Development of a selective matrix metalloproteinase 13 (MMP-13) inhibitor for the treatment of Osteoarthritis.

Abstract

Osteoarthritis (OA) is a chronic disorder that causes damage to the cartilage and surrounding tissues and is characterized by pain, stiffness, and loss of function. Current treatments for OA primarily involve providing only relief of symptoms but does not affect the overall trajectory of the disease. A major goal for treating OA has been to slow down or reverse disease progression. Matrix metalloproteinase-13 (MMP-13) is expressed by chondrocytes and synovial cells in human OA and is thought to play a critical role in cartilage destruction. Herein we report a new, allosteric MMP-13 inhibitor, AQU-019, that has been optimized for potency, metabolic stability, and oral bioavailability through a combination of structure activity relationship (SAR) and deuterium substitution as a potential disease modifying OA drug (DMOAD). The inhibitor was demonstrated to be chondroprotective when injected intraarticular (IA) in the monoiodoacetic acid (MIA) rat model of OA.

Graphical Abstract

Introduction:

Osteoarthritis (OA) is a serious and painful condition that is characterized primarily as a degenerative joint disease resulting from progressive loss of articular cartilage and subchondral bone (1). It is the most prevalent articular disease among those over the age of 65, with an estimated 32 million people, or 1 in 7 Americans being affected (2). Osteoarthritis accounts for 25% of visits to primary care physicians and half of all non-steroidal anti-inflammatory drug (NSAID) prescriptions (3). At present, there are no FDA approved disease modifying OA drugs (DMOAD) available. The current management of osteoarthritic pain and inflammation mostly involves the use of opioids, non-opioids, and NSAIDs, as well as intra-articular injections of hyaluronic acid, which in some retrospective studies have shown little to no clinical difference over placebo (4, 5). The use of analgesic medications gives only temporary relief and is not disease modifying. Ultimately, as the disease progresses, the only option remaining for most patients is joint replacement surgery. Clearly, there is an unmet clinical need for a DMOAD to slow disease progression and to mitigate pain.

OA is believed to result from both biomechanical and molecular changes in the joint brought about by ageing, obesity, injury, joint malalignment, and inflammation (6). In the clinic, histopathological analysis has found that a majority of OA patients suffer from inflammation of the synovium (synovitis) (7). As such, it has now been established that synovitis and the resultant proinflammatory mediators are important in the pathogenesis of OA with effects on articular cartilage (8–10). While there have been many target molecules (i.e., interleukin-1β, tumor necrosis factor alpha, basic fibroblast growth factor, et al.) that have been identified as potentially being involved with disease onset, they all have been found to converge to a few common downstream enzymes that break down articular cartilage central to OA, such as the matrix metalloproteinases (MMPs), of which there are approximately twenty-five known different subtypes. MMPs are zinc-dependent, potent endopeptidases belonging to a larger family of proteases known as the metzincin superfamily. Of all of the MMPs, attention as it relates to OA has focused on matrix metalloproteinase-13 (MMP-13)

There is a large body of data pointing to matrix metalloproteinase-13 (MMP-13) as being central to OA. For example, (a) over-expression of MMP-13 is found to occur in the majority of OA patients and can hardly be detected in normal adult tissues, (b) MMP-13 expression is largely limited to the articular joint particularly compared to other MMP’s, (c) MMP-13 transgenic mice develop cartilage defects similar to humans, (d) mutation in human MMP-13 causes the Missouri variant of spondyloepimethaphyseal dysplasia (SEMD) (abnormalities in development and growth of skeletal elements) and (e) inhibition of MMP-13 reduces disease progression in rodent animal models of OA (11–16).

Development of orally bioavailable, MMP inhibitors for OA involved utilizing broad-spectrum MMP inhibitors that exhibited high potency but had very poor selectivity for any one member of the MMP family. As such, early studies using these broad-spectrum MMP inhibitors were found to exhibit dose limiting toxicity in the form of musculoskeletal side effects, including joint stiffness and inflammation identified as Muscular Skeletal Syndrome (MSS) (17, 18). More than thirty broad spectrum MMP inhibitors were orally tested in the clinic (mostly for a cancer indication) and were discontinued due to MSS (19, 20). Because MSS was found to be dose and time related, patients in many of the trials either withdrew from treatment completely, moved to much lower doses for which there was no indication of expected efficacy, or skipped doses for a period of time allowing for the MSS symptoms to resolve (21). MSS is believed to be caused by a lack of selectivity and/or the incorporation of a hydroxamic acid functional group (found in many broad-spectrum MMP inhibitors) which can ubiquitously bind many different metal containing proteins (17, 19, 20).

The pyrimidine dicarboxamide 1 (Figure 1-A) is part of a class of molecules that have been reported to bind exclusively to an allosteric binding site (S1’ side pocket) of MMP-13 (22, 23). Good selectivity and potency toward MMP-13 can be obtained given the appropriate functionality around the phenyl moieties. However, some metabolic liabilities have been associated with this class of molecules (24). For example, the microsomal stability of compound 1 was measured using rat liver microsomes and found to be 0.6% (Figure 1-A) (25). In some cases, the lack of microsomal stability has been ameliorated through the selective substitution of a methyl group at one or more of the benzylic positions of the phenyl benzamides 2 (Figure 1-B) (26). The fact that substitution at the benzyl position can improve the microsomal stability may indicate that the benzyl carbon could be involved in cytochrome P450 (CYP) mediated oxidation. However, due to the restricted space within the S1’ binding pocket only small sized substituents are tolerated (methyl or ethyl) and they must be oriented in an S enantiomeric configuration (as opposed to R) in order to preserve the compound’s potency for inhibiting MMP-13 (Figure 1-B). Additionally, synthesizing compounds stereoselectivity can be a challenge especially upon their scale-up.

Figure 1.

A) Pyrimidine dicarboxamide 1 with a reported MMP-13 IC₅₀ = 8 nM. B) Face preference for MMP-13 inhibition for methyl substitutions at one of the benzyl positions of compound 2. C) Pyrimidine dicarboxamide AQU-019 with selective.

We were interested in testing whether substituting the hydrogens of the benzyl carbon with deuterium instead of an alkyl group could metabolically stabilize the pyrimidine dicarboxamide moiety while still maintaining its potency.

Kushner and coworkers have presented examples of how incorporating deuterium into a drug can often reduce the level of metabolic-induced transformations mediated by CYP (27). This phenomenon is due to the fact that atomic substitution of a hydrogen by a deuterium in a drug alters the strength of the carbon-deuterium bond of the drug, while keeping its 3D surface identical to that of its non-deuterated parent. Substitution of deuterium for hydrogen can give rise to an isotope effect that can alter the pharmacokinetics of the drug (28). In a reaction in which the cleavage of a C-H bond is rate determining, the same reaction of the C-D analogue will be reduced. Another advantage to using deuterium substitution to metabolically stabilize the benzyl carbon position is that one can substitute both benzyl hydrogens without affecting the compound’s potency, and as a result, avert the consequence of creating a chiral center.

Herein we report the synthesis of a potent pyrimidine dicarboxamide derivative, AQU-019, which has been optimized for potency, in-vitro microsomal stability and in-vivo oral bioavailability and has been tested for its ability to prevent cartilage degradation in the monoiodoacetic acid (MIA) rat model of OA (Figure 1-C).

Results

Chemistry.

Retrosynthetic analysis of AQU-019 indicates that a key step to introducing deuterium at the benzyl position involves the reduction of the nitrile intermediate 7 to the resulting deuterated benzyl amine 4 (Scheme 1). While conversion of phenyl nitriles to their resulting benzyl amine product is straightforward, the number of published methods available to accomplish this was surprisingly rather limited (29–32). We tried three different methods adapted from the literature to introduce deuterium into the benzylic position. While all of the approaches produced the desired deuterated product, we found that the yields varied depending on which types of substituents (R1, R2) were placed around the ring and the type of work-up procedure used (Scheme 2). Of the three methods tested, we found the palladium catalyzed deuteriumation of the phenyl nitrile 8 to the resulting benzyl amine 9 to be preferred due to its ease of work-up and high yield and purity of the resulting crude amine hydrochloride product. However, we found that the procedure needed to be optimized to provide for a high percent deuterium incorporation. For example, if the reaction was run in the presence of ethanol in addition to the desired deuterated product 11, a sizable percentage (39%) of the resulting hydrogenated product 12 would also be formed (Scheme 3). Fortunately, by replacing the ethanolic solvent with its deuterated version (i.e., 99% ethanol-D), we were able to afford the desired deuterated product 11 with a high percent (98%) deuterium incorporation (Scheme 3).

Scheme 1.

Retrosynthetic analysis of selectively deuterated pyrimidine dicarboxamide AQU-019.

Scheme 2.

Methods for incorporating deuterium via nitrile reduction. *Crude yield ≥95% purity determined by high pressure liquid chromatography (HPLC) and proton NMR spectroscopy (H-NMR).

Scheme 3.

Comparison of percent deuteration of benzyl position after deuteriumation of nitrile 7 using 5% palladium on carbon and deuterium gas in the presence of either deuterated or non-deuterated ethanol solvent to give benzyl amines 11 & 12. Crude yield ≥95% purity by HPLC and H-NMR. Percent deuteration determined by H-NMR of deprotected benzylamine.

The synthesis of AQU-019 begins with the palladium catalyzed deuteriumation of commercially available 4-cyano-benzoic acid methyl ester 7 to give the resulting benzyl amine; which was then trapped in-situ with di-tert-butyl decarbonate to give the resulting t-butyl-carbamate protected amine 10 in 95% crude yield (Scheme 4). The t-butyl-carbamate product was then deprotected using HCl to give the resulting hydrochloride salt 11 in 88% yield. The other benzylamine portion of the pyrimidine dicarboxamide was synthesized via the reduction of commercially available 4-fluoro-3-methyl-benzonitrile 5 using LiAlH₄ to give the resulting benzylamine, which was then subsequently treated with HCl to give the resulting benzylamine hydrochloride salt 13 in 95% yield. Commercially available pyrimidine-4,6-dicarboxylic acid dimethyl ester 6 was then partially deprotected with base to give the pyrimidine-4,6-dicarboxylic acid monomethyl ester compound 3 in 86% yield. The monoester 3 was then coupled to the 4-Fluoro-3-methyl-benzylamine hydrochloride portion 13 using 1-hydroxy-7-azabenzotriazole (0.82 mmol) (HOAT) and 2-(7-azabenzotriazole-1-yl)-N-N-N-N-tetramethyluronium-hexafluorophosphate (HATU) in dimethylformamide (DMF) in the presence of excess N-methyl morpholine (NMM) to give after column chromatography 6-(4-fluoro-3-methyl-benzylcarboamoyl)-pyrimidine-4-carboxylic acid methyl ester 14 in 66% yield. Deprotection of the methyl ester of 14 with base gave the resulting free acid 15 upon acidification in 70% yield. The free acid 15 was then coupled with the deuterated benzyl amine hydrochloride 11 using HOAT, HATU, and NMM in DMF to give the resulting bis-amide pyrimidine product 16 in 93% yield after column purification. Hydrolysis of the methyl ester 16 using NaOD in D₂O followed by acidification gave AQU-019 in 50% yield after column chromatography (Scheme 4).

Scheme 4.

Synthesis of pyrimidine dicarboxamide AQU-019: Reagents and conditions. (a) di-tert-butyl dicarbonate, 10% Pd on C, CH₃CH₂OD, D₂, 45 psi, 95%; (b) HCl, 88%; (c) LiAlH₄, THF, 0 °C. then NaOH/H₂O then HCl, 95%; (d) NaOH, 86%; (e) Compound 13, HOAT, HATU, NMM, DMF, 66%; (f) KOH, MeOH, 70%; (g) Compound 11, HOAT, HATU, NMM, DMF, 93%; (h) THF, NaOD, D₂O, then HCl, 50%.

Pharmacology.

The potency of AQU-019 against the catalytic domain of recombinant human MMP-13 was determined to be 4.8 nM. The compound is not only potent but highly selective with little inhibition for other MMPs (Table 1).

Table 1.

MMP IC₅₀ (nM) inhibitory activity profile for AQU-019.^{1, 2}

MMP-1	MMP-2	MMP-3	MMP-7	MMP-8	MMP-9	MMP-10	MMP-12	MMP-13	MMP-14
>100000	>100000	>100000	>100000	98000	74000	1000	>100000	4.8	>100000

¹12-dose IC₅₀ curve with 3-fold serial dilution starting at 100 μM for designated catalytic domain of specific MMP.

²IC₅₀ curve performed in duplicate with reported value being the mean.

AQU-019 showed a modest improvement in microsomal stability with a lower turnover (99% of parent remaining after 1 hour) as compared to the non-deuterated parent 18 (90% of parent remaining after 1 hour) (Table 2) (25). AQU-019 and its non-deuterated parent 18 were then orally dosed via gavage in male Lewis rats (3 rats/group) at 1 mg/kg. Both compounds were dosed as an aqueous suspension in 0.5% methyl cellulose. Whole blood samples were collected from all animals, pre-dose, and at 0.25, 0.5, 1, 2, 4, 8, 16, and 24 hours after dose administration. After compound extraction, pharmacokinetic (PK) parameters were then estimated using a non-compartmental approach consistent with the PO route of administration. The Cmax & AUC_0-t of AQU-019 was observed to increase 17% and 68% respectively over that of the non-deuterated parent 18 (Table 2).

Table 2.

Comparison of in-vitro rat microsomal stability using rat liver microsomes & in-vivo single oral (PO) dose pharmacokinetics (PK) (1 mg/kg) in male Lewis rats between AQU-019 and its non-deuterated parent 18.

Compound Structure & #	%Mean remaining parent with NADPH with RLM (%)1	AUC(0-t)2 (ng•hr/mL)	AUC (0-inf)3 (ng·hr/mL)	T1/24 (hr)	Cmax5 (ng/mL)	Tmax6 (hr)
	99	697	705	1.22	362	0.75
	90	414	428	1.84	310	0.41

¹at T = 60 minutes at 1 microMolar compound concentration using rat liver microsomes (RLM). Average of two runs.

²AUC(0-t) = area under the curve versus time curve from time zero to the time after dosing at which the last quantifiable concentration of the drug was observed for an oral gavage dose of 1 mg/kg.

³AUC(0-inf) =The area under the arithmetic mean concentration versus time curve from time zero to infinity for an oral gavage dose of 1 mg/kg.

⁴T1/2 = the apparent terminal elimination half-life.

⁵Cmax = maximum observed concentration, occurring at Tmax.

⁶Tmax = Time of maximum observed concentration.

In order to determine whether MMP-13 inhibition can translate to chondroprotection, AQU-019 was then tested in the monosodium iodoacetate (MIA) rat model of OA via once/week intraarticular (IA) injection. MIA inhibits glyceraldehyde-3 phosphate dehydrogenase, a key enzyme of the glycolytic pathway. Inhibition of this pathway within chondrocytes leads to chondrocyte apoptosis and subsequent cartilage degeneration. The injection of MIA results in an early severe inflammatory response that is characterized by joint swelling and infiltration by neutrophils, macrophages, and lymphocytes that then subsides within the first week (33–35). Five days after MIA treatment of the joint, AQU-019 (1 mg), Triamcinolone (0.06 mg) as positive control, and Vehicle were injected once per week for three weeks (Table 3). After twenty-six days post MIA injection, the animals were sacrificed and the joints removed and sectioned. Histology showed that the Vehicle group had very severe cell, proteoglycan, and matrix loss that was relatively consistent on the medial tibial plateau (MTP), medial femoral condyle (MFC), lateral tibial plateau (LTP), and lateral femoral condyle (LFC) (Fig. 3 & 4). Bone resorption was severe (93% overall) with lower severity on the MTP (83%) (Figure 5). Osteophytes had a mean measurement of 385 μm (Figure 5), All parameters except for subchondral bone sclerosis were significantly increased over the naïve controls.

Table 3.

IA Testing of Narp-13 inhibitor, AQU-019 in the MIA-Rat Model of OA.

Compound: AQU-019	Dosing:	IA injection on Day 5, 12 and 19 for 26 clays. Dosing began 5 days after MIA injection.
Gender = Male
Anirnal/Strain = Sprague-Dawley (175–200 grams)
Total test Animals = 29 MIA- Animals =24

Group	#Rats	Dose	IA Regimen	Compound
1	8	54µl	Days 5,12,19	Vehicle1
2	8	0.06nag	Days 5,12,19	Triamcinolone2
3	8	1mg	Days 5,12,19	AQU-0192
4	5	NA	NA	Naive3

¹Velicle composed of saline.

²Compound was injected as a suspension in 54µl of saline.

³Naive consists of no IA injections of either MIA or vehicle.

NA = not applicable

Figure 3. Histology of joints in MIA-rat model.

AQU-019 exhibited statistically significant reduction in cartilage degeneration in the MTP, LTP, and LFC compared to Vehicle control. (* P<0.05). Triamcinolone exhibited no reductions in cartilage degeneration. (MTP)=medial tibial plateau, (MFC) = medial femoral condyle, (LTP) = lateral tibial plateau, (LFC) = lateral femoral condyle, Mean = all surfaces. Post-test comparisons were made from each treatment group to the Vehicle control group. Vehicle = phosphate buffered saline.

Figure 4. Histology of joints in MIA-rat model.

AQU-019 displayed a trend of reduced resorption compared to Vehicle but was not statistically significant. Triamcinolone, exhibited significantly reduced bone resorption in the MFC and LFC (49–50%) with nearly significant reductions on the tibia (34–39%, *P = 0.054–0.056) contributing to a significant 43% reduction in the overall mean. (MTP)=medial tibial plateau, (MFC) = medial femoral condyle, (LTP) = lateral tibial plateau, (LFC) = lateral femoral condyle, Mean = all surfaces. Post-test comparisons were made from each treatment group to the Vehicle control group. Vehicle = phosphate buffered saline.

Figure 5. Histology of joints in MIA-rat model.

AQU-019 displayed no reduction in osteophyte measurement compared to Vehicle control. Triamcinolone exhibited a statistically significant (40–47%) reduction in osteophyte compared to Vehicle control (* P<0.05).

Animals treated with AQU-019 exhibited statistically significant reductions in proteoglycan (PG) loss on the MTP, LTP, and LFC (15–19%), as well as statistically significant reductions in amount of cell loss (8%) on the LTP as compared to Vehicle and positive controls (Fig. 3 & 4). This contributed to significant reductions in the mean loss on the LTP and LFC (21–22%), as well as the mean of all surfaces (16%) (Fig. 3 & 4). Bone resorption, while reduced, was not statistically significant compared to Vehicle control (Figure 5). Osteophyte scores were comparable to Vehicle control (Figure 5). Data was analyzed using a one-way analysis of variance (1-way ANOVA) or Kruskal-Wallis test (non-parametric ANOVA), along with the appropriate multiple comparison post-test. Animals treated with 0.06 mg of Triamcinolone had no significant reductions in cell loss compared to Vehicle control (Figure 4). However, triamcinolone exhibited significantly reduced bone resorption in the MFC and LFC (49–50%) with nearly significant reductions on the tibia (34–39%, p = 0.054–0.056), contributing to a significant 43% reduction in the overall mean (Figure 5). Osteophyte scores were significantly reduced by 40–47% compared to Vehicle control (Figure 5).

Discussion and Conclusion.

The incidence of OA increases with age, with approximately 80% of those that reach 65 years of age showing some radiographic evidence of disease (36). With the aging of the world’s population, the burden of OA will continue to rise, making the need for a DMOAD even greater. There has been a great deal of accumulated evidence to support the pivotal role that MMP-13 plays in breaking down articular cartilage in OA, making the use of a MMP-13 inhibitor seem like a logical strategy for a DMOAD (8–10). However, proving that inhibiting MMP-13 leads to chondroprotection in OA has proven difficult in a clinical setting due the oral toxicity exhibited by earlier non-specific MMP-13 inhibitors (17–20). The discovery of the pyrimidine dicarboxamide as well as other scaffolds capable of inhibiting MMP-13 via an allosteric, non-zinc binding site (S1’ side pocket) has given new impetus for reexamining the clinical use of a MMP-13 inhibitor as a DMOAD (22, 23, 37). In pursuit of this strategy, we took as a starting point the pyrimidine dicarboxamide 1, and using a combination of structure activity relationship (SAR) and deuterium substitution, produced the pyrimidine dicarboxamide derivative, AQU-019. We tried three different synthetic approaches toward selectively incorporating deuterium into the benzyl carbon position, and while each method gave the desired deuterated benzyl amine product 9, the palladium catalyzed deuteriumation of phenyl nitrile 8 to the corresponding benzylamine 9 was preferred due to the ease of work-up and relative high purity of the resulting crude product (Scheme 2). Following this procedure, we were able to synthesize the desired deuterated product AQU-019 in eight synthetic steps in 14.6 % overall yield starting from the commercially available 4-cyano-benzoic acid methyl ester 7 (Scheme 3). AQU-019 exhibits improved potency compared to pyrimidine dicarboxamide 1 (MMP-13 IC50 = 4.8 nM versus 8 nM) and rat microsomal stability (99% versus 0.6% remaining after 1 hour) as compared to the starting pyrimidine dicarboxamide 1. However, the large improvement in microsomal stability for AQU-019 is primarily a result of replacing the methyl and fluoro groups from one side of compound 1 with a carboxylic acid. This is apparent upon comparing the microsomal stability of AQU-019 (99% remaining after one hour) with its non-deuterated parent 18 (99% versus 90% remaining after 1 hour) (Table 2). However, when tested orally in rats, AQU-019 exhibited significantly more systemic exposure compared to its non-deuterated parent 18 as observed through a comparison of total drug exposure across time (68% improvement in AUC_(0-t); 64% improvement in AUC_(0-infinity)) (Table 2). These results are consistent with other reported deuterium substitution efforts that have shown that even small improvements in the in-vitro microsomal stability (lowering the intrinsic clearance rate) of a compound through selective deuteration can translate into observable improvements in in-vivo oral bioavailability (39).

We were then interested in determining if AQU-019 could be chondroprotective. We chose the MIA rat model due to its ability to induce cartilage degeneration via inhibition of aerobic glycolysis, which kills chondrocytes (33–35, 39). Depending on the concentration used and frequency, different degrees of chondrocyte death (degeneration) can be achieved. In rats, a single 25–50 μl injection of 10 mg/ml, sodium iodoacetate is sufficient to kill most of the chondrocytes (35, 39). The bone changes in this model are also quite striking and form the basis for cartilage lesions observed macroscopically.

While our PK results indicated that oral dosing with AQU-019 would provide good plasma exposure, it is known that several factors, such as plasma protein binding, could affect the ultimate level of compound penetration into synovial fluid, resulting in a low synovial fluid to plasma ratio (40–42). Since our primary goal was to validate that MMP-13 inhibition in-vivo can result in chondroprotection, it was decided that IA delivery as opposed to oral would be used in the model in order to ensure sufficient compound within the joint. In addition to testing AQU-019 and a Vehicle control, we also wanted to include a known anti-inflammatory compound as a comparator. We chose the corticosteroid, triamcinolone, because it is an anti-inflammatory compound commonly used in rodent models of OA and is FDA approved in an extended-release form (Zilretta®) for the treatment of OA pain and inflammation of the knee (43, 44). Lastly, we also included a small naïve group (n = 5, Table 3) as a negative control. Twenty-six days after MIA injection, histopathological analysis of sections from the group treated with AQU-019 demonstrated significant chondroprotection as compared to the Vehicle group, which showed severe cell, proteoglycan, and matrix loss (Figures 4). The group treated with triamcinolone while exhibiting significant reductions in bone resorption and osteophyte measurement exhibited no chondroprotection (Figures 4–5). This study helps to confirm the role that MMP-13 plays in cartilage degeneration, and more importantly, validates the use of the selective MMP-13 inhibitor, AQU-019, for chondroprotection. The fact that AQU-019 had no effect on osteophytes measurement is consistent with prior studies that show no difference in osteophyte maturity and size between MMP-13 knock-out and wild type mice having undergone medial meniscal destabilization of the right knee (45).

IA injection of AQU-019 was able to inhibit MMP-13 to a sufficient extent to produce chondroprotection in the MIA rodent model. However, additional work will need to be performed to determine if oral dosing can provide a similar benefit. There is prior published data showing that the plasma to synovial penetration index (measured as the ratio of the synovial concentration/plasma concentration x 100) for various small molecule drugs (i.e. antibiotics, anti-inflammatories) can vary considerably depending on the compound, but in general falls less than 50% (40, 42, 46, 47). As such, the plasma to synovial penetration index will need to be determined for AQU-019 in order to assess how much compound can penetrate into the joint under single and steady state dosing. An oral dose response will then need to be determined for AQU-019 with regards to chondroprotection, followed ultimately by a determination of the compound’s therapeutic index (48). While AQU-019 is a selective, allosteric inhibitor of MMP-13 and would not be expected to cause MSS, the compound’s toxicity profile would still need to be evaluated.

All previous OA clinical trials utilizing MMP inhibitors as a DMOAD followed the paradigm of oral dosing. This resulted in systemic exposure of the MMP inhibitor and a primary outcome measure determined by the compound’s ability to effect a change in the joint space width (JSW), determined via x-ray (49, 50). Given that any clinical testing of a DMOAD would necessitate 1–2 years to demonstrate chondroprotection, the potential for long-term toxicity rises. It would therefore seem that testing AQU-019, as a first-in-class MMP-13 inhibitor via IA administration could have some significant benefits. IA dosing offers high drug concentrations and localized action at specific sites with limited or no systemic exposure. This is a significant advantage since a major hurdle to obtaining U.S. Food and Drug Administration (FDA) approval for a DMOAD is acceptable long-term toxicity. IA dosing of a DMOAD would produce fewer off-target effects and adverse events because of the localized nature of the delivery. The potential cost would also be reduced since one is focusing the delivery of the DMOAD to a small area, resulting in less compound needed (51). If made thirty years ago, arguments for a DMOAD delivered via IA administration would have landed on deaf ears given the pharmaceutical community’s exclusive focus on oral delivery. However, given that it has been over twenty years since the FDA provided its first guidance on the development of DMOADs and that to this day none have been approved lends greater urgency to the situation (49). There is now a growing consensus that alternative approaches should be pursued, such as the use of qualified biomarkers as primary end-points to gauge human efficacy. Indeed, the FDA has recently issued new guidance encouraging the use of qualified molecular biomarkers as primary outcome measures for OA clinical trials (52, 53). Another approach is to target OA for early detection and intervention since the process of cartilage degradation is irreversible and the onset mechanism of OA may be controllable (52, 53). Among these strategies, we include the use of a MMP-13 inhibitor such as AQU-019 as a first-in-class DMOAD, clinically developed to first be delivered via IA injection, followed by an oral program. It is hoped that with these measures a DMOAD may be approved in the not-to-distant future.

Experimental Section.

Chemistry.

Reagents were obtained from commercial sources and used without further purification unless otherwise stated. All reactions were performed using glassware that was oven dried overnight (100 °C). All solvents are of reagent grade. All reactions were carried out under nitrogen atmosphere unless otherwise stated. Organic reaction mixtures were concentrated using a Buchi rotary evaporator. Proton NMR spectra were recorded on a Varian Nuclear Magnetic Resonance spectrometer at 300 MHz. Low resolution mass spectrometry was obtained for all compound intermediates. High resolution mass spectrometry (HRMS) was obtained for final compounds submitted for biological study. Column chromatography was performed using silica gel (40−63 μm), and the reaction progress was determined by thin layer chromatography. The purity of all final compounds was established by Liquid chromatography coupled to mass spectrometry (LC-MS) and proton NMR to be ≥ 95%. The following Instrument and specifications were used to analyze the various compounds. For HPLC liquid chromatography a Shimadzu LC-10AD VP with Agilent Zobax 3.5 SB-C18 Column (4.6 mm x 50 mm) was used. The gradient was 5% to 95% Acetonitrile and water both containing 0.1% formic acid over a period of 5 minutes (1.5 ml/minute flow rate at a maximum pressure and temperature of 4000 psi and 25 °C). The mass spectrometer used was a Waters brand Micromass Quatro Ultima LC/MS (triple-quad MS) with a CTC Analytics PAL autosampler.

Pyrimidine-4,6-dicarboxylic acid bis-(4-fluoro-3-methyl-benzylamide) (1).

To a thick walled glass vessel containing a stir bar and 23 mg (0.11 mmoles) of commercially available dimethyl pyrimidine-4,6-dicarboxylate (Oakwood Products) (6) was added a large excess (0.25 ml, 1.8 mmoles) of commercially available 4-Fluoro-3-methyl-benzylamine (2) (Sigma-Aldrich) and 0.5 ml of anhydrous dimethylformamide and mixture heated while stirring under closed nitrogen atmosphere at 85 °C using microwave radiation (Biotage Microwave Reactor) for 24 hours. The volatile components of the reaction mixture were removed under reduced pressure to give a solid which was recrystallized from diethyl ether to give 42 mg (87%) of Pyrimidine-4,6-dicarboxylic acid bis-(4-fluoro-3-methyl-benzylamide) (1) as a white crystalline solid. 1H NMR (300 MHz, CD₃OD) δ 2.23 (s, 6H), 4.55 (s, 4H), 6.85–7.30 (m, 6H), 8.66 (s, 1H), 9.34 (s, 1H). HRMS calculated for C₂₂H₂₀F₂N₄O₂; MW: [M + H]⁺, 411.1554; Found 411.1558

Deuteration Procedures

4-(Amino-dideutero-methyl)-benzoic acid methyl ester hydrochloride salt (11) (Method A)

Commercially available 4-Cyano-benzoic acid methyl ester (7) (4.0 mmoles) (Aldrich) is added to a 50 ml round bottom flask containing a stir bar. To the flask is then added NiCl₂*6D₂O (0.21 mmoles) [NiCl₂*6D₂O was synthesized in the following manner: To a 25 ml round bottom flask was added 0.5 grams of commercially available NiCl₂*6H₂O (Alfa Aesar) and dissolved with 5 ml of commercially available D₂O (Cambridge Isotope Laboratories) and the volatile components removed under reduced pressure to give a yellow solid. To the solid was again added 5 ml of D₂O and the volatile components removed under reduced pressure to give NiCl₂*6D₂O]. To the 50 ml flask was then added ditertbutlycarbonate (6.8 mmoles) and mixture dissolved in 12 ml of anhydrous CD₃OD (Acros Organics) and mixture stirred under nitrogen atmosphere until solution was complete. The solution was then cooled to 0 °C and to the solution was slowly added in portions a total of 0.25 grams of NaBD₄ (Alfa Aesar) making sure to keep the temperature around 0 °C. After addition was complete the reaction was stirred under nitrogen atmosphere at 0 °C for 1 hour and then at room temperature for 24 hours. The volatile components of the reaction mixture were then removed under reduced pressure to give a residue which was taken up in 100 ml of ethyl acetate and organic layer washed with 10% citric acid and then saturated sodium bicarbonate. The organic layer was separated and dried over anhydrous magnesium sulphate, filtered and the volatile components removed under reduced pressure to give a solid which was purified by column chromatography (SiO₂, Hexane: ethyl acetate 70:30) to give 0.72 grams (67% yield) of reduced 4-(tert-Butoxycarbonylamino-dideutero-methyl)-benzoic acid methyl ester (10). 1H NMR (300 MHz, CD₃OD) δ 1.46 (s, 9H), 3.90 (s, 3H), 4.90 (br s, 1 H), 7.34 (d, 2H, J = 8.1 Hz), 7.99 (d, 2H, J = 8.1 Hz). LC-MS calculated for 2C₁₄H₁₇D₂NO₄; MW: [2M + H]⁺, 535.3; Found 535.3

To 0.72 grams of 4-(tert-Butoxycarbonylamino-dideutero-methyl)-benzoic acid methyl ester (10) in a 25 ml round bottom flask was added 8 ml of a solution composed of 4 M HCl in anhydrous dioxane and mixture stirred under nitrogen atmosphere for 2 hours. The volatile components of the reaction mixture were then removed under reduced pressure to give a white solid which was triturated with diethyl ether and the resulting solid dried under vacuum to give 0.5 grams of the desired 4-(Amino-dideutero-methyl)-benzoic acid methyl ester (11) as the hydrochloride salt (92% yield). 1H NMR (300 MHz, d6-DMSO) δ 1.46 (s, 9H), 3.84 (s, 3H), 6.63 (d, 2H, J = 8.4 Hz), 7.97 (d, 2H, J = 8.4 Hz), 8.60 (br s, 2H). LC-MS calculated for 2C₉H₉D₂NO₂; MW: [2M + H]⁺, 334.2; Found 334.2.

C,C-Dideutero-C-(4-fluoro-3-methyl-phenyl)-methylamine hydrochloride (9, R¹ = CH₃, R² = F, Scheme 2) (Method B).

Commercially available 4-Fluoro-3-methyl-benzonitrile (5) (3.0 mmoles) (Oakwood) is added to a 100 ml round bottom flask containing a stir bar. The flask was placed under vacuum then nitrogen and then 10 ml of anhydrous tetrahydrofuran was syringed in and mixture stirred until solution was complete. The solution was then cooled to –10 °C and then added in small portions LiAlD₄ (2.85 mmoles) (Aldrich) making sure to keep the temperature ~ 0 °C. After addition was complete the reaction was stirred under nitrogen atmosphere at 0 °C for 1 hour and then at room temperature for 10 hours. To the reaction was then added 0.15 ml of D₂O and then 0.1 ml of 20% NaOD and then 0.3 ml of D₂O in that order and mixture allowed to stir for 1 hour at room temperature. The reaction mixture was then filtered through celite and washed with methylene chloride. To the filtered organic liquid was then added 3 ml of a solution composed of 4 M HCl in anhydrous dioxane and the volatile components of the reaction mixture were then removed under reduced pressure to give a while solid. The white solid was triturated with diethyl ether and then placed under vacuum to give 0.5 grams of C,C-Dideutero-C-(4-fluoro-3-methyl-phenyl)-methylamine (9, R¹ = CH₃, R² = F) as the hydrochloride salt (94% yield). 1H NMR (300 MHz, d6-DMSO) δ 2.21 (s, 3H), 7.10–7.49 (m, 3H), 8.50 (br s, 2H). LC-MS calculated for C₈H₈D₂FN; MW: [M + H]⁺, 142.0; Found 142.0.

4-(Amino-dideutero-methyl)-benzoic acid methyl ester hydrochloride salt (11) (Method C)

To one equivalent (37.2 mmoles) of 4-Cyano-benzoic acid methyl ester (7) in a thick walled glass vessel was added 1.1 equivalents of di-tert-butyl dicarbonate and 1 grams of 10% Palladium on activated carbon (Aldrich) and 70 ml of CH₃CH₂OD (Cambridge Isotope Laboratories) and mixture shaken at room temperature using a Parr hydrogenator in the presence of Deuterium gas (D₂) (Aldrich) at 45 psi for 20 hours. The mixture was then filtered through a medium porosity fritted glass funnel containing celite and the retentate washed with methylene chloride. The organic washes were combined and the volatile components of the reaction mixture were then removed under reduced pressure to give 9.4 grams (95% yield, 95% pure by HPLC and proton NMR) of the desired 4-(tert-Butoxycarbonylamino-dideutero-methyl)-benzoic acid methyl ester (10). 1H NMR (300 MHz, CD₃OD) δ 1.46 (s, 9H), 3.90 (s, 3H), 4.90 (br s, 1 H), 7.34 (d, 2H, J = 8.1 Hz), 7.99 (d, 2H, J = 8.1 Hz). LC-MS calculated for C₁₄H₁₇D₂NO₄; MW: [M + Na]⁺, 290.1; Found 290.2

To 9.4 grams (35.2 mmole) of 4-(tert-Butoxycarbonylamino-dideutero-methyl)-benzoic acid methyl ester (10) in a 250 ml round bottom flask was added 115 ml of a solution composed of 4 M HCl in anhydrous dioxane and mixture stirred under nitrogen atmosphere for 5 hours. The volatile components of the reaction mixture were then removed under reduced pressure to give ½ the original volume of reaction mixture. To the reaction mixture was then added 100 ml of diethyl ether and solid filtered through a medium porosity fritted glass funnel. The filtered solid was washed with another 50 ml of diethyl either and then dried under pump vacuum to give 6.3 grams (88% yield) of the desired 4-(Amino-dideutero-methyl)-benzoic acid methyl ester (11) as the hydrochloride salt. 1H NMR (300 MHz, d6-DMSO) δ 1.46 (s, 9H), 3.84 (s, 3H), 6.63 (d, 2H, J = 8.4 Hz), 7.97 (d, 2H, J = 8.4 Hz), 8.60 (br s, 2H). LC-MS calculated for C₉H₉D₂NO₂; MW: [M + H]⁺, 168.1; Found 168.3.

4-Fluoro-3-methyl-benzylamine hydrochloride salt (13).

Commercially available 4-Fluoro-3-methyl-benzonitrile (5) (0.47 g, 3.5 mmoles) (Oakwood) is added to a 100 ml round bottom flask containing a stir bar. The flask was placed under vacuum then nitrogen and then 10 ml of anhydrous tetrahydrofuran was syringed in and mixture stirred until solution was complete. The solution was then cooled to –10 °C and then added in small portions LiAlH₄ (3.4 mmoles) (Aldrich). After addition was complete the reaction was stirred under nitrogen atmosphere at 0 °C for 1 hour and then at room temperature overnight. To the reaction was then added 0.15 ml of H₂O and then 0.1 ml of 20% NaOH and then 0.3 ml of H₂O in that order and mixture allowed to stir for 1 hour at room temperature. The reaction mixture was then filtered through celite and washed with methylene chloride. To the filtered organic liquid was then added a 3 ml solution composed of 4 M HCl in anhydrous dioxane and the volatile components of the reaction mixture were then removed under reduced pressure to give a while solid. The white solid was triturated with diethyl ether and then placed under vacuum to give 0.58 grams of 4-Fluoro-3-methyl-benzylamine hydrochloride (13) (95% yield). 1H NMR (300 MHz, d6-DMSO) δ 2.21 (s, 3H), 7.10–7.49 (m, 3H), 8.50 (br s, 2H). LC-MS calculated for C₈H₁₀FN; MW: [M + H]⁺, 140.0; Found 140.0

Pyrimidine-4,6-dicarboxylic acid monomethyl ester (3).

To a 100 ml round bottom flask containing a stir bar is added 1.0 grams (5.1 mmoles) of commercially available pyrimidine-4,6-dicarboxylic acid dimethyl ester (6) (Sigma-Aldrich). To the solid was added a solution comprising 0.2 grams (5.0 mmoles) of sodium hydroxide dissolved in 10 ml anhydrous methanol and mixture stirred at room temperature under a nitrogen atmosphere for 1 hour. To the reaction mixture was then added 1.2 ml of a solution comprising 4 M hydrochloric acid in dioxane and mixture stirred for 10 minutes. To the reaction mixture was then added ~2 grams of silica gel (SiO2), and the volatile components removed under reduced pressure and solid added to a column and purified via column chromatography (SiO2, 40% ethyl acetate in hexane) to give 0.80 grams (86%) of Pyrimidine-4,6-dicarboxylic acid monomethyl ester compound (3). 1H NMR (300 MHz, CD3OD) δ 4.04 (s, 3H), 8.59 (s, 1H), 9.46 (s, 1H), LC-MS calculated for C₇H₆N₂O₄; MW: [M + H]^+, 183.0; Found 183.0

6-(4-Fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carboxylic acid methyl ester (14).

To a round bottom flask containing a stir bar was added 5.14 grams (28.2 mmoles) Pyrimidine-4,6-dicarboxylic acid monomethyl ester compound (3), 4.54 grams (25.8 mmoles) 4-fluoro-3-benzylamine hydrochloride (13), 3.0 grams (22 mmoles) of 1-hydroxy-7-azabenzotriazole (HOAT) (AK Scientific) and 8.9 grams (38.8 mmoles) of 2-(7-azabenzotriazole-1-yl)-N-N-N-N-tetramethyluronium-hexafluorophosphate (HATU) (AK Scientific). To the mixture was then added 80 ml of anhydrous dimethylformamide (DMF) and mixture stirred at room temperature under a nitrogen atmosphere for 5 minutes. Then 8.5 ml (77.4 mmoles) N-Methylmorpholine (NMM) was injected and mixture stirred under nitrogen for 24 hours. The volatile components were then removed under reduced pressure to give an oil residue which was purified by column chromatography (SiO2, 10–40% ethyl acetate: hexane) to give 5.2 grams (66%) 6-(4-Fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carboxylic acid methyl ester (14) as a colorless oil. 1H NMR (300 MHz, CDCl₃) δ 2.23 (s, 3H), 4.04 (s, 3H), 4.59 (d, 2H, J= 6.3Hz), 6.91–6.97 (m, 1H), 7.08–7.26 (m, 2H), 8.25 (br s, 1H), 8.77 (s, 1H), 9.34 (s, 1H), LC-MS calculated for C₁₅H₁₄FN₃O³; MW: [M + H]⁺, 304.1; Found 304.0

6-(4-Fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carboxylic acid (15).

To a 250 ml round bottom flask containing 5.2 grams (17.9 mmoles) of 6-(4-Fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carboxylic acid methyl ester (14) was added a stir bar and 50 ml of methanol and mixture stirred until solution was complete. To the solution was then added 4.0 grams (71.3 mmoles) of potassium hydroxide (KOH) in 10 ml methanol and mixture stirred for 10 hours. To the mixture was then added concentrated hydrochloride acid until mixture was pH ~1. The volatile components of the reaction mixture were then removed under reduced pressure to give a white solid. The solid was taken up in 100 ml of ethyl acetate and organic washed with 40 ml of saturated NaCl and then organic separated and dried over magnesium sulfate (MgSO4), filtered and then several grams of silica gel added and the volatile components removed under reduced pressure. The solid was then purified by column chromatography (SiO₂, 0–15% methanol: methylene chloride) to give 3.5 gram (70%) of 6-(4-Fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carboxylic acid (15) as a white solid. 1H NMR (300 MHz, d6-DMSO) δ 2.18 (s, 3H), 4.43 (d, 2H, J= 6.3Hz), 7.05–7.23 (m, 3H), 8.39 (s, 1H), 9.48 (s, 1H), 9.68 (br s, 1H), LC-MS calculated for C₁₅H₁₄FN₃O₃; MW: [M + H]⁺, 290.0; Found 290.0

4-(Dideutero-{[6-(4-fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carbonyl]-amino}-methyl)-benzoic acid methyl ester (16).

To a round bottom flask containing a stir bar and 6-(4-Fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carboxylic acid (15) (6 grams, 20.7 mmoles) was added 3.0 grams (14.7 mmoles) of the hydrochloride salt of 4-(Amino-dideutero-methyl)-benzoic acid methyl ester (11), 10 grams (42.5 mmoles) HOAT and 2.5 grams (18.3 mmoles) HATU. To the mixture was then added 80 ml of anhydrous dimethylformamide (DMF) and mixture stirred for a few minutes. To the mixture was then added 10.0 ml (90 mmoles) of N-methylmorpholine (NMM) and mixture stirred under nitrogen for 48 hours. The volatile components were then removed under reduced pressure to give an oil which was taken up in 400 ml of ethyl acetate and mixture washed twice with 200 ml of 10% sodium hydroxide solution then 200 ml of saturated NaCl solution. The organic layer was separated and dried over MgSO₄ and then filtered. To the solution was added 5 grams of silica gel and the volatile components removed under reduced pressure to a solid was purified by column chromatography (SiO₂, 0–20% ethyl acetate: methylene chloride) to give 6.0 grams (93%) of 4-(Dideutero-{[6-(4-fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carbonyl]-amino}-methyl)-benzoic acid methyl ester (16) as a white solid. 1H NMR (300 MHz, CDCl₃) δ 2.25 (s, 3H), 3.90 (s, 3H), 4.61 (d, 2H, J= 6.3Hz), 6.85–7.18 (m, 3H), 7.26–7.31, 7.41 (d, 2H, J =8.7 Hz), 8.02 (d, 2H, J=8.7 Hz), 8.23 (br s, 1H), 8.30 (br s, 1H), 8.93 (s, 1H), 9.18 (s, 1H), LC-MS calculated for C₂₃H₁₉D₂FN₄O₄; MW: [M + H]⁺, 439.1; Found 439.2

4-(Dideutero-{[6-(4-fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carbonyl]-amino}-methyl)-benzoic acid (AQU-019).

To a 250 ml round bottom flask containing 6 grams (13.68 mmoles) of 4-(Dideutero-{[6-(4-fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carbonyl]-amino}-methyl)-benzoic acid methyl ester (16) was added a stir bar and 90 ml of tetrahydrofuran (THF) and mixture stirred until solution was complete. To the solution was then added 6 ml of a 40% NaOD in D₂O solution. To the mixture was then added 50 ml of CH₃OD and solution stirred under nitrogen for 10 hours. To the mixture was then added concentrated hydrochloride acid until mixture was pH ~2. The volatile components of the reaction mixture were then removed under reduced pressure to give a solid which was purified by column chromatography (SiO₂, methylene chloride-MeOH, 9:1) to give 2.9 grams (50%) of the desired 4-(Dideutero-{[6-(4-fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carbonyl]-amino}-methyl)-benzoic acid (AQU-019) as a white solid. 1HNMR (300 MHz, d6-DMSO) δ 2.18 (s, 3H), 4.44 (d, 2H, J = 6.3 Hz)), 7.02–7.23 (m, 3H), 7.42 (d, 2H, J = 8.4 Hz), 7.87 (d, 2H, J = 8.4 Hz), 8.44, (s,1H), 8.46 (s, 1H), 9.45 (s, 1H), 9.65 (t, 1H, J = 6.0 Hz), 9.75 (s, 1H), HRMS calculated for C₂₂H₁₇D₂FN₄O₄; [M + H]⁺, 425.1514; Found 425.1510

4-({[6-(4-Fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carbonyl]-amino}-methyl)-benzoic acid (18).

This compound was synthesized and purified following the same manner as AQU-019 accept that commercially available methyl 4-(aminomethyl)benzoate was used in the coupling reaction instead of its deuterated derivative (10) to give to give after methyl ester deprotection and column purification the desired 4-({[6-(4-Fluoro-3-methyl-benzylcarbamoyl)-pyrimidine-4-carbonyl]-amino}-methyl)-benzoic acid (18) as a white solid. 1HNMR (300 MHz, d6-DMSO) δ 2.19 (s, 3H), 4.44 (d, 2H, J = 5.9 Hz), 4.57 (d, 2H, J = 5.9), 7.02–7.24 (m, 3H), 7.42 (d, 2H, J = 7.8 Hz), 7.88 (d, 2H, J = 7.8 Hz), 8.45, (s,1H), 8.46 (s, 1H), 9.46 (s, 1H), 9.67 (t, 1H, J = 6.6 Hz), 9.78 (s, 1H). 13CNMR (300 MHz, d6-DMSO), HRMS calculated for C₂₂H₁₉FN₄O₄; [M + H]⁺, 423.1390; Found 423.1391.

Pharmacology.

In-vitro rat microsomal stability.

Rat microsomal stability was determined for select compounds following the method of Houston (54). 1 μM concentration of compound and 0.3 mg/ml rat microsomes (BD bioscience) were used in the in-vitro assay. To ensure proper energy supply for microsomal degradation of compound, an energy regenerating system comprised of 100 mM potassium phosphate, 2mM NADPH, 3mM MgCl₂, pH = 7.4 and the microsomal protein is added to each sample and the resulting suspension is then incubated in duplicate for 60 min at 37° C. in a rotary shaker. A control is run for each test agent in duplicate omitting NADPH to detect NADPH-free degradation. At T=0 and T= 60 min., an aliquot is removed from each reaction and control reaction and then mixed with an equal volume of ice-cold Stop Solution (consisting of 0.3% acetic acid in acetonitrile containing haloperidol and diclofenac as internal standards). Stopped reactions are then incubated for at least ten minutes at –20 °C, and an additional volume of water is then added. The samples are then centrifuged to remove precipitated protein, and the supernatants are then analyzed by LC-MS/MS to determine the percentage of compound remaining. The LC-MS/MS system used was an Agilent 6410 mass spectrometer coupled with an Agilent 1200 HPLC and a CTC PAL chilled autosampler, all controlled by MassHunter software (Agilent), or an ABI2000 mass spectrometer coupled with an Agilent 1100 HPLC and a CTC PAL chilled autosampler, all controlled by Analyst software (ABI). After separation on a C18 reverse phase HPLC column (Agilent, Waters, or equivalent) using an acetonitrile-water gradient system, peaks were analyzed by mass spectrometry (MS) using ESI ionization in MRM mode. Verapamil (high metabolized) and Warfarin (low metabolized) are used as controls to test the activity of the microsomal proteins.

In-vitro measurement of MMP inhibition.

The MMP IC₅₀ was determined based on the change of rate of hydrolysis of a general MMP substrate: QXL520-γ-Abu-Pro-Cha-Abu-Smc-His-Ala-Dab(5-FAM)-Ala-Lys-NH2 (Smc= S-Methyl-L-cysteine) (Anaspec, cat# 60581–01). This 5-FAM/QXL™520-based fluorescence resonance energy transfer (FRET) substrate is a sensitive reagent for assaying MMP activity (55). It can be cleaved by MMP-1, 2, 3, 7, 8, 9, 12, 13 & 14 (human recombinant). This FRET peptide substrate incorporates QXL™520, a quencher available to pair with 5-FAM. When the peptide is intact, the fluorescence of 5-FAM (donor) is quenched by QXL™520 (“dark” acceptor) through FRET. Upon cleavage by MMPs into two separate fragments, the fluorescence of 5-FAM is recovered and can be detected at the emission wavelength of 520±20 nm, with excitation wavelength of 490±20 nm. A 1 μM DMSO stock solution of compound was made and diluted in a buffer composed of 50 μM HEPES (pH7.5), 10 μM CaCl2, 0.01% Brij-35, 0.1 mg/ml BSA and 5% DMSO at a concentration range of 1 to 100 μM. Fluorescence intensity was measured using a fluorescence spectrometer at every 5 min for 2 hours. Compound was tested in 12-dose IC₅₀ with 3-fold serial dilution starting at 100 μM in duplicate against human recombinant MMP-1, 2, 3, 7, 8, 9, 12, 13 & 14. A standard control compound, GM6001, was tested in 12-dose IC₅₀ with 3-fold serial dilution starting at 0.1 μM.

Assay Protocol:

1. Deliver 2X Enzyme

2. Deliver buffer into No Enzyme wells

3. Deliver compounds by using Acoustic technology (Echo550; nanoliter range)

4. Incubate for 10 min

5. Deliver 2X Substrate to initiate the reaction

6. Spin & shake, start measuring in EnVision brand Multimode Plate Reader at room temp; 5 min interval for 2 hours.

Data Analysis:

The protease activities were monitored as a time-course measurement of the increase in fluorescence intensity from fluorescently-labeled peptide substrate, and the initial linear portion of 2h measurement was taken as slope (FI signal/min) for further analysis. Therefore, time period taken for slope for analysis is different for each protease. Slope is calculated by using Excel, and curve fits are performed by using GraphPad Prism software.

In-vivo, single dose, oral rat pharmacokinetics (PK).

The basic study design and animal usage was approved by the Charles River Laboratory’s Institutional Animal Care and Use Committee (IACUC) for compliance with regulations prior to study initiation. Six (6) male Lewis rats (Charles River Laboratories) were used. Fasting was conducted at least 16 hours prior to dose administration. Food was returned at approximately 4 hours post dose. The animals were placed into 2 groups of 3 animals per group. The oral (PO) formulation for Groups 1 & 2 were prepared on the day of dosing at a target concentration of 0.5 mg/mL in 0.5% Methylcellulose (400 cps) to produce a white, homogeneous suspension. Each animal received compound (either AQU-019 or 18) by oral gavage administration at a target dose level of 1.0 mg/kg and at a dose volume of 2 mL/kg. Whole blood samples (0.250 mL; K₂EDTA anticoagulant) were collected from each animal through a jugular vein catheter. Whole blood samples were collected from all animals pre-dose, and at 0.25, 0.5, 1, 2, 4, 8, 16, and 24 hours after dose administration. All blood samples were immediately placed on ice until processing. Whole blood samples were centrifuged at 2200xg for 10 minutes in a refrigerated centrifuge (5±3°C) to isolate plasma. The plasma samples were transferred to individual polypropylene vials and immediately placed on dry ice before storage at nominally 20±5°C. The plasma samples were then thawed and extracted and analyzed by high pressure liquid chromatography (HPLC) coupled to Mass Spectrometry (MS). The HPLC system that was used was an Agilent 1200 Series Binary Pump, Leap CTC PAL autosampler, Supelco Discovery C18 column (50 x 2.1 mm), mobile phase: water (0.1% formic acid) and acetonitrile (0.1% formic acid); A 1.0 min gradient was utilized going from 1% to 98% of Mobile Phase B for a total run time of 2.40 minutes. The mass spectrometer used was an API 5000. Pharmacokinetic parameters were estimated from the plasma concentration-time data for each analyte using standard noncompartmental methods and utilizing the PK module of Watson Bioanalytical LIMS software (Version 7.2, Thermo Electron Corp).

In-vivo MIA rat model of OA via IA dosing.

The basic study design and animal usage was approved by the Bolder BioPATH’s Institutional Animal Care and Use Committee (IACUC) for compliance with regulations prior to study initiation. Twenty-nine male Sprague Dawley rats weighing 200–230 grams were obtained from Harlan Labs. Twenty-four Sprague Dawley rats were injected in the right knee with 2 mg of MIA in a 40 μl volume to induce cartilage damage. Five, twelve, and nineteen days after MIA injection, rats were dosed via intra-articular (IA) injection with 50 μl of Vehicle, AQU-019 (1 mg), or Triamcinolone (0.06 mg). Rats were terminated on day 26 and right knees were collected and prepared for histological evaluation. Specifically, the right knee joints from all animals were trimmed of muscle and connective tissue and collected into 10% neutral buffered formalin. The patella was removed to allow proper fixation of the joints. Following 1–2 days in formalin fixative and then 4–5 days in decalcifier. Preserved and decalcified knees were then trimmed into 2 approximately equal frontal halves, processed through graded alcohols and a clearing agent, infiltrated and embedded in paraffin, sectioned, and stained with Toluidine Blue (T Blue) and then examined microscopically by a board certified veterinary pathologist and observations were entered into a computer-assisted data retrieval system. Scoring of joints was done using the Osteoarthritis Research Society International (OARSI) grading system (56, 57). Approximate percent of total loss of articular chondrocytes, loss of proteoglycan and loss of collagenous matrix were determined for each articular surface (MTP, LTP, MFC, LFC) (56, 57). Chondrocyte loss was determined by estimating the area of cartilage in which there was no viable appearing chondrocytes. Proteoglycan loss was determined by estimating differences in intensity of toluidine blue matrix staining (56, 57). Collagen matrix loss was an indication of disruption and loss of type II collagen.

Statistical Analysis.

Data was analyzed using a one-way analysis of variance (1-way ANOVA) or Kruskal-Wallis test (non-parametric ANOVA), along with the appropriate multiple comparison post-test. Significance for all tests was set at p<0.05.

Figure 2. MIA rat model: Histological section of select rat knee joint showing the MTP, LTP, LFC and MFC.

A) Histological section of a rat knee from naïve rat group; B) Histological section of a rat knee joint from the Vehicle group after 4 weeks post MIA. (note severe loss of cartilage at the LFC and MFC) and C) Histological section of a rat knee joint from the AQU-019 group after four weeks post MIA (note less severe cartilage degradation at LFC and LTP compared to Vehicle).

• A novel, selectively deuterated pyrimidine dicarboxamide AQU-019, was synthesized.

• Methods for the selective deuteration of the benzyl position of the pyrimidine dicarboxamide 1 were examined.

• AQU-019 exhibited improved potency toward inhibiting MMP-13 and microsomal stability compared to the starting pyrimidine dicarboxamide 1.

• AQU-019 demonstrated selective inhibition for matrix metalloproteinase-13 (MMP-13) over that of MMP-1, MMP-2, MMP-3, MMP-7, MMP-8, MMP-9, MMP-10, MMP-12 and MMP-14.

• AQU-019 exhibited improved oral bioavailability in rats compared to its non-deuterated parent.

• AQU-019 was demonstrated to be chondroprotective when dosed via intraarticular (IA) injection in the monoiodoacetic acid (MIA) rat model of OA.