mL, the flow rate was set at 1 mL/min, and the column effluent was monitored at 274 nm. Retention times were 4.0 (diclofenac), 5.0 (indolinone), 7.9 ((1-methyl-1H-imidazol-2-yl)methyl 2-(2-(2,6 dichlorophenyl)amino)phenylacetate [DP-1]), 6.6 (DP-2), and 7.0 min (DP-3). Calibration curves were constructed by diluting stock solutions (in acetonitrile) of diclofenac, 1-(2,6dichlorophenyl)-2-indolinone, or prodrug into aqueous buffer solutions. Further dilutions were made in buffer solutions (kinetic studies) or in a 1:2 vol/vol mixture of 67 mM phosphate buffer solution (PBS) (pH 7.4) and acetonitrile (studies of stability in 80% (vol/vol) plasma and 80% (vol/vol) SF).

Potentiometric Determination of pKa pKa values of DP-1, DP-2, and DP-3 were determined at 25 ± 1 C using the Sirius GLpKa titrator (Sirius Analytical Instruments Ltd, East Sussex, UK) equipped with a pH electrode. The titrations were performed in the pH range 2.5-10 at an ionic strength of 0.15 M. The co-solvent dissociation constants (pKa values) were determined at methanol contents of 29.2%, 39.1%, and 49.5% (vol/vol). pKa values at zero methanol concentration were achieved using the Yasuda-Shedlovsky extrapolation method.17 Solubility Experiments The determination of the solubilities of DP-1, DP-2, and DP-3 at pH 7.4 and 37 C was carried out as previously described for DP-2.4 Release Studies The release experiments (n ¼ 3) were carried out in the rotating dialysis cell model at 37 ± 0.5 C as previously reported.18 In short, to the dialysis cell containing 4.0 mL 80% SF (vol/vol) in 290 mM HEPES buffer solution pH 7.4 1.0 mL preformed prodrug suspension (10 mg/mL of DP-1 or DP-2) was added. The solid particles were dispersed in 67 mM PBS (pH 7.4) containing 0.01% (vol/vol) Tween® 80 and left in an ultrasonic bath for 15-30 min. Mean particle size of DP-1 and DP-2 was determined by laser diffraction with Malvern Mastersizer 2000 particle size analyzer (Malvern Instruments Ltd., Worcestershire, UK).4 In case of DP-3, 10 mg solid prodrug followed by 1 mL 0.01% Tween 80 in 67 mM PBS (pH 7.4) was added to the above-mentioned release medium. At time zero the dialysis cell was placed inside a round-bottomed vessel containing 1000 mL acceptor phase consisting of preheated 67 mM PBS (pH 7.4). The revolution speed of the dialysis cell was set to 50 rpm. Samples from the acceptor phase were withdrawn at appropriate time intervals and analyzed by HPLC. The cumulated amount of diclofenac, MA,t, released into the acceptor compartment was calculated from the following equation:

MA;t ¼ VS

n X

Ci1 þ VA Cn

i¼1

where VS and VA are the volume 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. Results and Discussion Chemical structures of the synthesized diclofenac ester prodrugs are depicted in Figure 1. The synthesis processes were not optimized and yields of the diclofenac ester derivatives were in the range 67%-80%. Prodrug structures and crystallinity were confirmed by NMR and X-ray powder diffraction according to Thing et al.4 Purity of the synthesized compounds exceeded 98% as estimated from peak area measurements (HPLC). The prodrug derivatives are weak bases. Their solubilities (at pH 7.4) and pKa values (determined by potentiometric titration) are presented in Table 1. Synthesis and physicochemical properties of DP-2 were previously reported.4 Diclofenac prodrugs with pKa values of about 6-8 were selected to combine a relatively high prodrug solubility at slightly acidic pH (the protonated species) with a

N. Mertz et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-9

5

Table 1 Solubility, pKa, Second-Order Rate Constants, and pHmin for the Degradation of DP-1, DP-2, and DP-3 in Aqueous Solution at 37 C Variable

DP-1 DP-2 DP-3 a b

Solubility pH 7.4 (mg/mL)

0.84 ± 0.03 4.3 ± 0.4 4.6 ± 1.7

pKa 25 Ca

5.72 7.66 7.43

pKa 37 Cb

6.04 7.31 7.64

pHmin

kHþ

kH2 O

kOH

k0 OH

0.119 0.139 0.239

3.06  104 1.70  105 1.05  105

3.76  106 5.86  105 1.38  105

1.03  105 3.14  104 2.52  104

3.19 3.63 4.06

Determined by potentiometric titration. Estimated by fitting of the experimental data to Equation 12 upon extrapolation to zero buffer concentration.

lower prodrug solubility at physiological pH 7.4 due to the presence of the neutral species. Degradation Kinetics of Diclofenac Ester Prodrugs The degradation of the prodrugs was studied over the pH range 1-10 in aqueous buffer solutions at 37 C. At fixed pH and an ionic strength of 0.3 M, the overall degradation reactions obeyed pseudofirst-order kinetics when measuring the concentration of intact prodrug as function of time over 1-2 half-lives. For the slowly proceeding hydrolysis reactions at 2  pH  5.8, the pseudo-firstorder rate constants were obtained using the initial rate method, that is, by measurement of the initial formation of diclofenac (corresponding to 1%-2% of the initial prodrug concentration).14 At pH 1-1.5, however, the initial rate method could not be used because diclofenac at low pH spontaneously cyclizes to form an indolinone derivative.15 Accordingly, at low pH the degradation pattern of the diclofenac ester prodrugs involves formation of diclofenac and the indolinone derivative in a consecutive manner (Fig. 2a). Instead, the pseudo-first-order rate constant for the degradation of prodrug to diclofenac (kC1) at pH 1-1.5 was determined from the general expression valid for consecutive reactions (Eq. 2) where the pseudo-first-order rate constant for the diclofenac cyclization to the indolinone derivative, kC2, at pH 1.12 (kC2 ¼ 0.089 h1) and 1.50 (kC2 ¼ 0.045 h1) were obtained from separate diclofenac cyclization experiments. For the faster degradation reactions above pH 5.8, the kobs values were determined by monitoring the concentration of intact prodrug as a function of time. Various buffer substances were used to maintain constant pH of the reaction solutions. It was observed that the degradation of the prodrugs, to various extents, was subject to general acid-base catalysis by the employed buffer species. Hence, the kobs values at zero buffer concentration, used for construction of the pH rate profiles, were obtained from the intercepts of linear plots of kobs against the total buffer concentration at 3-4 buffer concentrations in the range 0.04-0.1 M. The influence of pH on the rates of degradation of DP-1, DP-2, and DP-3 at 37 C and ionic strength 0.3 M over the pH range 1-10 is presented in Figure 3 in which the logarithm of the buffer-independent pseudo-first-order rate constants kobs are plotted against pH. All the prodrugs exhibited pKa values at about 7 (Table 1). Thus, at acidic to slightly acidic pH the prodrugs exist predominantly on their cationic forms (DP-Hþ), whereas the uncharged species (DP) is dominating at alkaline pH. Accordingly, the shape of the pH rate profiles suggest that in the pH range 1-10 four catalytic reactions contribute to the overall rate of degradation of the individual prodrug:

1 2 3 4

Second-Order Rate Constants (M1$h1)

Reactions

Second-Order Rate Constants

DP-Hþ þ Hþ / Diclofenac DP-Hþ þ H2O / Diclofenac DP-Hþ þ OH / Diclofenac DP þ OH / Diclofenac

kH þ kH2 O kOH k0OH

It should be stressed that the proposed Reactions 2 and 3 are kinetically equivalent19 to the reactions involving a hydrogen ion-catalyzed and a water-catalyzed degradation of the neutral prodrug species (DP), respectively. Under pseudo-first-order degradation kinetics, the overall rate of degradation of prodrug, v, is given by

d½DPT ¼ kobs $½DPT v¼ dt

(4)

where kobs is the pH-dependent pseudo-first-order rate constant and ½DPT refers to the total prodrug concentration according to

i h ½DPT ¼ DP  Hþ þ ½DP

(5)

At pH <2, a linear relationships between log kobs and pH with slope close to 1 were observed for each of the 3 prodrugs strongly suggesting that in this pH region the dominating degradation

Figure 3. pH rate profiles for the degradation of the diclofenac prodrugs in aqueous solution at 37 C (I ¼ 0.3 M). kobs refers to the estimated pseudo-first-order rate constant at zero buffer concentration in the pH range 2-10 and in dilute HCl, pH 1-1.5. The full lines are obtained by fitting the data to Equation 12 using a Kw of 1.297  1014 (37 C, I ¼ 0.3 M).16

6

N. Mertz et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-9

reaction contributing to kobs is a hydrogen ion-catalyzed reaction of the cationic species (Reaction 1):

i h kobs $½DPT ¼ kHþ $aHþ $ DP  Hþ

(6)

where kHþ is the second-order rate constant for specific acid catalysis of degradation of the cationic prodrug species and aHþ refers to the hydrogen activity. In the pH ranges of about 2-4.0 (DP-1), 2.6-3.6 (DP-2), and 3.8-4.3 (DP-3), the kobs values were relatively insensitive to changes in pH indicating that the major contributor to kobs was a water-catalyzed or a spontaneous degradation of the cationic species (Reaction 2). Hence, in the pH range 1 to about 4 the overall rate of degradation can be expressed as follows:

i i h h kobs $½DPT ¼ kHþ $aHþ $ DP  Hþ þ kH2 O $½H2 O$ DP  H þ

(7)



or

pHmin ¼ ½$pKW þ ½$log

  kobs ¼ kHþ aHþ þ kH2 O $½H2 O fDPHþ

(8)

where kH2 O is the second-order rate constant for the watercatalyzed degradation of the prodrugs and fDPHþ is the fraction of the cationic prodrug form is given by

fDPHþ ¼

aHþ aHþ þ Ka

(9)

where Ka is the ionization constant for the prodrug in question. At pH >5, the pH rate profiles consisted of 2 line segments with inflection points observed in the pH ranges 5.8-7.9 (DP-1), 7.2-8.8 (DP-2), and 7.1-8.9 (DP-3) (Fig. 3). The line segments below and above the inflection points exhibited slopes close to unity suggesting the involvement of Reactions 3 and 4, that is, hydroxide ion-catalyzed degradation of the cationic and unionized prodrug, respectively. Hence, the kinetic data obtained can be described as follows:

kobs ¼ kOH $aOH $fDPHþ þ k0OH $aOH $fDP

(10)

where kOH and k0 OH are the second-order rate constants for the specific base catalysis of the cationic and unionized prodrug, respectively. The fraction of uncharged prodrug is calculated from

fDP ¼

The pKa value estimated for DP-1 was significantly lower than those of DP-2 and DP-3 in accordance with the fact that the inductive effect of the electronegative diclofenac ester function on the imidazole nitrogen atom is diminished when the carbon chain spacer arm between the ester group and the imidazole moiety comprises 2 or more methylene groups.20 This difference in pKa values is in agreement with the observation that the inflection point of DP-1 was found at a lower pH value compared to those of the 2 other prodrugs. Despite differences in methodology and experimental conditions, it appears that the fitted pKa values were in fine agreement with those determined by potentiometric titration (Table 1) and those of imidazole derivatives reported in the literature,21,22 thus further supporting the adequacy of Equation 12. In solution, all the prodrugs were most stable in the pH range 3-4 where the prodrugs predominantly exist on the cationic form. Thus, the pH values where the prodrugs are most stable can be derived using the following expression23:

Ka aHþ þ Ka

(11)

Substituting aOH with aKwþ and inserting Equation 8 into H Equation 7 and Equations 9 and 11 into Equation 10 lead to Equation 12, that is, the expression for the overall apparent firstorder rate constant for the degradation of the individual diclofenac ester prodrug in the pH range 1-10:

 kobs ¼

kHþ $ aHþ þ kH2 O $½H2 O þ kOH $ Kw þ k0OH $ $f aHþ DPX

kHþ kOH

 (13)

The calculated pHmin values are presented in Table 1. At pH 4 and 37 C, times for 5% degradation of the prodrug solutions can be calculated to 2.7 (DP-1), 20 (DP-2), and 41 days (DP-3). From these stability data it is most likely that injectable formulations of the diclofenac ester prodrugs have to be in the form of dry powders for reconstitution prior to use. Mechanism of Degradation of the Diclofenac Ester Prodrugs in Aqueous Solution (pH 1-10) The HPLC procedure used for the diclofenac prodrug degradation experiments was capable of monitoring the decrease in intact prodrug along with the formation of the major degradation products (diclofenac and an indolinone structure) as a function of time. At pH 5, the prodrugs underwent hydrolysis to give diclofenac and the promoiety. At pH below 2, however, the emergence of an additional degradation product was observed after an initial lag time. Formation of this degradation product was expected because it was previously reported that diclofenac in strongly acidic solution undergoes an intra-molecular cyclization reaction to yield an indolinone derivative.15 Confirmation of the chemical structure of the second degradation product was done by HPLC analysis of a reference sample of the indolinone derivative 1-(2,6-dichl orophenyl)-2-indolinone (ILO) synthesized according to Larsen and Bundgaard.15 Thus, the experimental data suggested that at

 Kw $fDPXHþ aHþ (12)

In Figure 3, the solid lines represent the theoretical curves obtained by fitting the experimental data according to Equation 12 using a value of the ion product of water (Kw) of 1.297  1014 at 37 C and an ionic strength of 0.3 M.16 The closed symbols represent the experimentally determined apparent first-order rate constants. As apparent from Figure 3, good agreement between the solid curves and the experimentally determined rate constants was achieved (R2 >0.997). The fitted values of Ka and the second-order rate constants associated to the individual prodrugs are presented in Table 1.

Figure 4. Degradation of DP-2 (-) and formation of diclofenac (:) and indolinone (1-(2,6-dichlorophenyl)-1,3-dihydro-2H-indol-2-one) (C) in 0.1 M borate buffer solution, pH 8.93 (37 C).

N. Mertz et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-9

low pH the formation of ILO took place in a consecutive manner as outlined in Figure 2a. In the pH range 6-10, HPLC analyses revealed that the prodrugs were subject to parallel degradation to yield diclofenac and ILO (Fig. 2b). A representative degradation profile is depicted in Figure 4 revealing that the decrease in the DP-2 concentration is accompanied by simultaneous increase in the concentrations of the 2 degradation products, that is, diclofenac and ILO, as a function of time in 0.1 M borate buffer (pH 8.9). At each sampling point, the sum of the concentrations of the 3 species was equal to the initial prodrug concentration ±5%. The first-order rate constants kP1 and kP2 (Fig. 2b) were readily estimated by combining the relationships valid for parallel reactions (i) kobs ¼ kP1 þ kP2 and (ii) kP1/kP2 ¼ [diclofenac]t/[ILO]t, where [diclofenac]t and [ILO]t refer to the concentrations at time t for diclofenac and ILO, respectively. Thus, with reference to Figure 4 the ratio kP1/kP2 was calculated as the average value of the (kP1/kP2)t ratios determined at each of the 4 sampling points between 0.7 to 6.1 h were 0.449 ± 0.006. The obtained kP1 and kP2 values for the degradation of the 3 prodrugs at pH >6 are compiled in Table S1 in Supplementary Information. It is apparent that kP1 > kP2 at slightly acidic to neutral pH, whereas the indolinone formation is the dominating degradation reaction for the investigated prodrugs at more alkaline reaction conditions. Significant indolinone formation appears unlikely to occur in the in vivo situation because enzyme-mediated cleavage proceeds more than 10-fold faster than diclofenac ester prodrug degradation in aqueous buffer (Table 2). On the other hand, the utility of macromolecular prodrugs involving diclofenac attachment to the polymeric matrix in the form of ester bonds might be less straightforward in cases where steric hindrance of the polymer backbone constitutes an impediment for enzyme access to the diclofenac ester bonds.24,25 In such cases the degradation pattern is expected to be similar to that observed in PBS (pH 7.4), that is, parallel formation of diclofenac and the indolinone structure in ratios close to unity. The indolinone, stable in buffer pH 7.4 and in human serum, is reported to exhibit anti-inflammatory as well as analgesic activity.26 The pH rate profile for the hydrolysis of DP-2 to diclofenac (6  pH  10) is depicted in Figure 5 where the logarithm to kP1 has been plotted against pH. Quite similar profiles were observed for DP-1 and DP-3. The shape of the profile (comprising 2 line segments with slopes close to þ1 connected by an inflection) strongly suggests that specific base-catalyzed hydrolysis of the protonated and neutral prodrug species occurs at different rates with the expression for the first-order rate constant kP1 given by

kP1 ¼ kOH $ aOH $fDP2Hþ þ k0OH $aOH $fDP2

(14)

where kOH and k0 OH (see Table S2 in Supplementary Information) are the second-order rate constants for hydroxide ion-catalyzed hydrolysis of the protonated and neutral prodrug species, respectively. The fractions of protonated and neutral prodrug are represented by fDP-2Hþ and fDP-2. The shape of the plot of log kP2 against Table 2 Half-Lives (mean ± standard deviation, n ¼ 3) of Prodrugs in Aqueous Solution at pH 7.4, 80% (vol/vol) SF, and 80% (vol/vol) Plasma at 37 C

7

Figure 5. pH rate profiles for formation of diclofenac (B) and indolinone (1-(2,6dichlorophenyl)-1,3-dihydro-2H-indol-2-one) (-), respectively, from DP-2 at pH >6 and 37 C.

pH was comparable to that of log kP1 (Fig. 5). For the ILO formation, therefore, the first-order rate constant kP2 can be expressed as

kP2 ¼ kDP2Hþ $ aOH $fDP2Hþ þ k0DP2 $aOH $fDP2

(15)

where kDP-2Hþ and k0 DP-2 represent the second-order rate constants for hydroxide ion-catalyzed ILO formation of the protonated and neutral prodrug species, respectively. Separate investigations of the stability of the diclofenac methyl ester (synthesized according to Sriram et al.27) resulted in a degradation profile quite similar to those observed for DP-1, DP-2, and DP-3 in aqueous solution at pH 6-10 suggesting that the promoiety structure was not a major determinant as regard the rate of ILO formation. It should be emphasized that the diclofenac methyl ester is poorly soluble in water and does not exhibit pH-dependent solubility and is thus not of interest in regard to the investigated prodrug approach. Mechanistically, these results can be rationalized in terms of a kinetic scheme in which the indolinone formation for both the protonated and neutral prodrug involves an intra-molecular nucleophilic attack of the diarylamine anion upon the neighboring ester carbonyl moiety (see Fig. S1 in Supplementary Information). A similar mechanism was previously reported for the cyclization of methyl esters of succinamic and glutaramic acids to the corresponding imides.28 The ionization constant for deprotonation of the diaryl moiety is unknown but expected to be of the order of 1023.22

Variable Chemical Hydrolysis, pH 7.4a (min) 80% SF (min) 80% Plasma (min) DP-1 DP-2 DP-3

4.97  102 4.55  102 1.32  103

10.5 ± 0.9 115 ± 6.9 18.9 ± 1.2

1.7 ± 0.1 16.5 ± 0.7 3.0 ± 0.1

a Calculated using Equation 12 and estimated second-order rate constants at zero buffer concentration.

Conversion of Prodrugs to Diclofenac in Human Synovial Fluid and Plasma The stability of the diclofenac ester prodrugs after incubation in human 80% (vol/vol) SF and 80% (vol/vol) plasma was studied at

8

N. Mertz et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-9

37 C. Concomitant determination of the concentrations of intact prodrug and released diclofenac was done by HPLC. In all cases, degradation kinetics was applied to apparent first-order kinetics. Compared to the stability in PBS (pH 7.4), the prodrugs underwent much faster degradation in the biological media indicating the involvement of enzyme-mediated prodrug conversion (Table 2) to yield the active diclofenac. Mass balance calculations related to the 6 experiments (done in triplicate) revealed that reconversion of the prodrugs to diclofenac was almost quantitative with the average sum of the concentrations of intact prodrug and released diclofenac amounting to about 86 ± 11% of the initial prodrug concentration at each sampling point. Formation of the indolinone was not observed after prodrug incubation in the 2 biological matrixes. The relatively high standard deviation should been seen in the light of that the degradation experiments were carried out in very dilute solutions as a consequence of the poor solubility of the prodrugs at pH 7.4. For the 3 prodrugs, the susceptibility of the prodrugs to undergo enzyme-catalyzed cleavage in 80% (vol/vol) plasma exceeded that determined in 80% (vol/vol) SF by a factor of about 6 (Table 2). Previously, using another SF pool, a ratio of about 2 for the cleavage rate of DP-2 incubated in 80% (vol/vol) plasma and 80% (vol/vol) SF, respectively, was reported.4 Because macromolecules (MW of about 90 kDa or less) can be transported from the blood to SF, the apparent lower enzyme activity in SF may reflect a quantitative difference in the esterase content of the 2 media in accord with previous findings that the protein content in an SF pool (4.2 ± 0.3% (wt/vol)) was approximately half of that in plasma (8.2 ± 0.2% (wt/ vol)).29 Although SF is considered an ultra-filtrate of plasma,2 the presence of different esterases may also contribute to the differences with respect to prodrug degradation observed in the 2 media. In Vitro Release Diclofenac Profiles From Suspensions of DP-1, DP-2, and DP-3 at Buffer (pH 7.4) As already mentioned, the diclofenac ester prodrug approach aims at providing local and sustained diclofenac action after injection into joints in the form of a slightly acidic prodrug solution, which upon contact with the SF precipitate lead to suspension formation.4 In order to obtain basic knowledge on the potential of the prodrugs in relation to a suspension-based depot strategy, the release from pre-formed prodrug suspensions was investigated. The application of pre-formed suspensions rather than solutions allows the separation of the prodrug dissolution and hydrolysis/ drug conversion effects from supersaturation and precipitation events. In analogy to marketed aqueous microcrystalline corticosteroid ester prodrug suspension,30,31 the onset and duration of action in the joint cavity of diclofenac released from in situ formed diclofenac ester prodrug suspensions is expected to be influenced by inherent ester prodrug properties, that is, pH-dependent solubility and charge as well as rate of esterase-facilitated reconversion to diclofenac.4 DP-1, DP-2, and DP-3 differ with respect to these latter basic characteristics as apparent from Table 1 and Table 2. For comparative purposes, the reaction conditions for the preliminary release studies using the rotating dialysis cell were kept almost constant. Employing a continuous phase comprised of a 67 mM PBS (pH 7.4) containing 0.01% (vol/vol) Tween 80, 1.0 mL of preformed suspensions (approximately 25 mmol/mL) of DP-1 (mean particle size 18 mm) or DP-2 (mean particle size 7 mm) was added to 4.0 mL 80% (vol/vol) SF in the dialysis cell. In contrast to crystalline nature of the base form of DP-1 and DP-2, the neutral form of DP-3 existed as a semi solid. Thus, in case of DP-3, 10 mg solid prodrug (as the nitrate salt) followed by 1.0 mL 0.01% Tween 80 in 67 mM PBS (pH 7.4) was added to the above-mentioned release medium. A common feature of the 3 release profiles is an initial lag time of diclofenac appearance in the acceptor phase (Fig. 6). This phenomenon

Figure 6. Release profiles of diclofenac from prodrug suspensions of DP-1 (:, n ¼ 3), DP-2 (B, n ¼ 3), and DP-3 (-, n ¼ 2) in 80% (vol/vol) human synovial fluid at 37 C in the rotating dialysis model. A total dose of approximately 25 mmol prodrug was added to the donor compartment. Bars represent standard deviation.

can be ascribed to that diclofenac that is highly bound to serum albumin in human SF. Thus, passive transport of diclofenac, formed from enzyme-mediated cleavage of the prodrugs, from the donor to the acceptor compartment is impeded as long as the protein saturation has not been reached. In comparison to the behavior of the DP-2 and DP-3 suspensions, the diclofenac release from the DP1 suspension proceeded much faster despite the fact that DP-1 is 5 times less soluble than DP-2 and DP-3 at physiological pH 7.4 (Table 1). On the other hand, the DP-1 cleavage rate in 80% (vol/vol) SF exceeded that of DP-2 by a factor 12. The indolinone derivative was not detected in the acceptor phases in accordance with the fact that in all cases the enzymatic cleavage of the prodrugs proceeded much faster than the intra-molecular cyclization reaction. Although preliminary in nature these findings may suggest that the shape of the diclofenac release profiles can be optimized by choosing diclofenac ester prodrugs exhibiting feasible combinations of solubility and rate of diclofenac reconversion in SF. Future studies in our laboratory will be directed on characterizing the in situ suspension-forming properties of the prodrug derivatives. Conclusion pH rate profiles for the degradation of the prodrugs DP-1, DP-2, and DP-3 at 37 C were constructed and rate equations for prodrug degradation in aqueous solution in the pH range 1-10 were derived (Eq. 12). Qualitatively, the 3 pH rate profiles were comparable but differed with respect to absolute rates of degradation most pronounced in the pH range 2-4 including the pHmin for the 3 prodrugs. At pH 5, the prodrugs underwent hydrolysis to give diclofenac and promoiety. In the pH range 6-10, the prodrugs were subject to parallel degradation to yield diclofenac and the indolinone derivative 1-(2,6-dichlorophenyl)-2-indolinone. The stability studies suggest that the prodrugs should be formulated as dry powders for reconstitution. The rate of degradation of the prodrugs was about 6-fold faster in 80% (vol/vol) human plasma than in 80% (vol/vol) human SF, which might reflect the higher protein content of human plasma relative to SF or the presence of different esterases in the biological media. In SF and plasma, DP-2 was observed to degrade more slowly than DP-1 and DP-3. In vitro release studies performed on pre-formed suspensions of the prodrugs revealed significantly different release behavior among

N. Mertz et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-9

the prodrugs and indicate a complex relationship between prodrug solubility, dissolution, and cleavage rate. Thus, a more detailed understanding of the inter-relationship between these parameters together with efficacy data is required to select the optimal prodrug structure with respect to onset and duration of diclofenac action in the joint.

References 1. Aly MNS. Intra-articular drug delivery: a fast growing approach. Recent Pat Drug Deliv Formul. 2008;2(3):231-237. 2. Larsen C, Ostergaard J, Larsen SW, et al. Intra-articular depot formulation principles: role in the management of postoperative pain and arthritic disorders. J Pharm Sci. 2008;97(11):4622-4654. 3. Evans CH, Kraus VB, Setton LA. Progress in intra-articular therapy. Nat Rev Rheumatol. 2014;10(1):11-22. 4. Thing M, Agardh L, Larsen S, et al. A prodrug approach involving in situ depot formation to achieve localized and sustained action of diclofenac after joint injection. J Pharm Sci. 2014;103(12):4021-4029. 5. Caldwell JR. Intra-articular corticosteroids. Guide to selection and indications for use. Drugs. 1996;52(4):507-514. 6. Gerwin N, Hops C, Lucke A. Intraarticular drug delivery in osteoarthritis. Adv Drug Deliv Rev. 2006;58(2):226-242. 7. Hunter JA, Blyth TH. A risk-benefit assessment of intra-articular corticosteroids in rheumatic disorders. Drug Saf. 1999;21(5):353-365. 8. Anthony D, Jasinski DM. Postoperative pain management: morphine versus ketorolac. J Perianesth Nurs. 2002;17(1):30-42.
pez JE, Lo
pez-Mun ~ oz FJ, Castan ~ eda-Herna
ndez G, Flores-Murrieta FJ, 9. Torres-Lo Granados-Soto V. Pharmacokinetic-pharmacodynamic modeling of the antinociceptive effect of diclofenac in the rat. J Pharmacol Exp Ther. 1997;282(2):685-690. 10. Hiestand EN. Theory of coarse suspension formulation. J Pharm Sci. 1964;53(1): 1-18. 11. Larsen C, Larsen S, Thing M, Kristensen J, Jensen H, Østergaard J, inventors. Prodrugs of Non-steroid Anti-inflammatory Agents (NSAIDs) Patent. 2012. WO 2013/064153 A1. 12. Perrin DD. Dissociation Constants of Inorganic Acids and Bases in Aqueous Solution. Canberra: Australian National University; 1969. 13. Kortüm G, Vogel W, Andrussow K. Dissociation constants of organic acids in aqueous solution. Pure Appl Chem. 1960;1(2-3):187-536. 14. Connors KA. Reaction Mechanisms in Organic Analytical Chemistry. New York, NY: John Wiley & Sons, Inc.; 1973:41-60. 15. Larsen C, Bundgaard H. Kinetics of the acid-catalyzed cyclization of diclofenac to an indolinone in aqueous solution. Arch Pharm Chem Sci Ed. 1980;8:100-108.

9

16. Harned HS, Hamer WJ. The inoization constant of water and dissociation of water in potassium chloride solutions from electromotive forces of cells without liquid junction. J Am Chem Soc. 1933;55(6):2203-2204.
cs-Nov
17. Taka ak K, Box KJ, Avdeef A. Potentiometric pKa determination of waterinsoluble compounds: validation study in methanol/water mixtures. Int J Pharm. 1997;151(2):235-248. 18. Larsen S, Østergaard J, Yaghmur A, Jensen H, Larsen C. Use of in vitro release models in the design of sustained and localized drug delivery systems for subcutaneous and intra-articular administration. J Drug Deliv Sci Technol. 2013;23(4):315-324. 19. Van der Houwen O, De Loos M, Beijnen J, Bult A, Underberg W. Systematic interpretation of pH-degradation profiles. A critical review. Int J Pharm. 1997;155(2):137-152. 20. Connors KA. Reaction Mechanisms in Organic Analytical Chemistry. New York, NY: John Wiley & Sons, Inc.; 1973:120-140. 21. Geibel J, Cannon J, Campbell D, Traylor T. Model compounds for R-state and Tstate hemoglobins. J Am Chem Soc. 1978;100(11):3575-3585. 22. Bordwell F, Cheng J, Ji GZ, Satish A, Zhang X. Bond dissociation energies in DMSO related to the gas phase values. J Am Chem Soc. 1991;113(26):9790-9795. 23. Larsen C, Bundgaard H. Kinetics and mechanism of degradation of mecillinam in aqueous solution. Arch Pharm Chem Sci Ed. 1977;5:66-86. 24. Wang F, Finnin J, Tait C, et al. The hydrolysis of diclofenac esters: synthetic prodrug building blocks for biodegradable drug-polymer conjugates. J Pharm Sci. 2016;105:773-785. 25. Stella V, Borchardt R, Hageman M, Oliyai R, Maag H, Tilley J. Controlled ReleasedMacromolecular Prodrugs. Prodrugs: Challenges and Rewards. USA: Springer Science & Business Media; 2007:379-416. 26. Chung MC, Dos Santos JL, Oliveira EV, Blau L, Menegon RF, Peccinini RG. Synthesis, ex vivo and in vitro hydrolysis study of an indoline derivative designed as an anti-inflammatory with reduced gastric ulceration properties. Molecules. 2009;14(9):3187-3197. 27. Sriram D, Yogeeswari P, Devakaram RV. Synthesis, in vitro and in vivo antimycobacterial activities of diclofenac acid hydrazones and amides. Bioorg Med Chem. 2006;14(9):3113-3118. 28. Bundgaard H, Larsen C. Pro-drugs as drug delivery systems. V. Cyclization of methyl esters of succinamic and glutaramic acids to the corresponding imides (phensuximide and glutethimide) in aqueous solution. Acta Pharm Suec. 1978;16(5):309-318. 29. Thing M, Jensen SS, Larsen C, Ostergaard J, Larsen SW. Modification of concomitant drug release from oil vehicles using drug-prodrug combinations to achieve sustained balanced analgesia after joint installation. Int J Pharm. 2012;439(1-2):246-253. € llmann H, Grüner A, Haack D, Gyselby G. Pharmacokinetics 30. Derendorf H, Mo and pharmacodynamics of glucocorticoid suspensions after intra-articular administration. J Clin Pharmacol. 1986;39(3):313-317. € llmann H, Voortman G, et al. Pharmacokinetics of rimexolone 31. Derendorf H, Mo after intra-articular administration. J Clin Pharmcol. 1990;30(5):476-479.