Potent multitarget FAAH-COX inhibitors: Design and structure-activity relationship studies

Non-steroidal anti-in ﬂ ammatory drugs (NSAIDs) exert their pharmacological effects by inhibiting cyclooxygenase (COX)-1 and COX-2. Though widely prescribed for pain and in ﬂ ammation, these agents have limited utility in chronic diseases due to serious mechanism-based adverse events such as gastrointestinal damage. Concomitant blockade of fatty acid amide hydrolase (FAAH) enhances the therapeutic effects of the NSAIDs while attenuating their propensity to cause gastrointestinal injury. This favorable interaction is attributed to the accumulation of protective FAAH substrates, such as the endocannabinoid anandamide, and suggests that agents simultaneously targeting COX and FAAH might provide an innovative strategy to combat pain and in ﬂ ammation with reduced side effects. Here, we describe the rational design and structure-active relationship (SAR) properties of the ﬁ rst class of potent multitarget FAAH-COX inhibitors. A focused SAR exploration around the prototype 10r (ARN2508) led to the identi ﬁ cation of achiral ( 18b ) as well as racemic ( 29a - c and 29e ) analogs. Absolute con ﬁ gurational assignment and pharmacological evaluation of single enantiomers of 10r are also presented. ( S )-( þ )- 10r is the ﬁ rst highly potent and selective chiral inhibitor of FAAH-COX with marked in vivo activity, and represents a promising lead to discover novel analgesics and anti-in ﬂ ammatory drugs.


Introduction
Non-steroidal anti-inflammatory drugs (NSAIDs) are widely utilized to treat pain and inflammation [1], but their chronic use is hindered by a variety of potentially serious adverse events that include gastrointestinal (GI) mucosal lesions, bleeding and perforations [2e5]. Conventional NSAIDs inhibit the two isoforms of cyclooxygenase (COX), COX-1 and COX-2, which catalyze the first committed steps in the biosynthetic pathway that converts arachidonic acid (AA) into inflammatory prostanoids such as prostaglandin E 2 (PGE 2 ) and thromboxane A 2 (TXA 2 ) [6]. The dual role of COX-1-derived PGE 2 as inflammation promoter and mucosal tissue protectant explains, at least in part, why NSAIDs cause damage to the GI tract [7e10]. Efforts to overcome this problem have led to the development of selective COX-2 inhibitors, which combine a high level of anti-inflammatory efficacy with a reduced propensity to cause injury to the GI mucosa [6]. Nevertheless, the use of COX-2 inhibitors has been linked to a distinctive set of adverse cardiovascular effects [11,12]. Thus, the need for safe and effective drugs that can be used in the treatment of chronic inflammatory disorders remains urgent.
A promising approach to meet this need is offered by targeting with a single agent more than one component of the inflammatory cascade [13e15]. Agents designed to achieve this objective include nitric oxide (NO) donors-NSAIDs [16,17], COX-2 inhib-itorseNOedonors [18,19], hydrogen sulfide (H 2 S) donors-NSAIDs [20e22], as well as compounds that block distinct enzymes of the AA pathway, such as COX/lipoxygenase [23,24] and COX-2/soluble epoxy hydrolase (sEH) [25]. Another potential multitarget strategy to treat inflammation is the concomitant inhibition of COX and fatty acid amide hydrolase (FAAH) [26] [27e33], a serine hydrolase that deactivates a family of analgesic and anti-inflammatory lipid amides that are produced by host-defense cells and other cells in the body [34,35]. These lipid mediators include the endocannabinoid anandamide (arachidonoylethanolamide) e which engages cannabinoid-1 (CB 1 ) and CB 2 receptors to suppress neutrophil migration [36] and prevent immune-cell recruitment [37,38] e as well as the endogenous peroxisome proliferator-activate receptor-a (PPAR-a) agonists, palmitoylethanolamide (PEA) and oleoylethanolamide (OEA) [39e41]. In addition to opposing pain and inflammation, these FAAH substrates are also protective of the GI mucosa [42,43]. Indeed, studies in animal pain models have shown that co-administration of FAAH and COX inhibitors results in a synergistic potentiation of analgesia along with reduced gastric damage [44e46].
In several chronic inflammatory conditions, including inflammatory bowel disease (IBD), FAAH [47e49] and COX-2 [50] are expressed at abnormally high levels. This simultaneous upregulation may help establish a pathological state that exacerbates inflammation by amplifying inflammatory COX-dependent signals at the expense of defensive FAAH-regulated mediators. This hypothesis predicts that drugs targeting both COX and FAAH should have substantial anti-inflammatory efficacy combined with reduced GI toxicity. In a recent study, we provided support to this hypothesis using a multitarget modulator based on the hybrid scaffold 1 (Fig. 1) [51]. This scaffold merges key pharmacophores of two known classes of FAAH and COX inhibitors e O-aryl carbamates [52e58] such as [3-(3-carbamoylphenyl)phenyl] N-cyclohexylcarbamate (URB597, 2) [54,57], and 2-aryl propionic acids [6] such as flurbiprofen, 3a [59e61] e which share a biphenyl core as a common structural motif (A and B rings, Fig. 1). Moreover, structure-activity relationship (SAR) studies of these scaffolds supported the hypothesis of additional elements of structural overlapping, such as the oxygenated substituents at the 3 0 -position of the A phenyl ring, corresponding to the carbamate functionality of 2 [53,54,56] and the ether moieties of 3b or 3c [61], respectively ( Fig. 1).
This SAR work led to the identification of compound 10r ((±)-2-[3-fluoro-4-[3-(hexylcarbamoyloxy)phenyl]phenyl]propanoic acid, ARN2508) [51] as a potent in vivo active inhibitor of intracellular FAAH and COX activities, which exerts profound anti-inflammatory effects in mouse models of IBD without causing COX-dependent gastric toxicity [51]. In the present study, (a) we outline the indepth SAR investigations that led to the discovery of compound 10r [51]; (b) we report an expansion of this SAR work, which culminated in the identification of several new and potent multitarget inhibitors (18b, 29a-c and 29e); and, finally (c) we describe the absolute configurational assignment and pharmacological properties of single enantiomers of 10r, identifying (S)-(þ)-10r as the first chiral inhibitor of FAAH-COX with marked in vivo activity.

Chemistry
Compounds 10a-t were synthesized from the corresponding phenol 8 through a carbamoylation reaction, using commercially available isocyanates, followed by the hydrolysis of the methyl esters 9a-t, under acidic conditions (Scheme 1).
The intermediate 8 was prepared in four steps, starting from the acid 4, obtained as previously described [62]. Compound 4 was converted to the corresponding methyl ester 5, under standard acidic conditions, to afford, after catalytic hydrogenation with ammonium formate in the presence of Pd/C, the resulting aniline 6. Compound 6 was then transformed into the corresponding diazonium salt, that was reacted in situ with NaI to obtain the phenyl iodide 7 in good yield, which was converted, under ligand less Suzuki cross coupling conditions [63], to the biphenyl derivatives 8 and 13a-c in excellent yield (Schemes 1e3).
3-Hydroxypropyl derivative 12 was synthesized by reduction of the methyl ester 8 to the alcohol 11 (Scheme 1). Although lithium aluminum hydride succeeded in reducing the ester 8, a significant des-fluorinated side product was observed and separation of the two compounds was troublesome. Therefore, a milder reducing agent, such as zirconium borohydride generated in situ, was used to afford a clean conversion of 8 to 11 [64], which was then converted to 12 under standard carbamoylation reaction conditions (Scheme 1).
Carbamates 15a-b and urea 15c were prepared from the corresponding phenols 13a-b and aniline 13c, respectively, through a carbamoylation reaction using n-hexyl-isocyanate, followed by acidic hydrolysis of the methyl esters 14a-c (Schemes 2 and 3). The reverse carbamate 15d was prepared upon activation of the aniline 13c with triphosgene, and, then, reaction with n-hexanol, followed by acidic hydrolysis of the methyl ester 14d (Scheme 3).
Compounds 18a and 21a-b were synthesized by reacting the phenyl iodides 16b and 19a-b, with (3-hydroxyphenyl)boronic acid under Suzuki cross coupling conditions, followed by carbamoylation reaction of phenols 17 and 20a-b under standard conditions (Schemes 4 and 5).
Compounds 18a and 21b were then transformed into the corresponding acids 18b and 21c under standard acidic hydrolysis (Schemes 4 and 5).
Compounds 29a-g were synthesized following the synthetic sequence described in Scheme 6 p-Nitrofluorobenzenes 22ad were reacted with diethyl methylmalonate followed by decarboxylation to the corresponding acids 23a-d. 23a-d and the commercially available 23e were converted into methyl esters 24ae in acidic MeOH. In addition, the phenolic intermediate 24d was directly converted into the corresponding O-Bn protected 24f, Fig. 1. Rational design of a 0 hybrid scaffold 0 for FAAH and COX inhibition. under standard reaction conditions. Reduction of the nitro group was carried out using iron in presence of HCl for compounds 24a and 24f, and ammonium formate in the presence of Pd/C for compounds 24b-c and 24e. Compound 25f was obtained from 25e by standard nitration reaction. Diazotation/Sandmayer reaction of the anilines 25a-f gave the iodides 26a-f, which were converted to carbamates 28a-f via Suzuki and carbamoylation reactions. Compounds 28a-f were then transformed into the corresponding acids 29a-f under standard acidic hydrolysis. Finally, the aniline 29g was obtained from the nitrophenyl 29f, under standard hydrogenation conditions. We started our SAR exploration with compound 10a [51], which was designed by merging essential pharmacophores of the FAAH inhibitor, URB597, 2, and those of the NSAID, flurbiprofen, 3a (Fig. 1). The inhibitory potencies of 2, 3a and 10a against rat brain FAAH, ovine testis COX-1 and human recombinant COX-2 are reported in Table 1.
Compound 10a inhibited FAAH and COX activities with relatively weak potencies (IC 50 values, in mM: COX-2 > 100). Nevertheless, these initial results encouraged us because 10a was one of the most potent FAAH/COX-1 inhibitors previously reported [27,28,30,32,33,65]. We started, therefore, an SAR exploration around 10a with the objective of identifying chemical and structural determinants that might improve potency on the three targets in a balanced manner.

Study of the effect of the nature of R group: cycloalkanes, small-branched alkanes and phenyls
We prepared a series of analogs bearing cycloalkyl groups with different ring size at the N-terminal of the carbamate functionality (Table 1).
We observed that, while the potency against FAAH was retained with the c-pentyl analog 10b (IC 50 ¼ 4.8 mM), a 10-fold loss in potency occurred with the c-butyl derivative 10c (IC 50 ¼ 48.7 mM) and complete loss of activity (IC 50 > 100 mM) with the c-propyl derivative 10d. With regard to COX activity, while the c-pentyl analog 10b showed a comparable potency against COX-1 (IC 50 the c-butyl analog 10c was 10-fold more potent than compound 10a (IC 50 ¼ 0.72 mM). Conversely, the c-propyl analog 10d displayed an IC 50 value similar to compounds 10a and 10b against COX-1 (¼5.4 mM) and was indeed the only compound in this series that showed modest activity against COX-2 (IC 50 The N-terminal region of the carbamate functionality in 10a may engage in beneficial interactions with the acyl chain-binding domain of FAAH [26] [56,57], as well as the hydrophobic channels present in COX-1 and COX-2 [6]. [61] To capture such interactions, we prepared a series of analogs bearing lipophilic aliphatic and aromatic N-terminal substituents with diverse steric properties  ( Table 1).
The insertion of a methylene group adjacent to the c-hexyl ring of 10a -compound 10e-led to a significant increase of potency toward FAAH (23-fold) and COX-1 (10-fold), but no COX-2 inhibition (IC 50 > 100 mM). A further homologation, compound 10f, showed a 400-fold increase in potency toward FAAH and a 50-fold increase in potency toward COX-1, compared to 10a. Interestingly, 10f also inhibited COX-2 with an IC 50 of 10.8 mM.
Next, we investigated the effects of small and branched alkyl groups, the iso-propyl 10g and the iso-butyl 10h -as truncated analogs of 10a and 10e, respectively. These modifications were detrimental for FAAH and COX inhibitory activities compared to 10a and 10e, respectively.
While the replacement of the c-hexyl ring with a phenyl group (10i) was not tolerated by FAAH, in analogy to previous reports on the class of O-aryl carbamates [56,57], this modification led to a gain in inhibitory activity toward COX-1 and COX-2. The insertion of a methylene group adjacent to the phenyl ring of 10i -compound 10j-caused a 10-fold increase in potency toward FAAH, compared to 10i, but had almost no impact on COX-1 activity and dramatic loss on COX-2. Homologation (10k-m) resulted in a progressive enhancement of the inhibitory potency toward FAAH, but this trend was more erratic for COX-1 and COX-2: compound 10l was most active analog with IC 50 ¼ 0.58 mM and 6.2 mM against COX-1 and COX-2, respectively.
These findings might reflect differences in the depth of lipophilic pockets of FAAH and COX enzymes [6,26].

2.2.3.
Study of the effect of the nature of the R group: linear alkanes. Identification of 10r (ARN2508) Since the (CH 2 ) n homologation at the N-terminal site appeared to be critical for the modulation of the biological activities at both targets, we prepared a series of carbamates bearing linear alkyl groups (alkyl ¼ (CH 3 (CH 2 ) n ) with n ¼ 1 to 7) at N-terminal region ( Table 2).
In analogy to the reported SAR results on the class of O-aryl carbamates [56], potency toward FAAH increased with increased length of the (CH 2 ) n chain (n ¼ 1e7). A different trend was observed for COX-1 and COX-2, where insertion of short (CH 2 ) n chains (n ¼ 1e2) led to compounds (10n-o) that were weak COX-1 inhibitors and had no activity against COX-2. On the other hand, insertion of n ¼ 3e5 (CH 2 ) n chains (10p-r) increased the inhibitory potencies for COX-1 and COX-2 from sub-micromolar to nanomolar IC 50 , whereas insertion of n ¼ 6e7 (CH 2 ) n chains (10s-t) was detrimental.
These results are in agreement with those above reported in the homologation of the Ph(CH 2 ) n chain series (n ¼ 1e4, compounds 10i-m, Table 1).
From this SAR exploration, we identified 10r (ARN2508) [51], which bears a n-hexyl chain at the N-terminal site, as a potent multitarget inhibitor of FAAH, COX-1 and COX-2 (IC 50 : Table 2). In addition to its high balanced potency, the highest reported thus far [27,28,30,32,33,65], we found that 10r displays no offetarget activities on a panel of >90 biologically relevant targets, and effectively engages its intended targets after oral administration in mice [51].
These results encouraged us to initiate a more focused SAR exploration to define the effect of additional chemical and structural modifications in various regions of 10r scaffold.
2.2.4. Focused SAR exploration around 10r (ARN2508) and identification of 18b, 29a-c, e and (S)-(þ)-10r In particular, we focused our interest on the role and position of carbamate group in the A phenyl ring (Table 3 and Table 4), as well as the role of the propionic acid functionality and the fluorine atom in the B phenyl ring (Table 5 and Table 6).

Role and position of carbamate group in the A phenyl ring.
We first investigated the effect of the position of the carbamate group in the A phenyl ring, which indeed appeared to play an important role in the inhibition of both FAAH and COX (Table 3). In agreement with the rational design of our hybrid scaffold 1 (Fig. 1), the C(2 0 )-derivative 15a (ortho derivative) showed a 70-fold decrease in potency toward FAAH, a 60-fold decrease in potency toward COX-1, and a complete loss of activity toward COX-2, when compared to the C(3 0 )-isomer 10r (meta derivative) ( Table 3).
On the other hand, the C(4 0 )-derivative 15b (para derivative) exhibited a slight loss of potency toward FAAH compared to 10r, but both COX inhibitions were completely suppressed (Table 3). These results support the hypothesis that the bent shape of the Obiphenyl moieties, which is known to better fit the FAAH enzyme surface [53], is also important in the recognition by COX-1 and COX-2, possibly through a better superimposition to the conformations adopted by the fatty acyl chain of the natural substrate/product (the first two cis-double bonds of AA) when bound to COX-1 [66] and COX-2 [67].
Next, we replaced the carbamate moiety with alternative functional groups, such as urea (15c) [51] and reversed carbamate (15d) [51] (Table 4). As expected from the rational design of our class of multitarget inhibitors, 15c and 15d showed a significant decrease in potency toward FAAH, whilst retaining COX-1 and COX-2 inhibitory activities compared to 10r.
These results support the hypothesis that the mechanism of action of this class of compounds is similar to the one reported for the O-aryl carbamates (acylation of FAAH Ser 241) [57] and that COX inhibition does not rely on any irreversible binding mode at the expense of the carbamate group of 10r. Reported dialysis experiments on 10r are in agreement with this mechanistic speculation [51].

2.2.4.2.
Role of the propionic acid functionality in the B phenyl ring. We then turned our attention to the role of the propionic acid in the B phenyl ring (Table 5).
Replacing the propionic acid group of 10r with several substituents had only a minor impact on the potency toward FAAH, compared to the effect observed on COX activities. In fact, methyl ester 9r retained FAAH inhibitory activity, compared to 10r, but completely lost activity toward both COX-1 and COX-2. Replacement of the carboxylic acid of 10r with the corresponding primary alcohol 12 resulted in a 10-fold improvement in potency toward FAAH (IC 50   COX-1 (IC 50 ¼ 2.1 mM and 12 nM, respectively). The activity against COX-2 was slightly improved (IC 50 ¼ 0.24 mM and 0.43 mM respectively). The methyl analog 21a [51] was active against FAAH in the same potency range of 10r (IC 50 ¼ 26 nM and 31 nM, respectively), while a completely loss of activity against COX enzymes was observed. A similar result was obtained with the carboxylic analog 21c, which also showed a 3-fold reduction in potency toward FAAH, compared to 10r (IC 50 ¼ 85 nM and 31 nM, respectively).
We conclude that FAAH tolerates substituents with different steric and electronic properties at the 4-position of the B phenyl ring, while COX-1 and COX-2 display a stringent requirement for a propionic or acetic acid groups in the same position.  (Table 6).
Substituting the fluorine with chlorine was tolerated: indeed, 29a was virtually equipotent against FAAH and COX-1, and marginally less potent on COX-2, compared to 10r. The same trend was observed with the methyl derivative 29b, which was slightly more potent than 10r against FAAH and equally potent on COX-1, but less active against COX-2. The CF 3 derivative 29c showed a 6fold and 2-fold increase in potency toward FAAH and COX-2, respectively, and was as potent as 10r on COX-1.
Removal of the fluorine atom (29e) resulted in a 10-fold increase in potency toward FAAH, compared to 10r, and a slight decrease in activity for COX-1 and COX-2.
Compounds 29d and 29g, which beareOH oreNH 2 groups, respectively, inhibited FAAH with potencies similar to that of 10r, whereas a clear loss in potency for both COX-1 and COX-2 was observed. On the other hand, the NO 2 derivative 29f had higher Table 2 SAR exploration on the nature of the R group: linear alkanes.

Table 3
Effect of the position of the carbamate functionality on the A phenyl ring.     potency toward FAAH but loss lower potency toward both COX-1 and COX-2. We interpret these results to suggest that the electronic and steric properties of the substituents in the 3-position of the B phenyl ring affect FAAH recognition only slightly, whereas these same substituents influence COX-1 and COX-2 more markedly, with lipophilic groups being better tolerated than polar or Hbond donator groups.

2.2.4.4.
Stereochemical and pharmacological studies of 10r enantiomers. Finally, we subjected the best studied member of this class of inhibitors, the racemic compound 10r [51], to chiral HPLC separation and tested each of its enantiomers e (À)-10r (first eluted) and (þ)-10r (second eluted) e for the ability to inhibit FAAH, COX-1 and COX-2 (Table 7). FAAH showed no preference for either enantiomer, with each being more active than the racemate 10r. By contrast, in analogy to prior studies on different classes of FAAH/COX inhibitors [30,33], substantial differences were observed on COX-1 and COX-2. Compound (þ)-10r was highly potent on both COX-1 (IC 50 ¼ 0.29 nM) and COX-2 (IC 50 ¼ 50 nM), whereas (À)-10r was weakly active on either target. We completed our exploration on the two enantiomers of 10r by assigning their absolute stereo-configurations. As reported in Supporting Information (Scheme S1), a stereochemical correlation study allowed us unambiguously to assign the absolute stereochemistry of (À)-10r and (þ)-10r to the (R)-and (S)-configurations, respectively. These results are in agreement with earlier reports showing that the (S)-enantiomer is responsible for the COXinhibiting activity of aryl-propionic acid derivatives such as flurbiprofen [29,31,33,59].

In vivo experiments on (S)-(þ)-10r.
Finally, pharmacological experiments indicate that compound (S)-(þ)-10r strongly engages its intended molecular targets in live mice. Intravenous administration of (S)-(þ)-10r (1 mg/kg) lowered the concentrations of two COX products in circulation, prostacyclin and TXA 2 , as assessed surveying the stable metabolites, 6-keto-PGF 1a and TXB 2 ( Fig. 2A and B). Moreover, (S)-(þ)-10r increased plasma levels of the FAAH substrate, OEA (Fig. 2C). In addition, (S)-(þ)-10r demonstrated no offetarget activities on a panel of >90 biologically relevant receptors, enzymes [including N-acylethanolamine amide hydrolase (NAAA), which is the primary enzyme involved in the deactivation of PEA and OEA in innate immune cells] and ion channels (Table S1). Further pharmacological studies on the (R)and (S)-series of this class of inhibitors will be reported in due course.

Conclusions
The present study outlines key SAR properties of a novel class of dual inhibitors of intracellular FAAH and COX activities, which are based on the hybrid scaffold 1. Several chemical variations of this scaffold were considered, which involved the carbamate moiety at the 3 0 -position of the A phenyl ring, the R groups, and the propionic acid moiety and fluorine atom in the B phenyl ring. Introduction of different alkyl and aromatic groups in the N-terminal region of the carbamate functionality improved inhibitory potency toward both FAAH and COX. A more focused exploration around the potent, selective and orally available racemic inhibitor 10r [51] led to the identification of novel potent analogs, 29a-c, and e. Because of the problems associated with the development of racemic compounds, we extended our studies and identified two additional molecules, the achiral compound 18b and the enantiomer (S)-(þ)-10r, which also display high inhibitory potency for FAAH/COX-1/COX-2.
The in vivo activity of (S)-(þ)-10r suggests that this agent may be used to probe the therapeutic utility of simultaneous FAAH-COX inhibition, especially in pathologies in which these enzymes are abnormally expressed.

Synthesis
Solvents and reagents were obtained from commercial suppliers and were used without further purification. URB597 was prepared following a reported procedure [54]. Flurbiprofen was purchased from SigmaeAldrich (Milan, Italy). Melting points were determined on a Büchi MÀ560 capillary melting point apparatus and are uncorrected. Automated column chromatography purifications were done using a Teledyne ISCO apparatus (CombiFlash ® Rf) with prepacked silica gel columns of different sizes (from 4 g until 120 g). Mixtures of increasing polarity of Cy and EtOAc or DCM and MeOH were used as eluents. Preparative TLC analyses were performed using MachereyeNagel pre-coated 0.05 mm TLC plates (SIL G-50 UV 254 ). 1 H and 13 C NMR experiments were run on a Bruker Avance III 400 system (400.13 MHz for 1 H, and 100.62 MHz for 13 C), equipped with a BBI probe and Z-gradient coil. 19 F NMR experiments were run on a Bruker Avance III 600 system (546.6 MHz for 19 F), equipped with a 5 mm CryoProbe QCI 1 H/ 19 Fe 13 C/ 15 NeD quadruple resonance and a Z-gradient coil. Spectra were acquired at 300 K, using deuterated dimethylsulfoxide (DMSO-d 6 ) or deuterated chloroform (CDCl 3 ) as solvents. Chemical shifts for 1 H and 13 C spectra were recorded in parts per million using the residual non-deuterated solvent as the internal standard (for DMSOd 6 : 2.50 ppm, 1 H; 39.52 ppm, 13 C; for CDCl 3 : 7.26 ppm, 1 H and 77.16 ppm, 13 C). Data are reported as follows: chemical shift (ppm), multiplicity (indicated as: bs, broad signal; s, singlet; d, doublet; t, triplet; q, quartet; p, quintet, sx, sextet; m, multiplet and combinations thereof), coupling constants (J) in Hertz (Hz) and integrated Table 7 Evaluation of the enantiomers of 10r. Compound Compound 4 was obtained as brown clear oil (4.50 g, 81%), according to the procedure reported in the literature starting from 2,4-difluoronitrobenzene (4.77 g, 30 mmol) [62].
To a solution of 4 (4.50 g, 21.11 mmol) in MeOH (40 mL), concentrated H 2 SO 4 (0.1 mL) was added and the resulting solution was stirred at rt overnight. After solvent evaporation, the crude oil was diluted with Et 2 O (15 mL) and filtered through a pad of SiO 2 to afford 5 as orange-brown oil (4.45 g, 93%).

(±)-Methyl 2-(4-amino-3-fluoro-phenyl)propanoate (6)
To a solution of 5 (12.60 g, 55.46 mmol) in MeOH (222 mL) was added 10% Pd/C (2.35 g, 2.22 mmol) followed by addition of HCO 2 NH 4 (20.98 g, 332.8 mmol). The solution was stirred at rt for 3 h, then, filtered through a pad of Celite and the filtrate was concentrated under reduced pressure. The residue was dissolved in EtOAc and filtered through a pad of SiO 2 to afford 6 as an orange oil (10.33 g, 94%). (7) A solution of NaNO 2 (0.70 g, 10.21 mmol) in H 2 O (1.5 mL) was added slowly to a solution of 6 (1.75 g, 9.76 mmol) in a 3N HCl solution (29 mL) at 0 C. After 30 min, NaI (1.54 g, 10.25 mmol) was added at 0 C under stirring. The resulting mixture was slowly warmed to rt in 5 min, and then heated at 60 C for 3 h. After cooling down to rt, the mixture was extracted with Et 2 O and the organic phase was then washed with a 1 M solution of Na 2 SO 3 (15 mL) and dried over Na 2 SO 4 . The residue was dissolved in EtOAc (50 mL), treated with activated carbon and then filtered through a pad of Celite. The filtrate was concentrated under reduced pressure and the yellow oil was purified by column chromatography (Cy: EtOAc, 95:5) to give 7 as a pale yellow oil (1.70 g, 55%).

(±)-2-(3-chloro-4-nitro-phenyl)propanoic acid (23a)
Step 1: To a solution of 22a (4.70 g, 27.0 mmol) and diethyl methylmalonate (4.13 mL, 25.0 mmol) in DMF (31 mL), NaOH (1.11 g, 28 mmol) was added. The mixture was stirred at rt for 15 h. The dark red solution was poured into ice, acidified with concentrated HCl (4 mL) and extracted with TBME. The organic solvent was washed with brine, dried over Na 2 SO 4 and evaporated under reduced pressure to give orange oil (8.24 g) which was used for the next step without any further purification.
Step 2: H 2 O (25 mL), AcOH (38 mL) and H 2 SO 4 (13 mL) were added to the orange oil (8.24 g, 25 mmol) and the reaction mixture was refluxed for 24 h. AcOH was removed under reduced pressure, and the mixture was extracted with DCM. The organic layer was then extracted with a saturated aqueous Na 2 CO 3 solution and the aqueous layer was acidified with 1N HCl, and extracted with DCM. The organic layer was dried over Na 2 SO 4 and evaporated to give 23a as an orange oil (5.00 g, 87%).

(±)-Methyl 2-(3-chloro-4-nitro-phenyl)propanoate (24a)
23a (5.00 g, 21.78 mmol) was dissolved in MeOH (27 mL), H 2 SO 4 (58 mL, 1.09 mmol) was added and the mixture was stirred for 15 h. The solvent was removed under reduced pressure and the residue taken up in TBME, activated carbon was added and then the mixture passed through an allumina pad. The solvent was removed under reduced pressure to give 24a as a yellow oil (4.57 g, 86%).

(±)-Methyl 2-(4-amino-3-chloro-phenyl)propanoate (25a)
Iron powder (4.19 g, 75 mmol) was added to a solution of 24a (4.57 g, 18.76 mmol) in MeOH/HCl (7:1, 40 mL). The mixture was refluxed for 2 h, then filtered through a pad of Celite. The solvent was removed under reduced pressure and taken up in H 2 O, the thick slurry was basified with K 2 CO 3 , EtOAc was added, filtered through a pad of Celite and the two phases separated. The organic layer was dried over Na 2 SO 4 and evaporated to give 25a as an orange oil (2.57 g, 64%).

(±)-2-(3-methyl-4-nitro-phenyl)propanoic acid (23b)
Step 1: To a solution of 22b (3.26 mL, 26.75 mmol) and diethyl methylmalonate (4.59 mL, 25 mmol) in DMF (31 mL), NaOH (1.11 g, 27.75 mmol) was added. The mixture was stirred at rt for 15 h. The dark red solution was poured into ice, acidified with concentrated HCl (10 mL) and extracted with Et 2 O. The organic solvent was washed with brine, dried over Na 2 SO 4 and evaporated under reduced pressure to give yellow oil (7.73 g) which was used for the next step without any further purification.
Step 2: H 2 O (25 mL), AcOH (41 mL) and H 2 SO 4 (12 mL) were added to the oil (7.73 g, 25 mmol) and the mixture was refluxed for 24 h. AcOH was removed under reduced pressure and the product was extracted with DCM and washed with brine. The organic layer was treated with aqueous K 2 CO 3 , and the separated aqueous layer was acidified with concentrated HCl, extracted with DCM, washed with brine and dried over Na 2 SO 4 . After removal of the solvent 23b was obtained as brown clear oil (2.50 g, 48%) which was used for the next step without any further purification.

(±)-Methyl 2-(3-methyl-4-nitro-phenyl)propanoate (24b)
To a solution of 23b (2.5 g, 11.95 mmol) in MeOH (120 mL), H 2 SO 4 (0.22 mL, 1.2 mmol) was added and the mixture was stirred at rt for 15 h. The solvent was removed under reduced pressure and the residue was dissolved in Et 2 O and passed through a pad of alumina. The solvent was evaporated to obtain 24b as a yellow oil (2.17 g, 81%).

(±)-Methyl 2-(4-iodo-3-methyl-phenyl)propanoate (26b)
A solution of NaNO 2 (0.71 g, 10.24 mmol) in H 2 O (2 mL) was added slowly to a solution of 25b (1.85 g, 9.57 mmol) in 2N HCl (43 mL) at 0 C. After stirring for 30 min a solution of NaI (2.15 g, 14.36 mmol) was added dropwise and the mixture was allowed to reach rt and stirred for 2 h, then warmed to 60 C for other 2 h Na 2 SO 3 was added and the product was extracted with Et 2 O, dried over Na 2 SO 4 and evaporated. The residue was purified by column chromatography (Cy/EtOAc, 95: 5) to give 26b as clear oil (1.14 g, 39%).

(±)-2-[4-nitro-3-(trifluoromethyl)phenyl]propanoic acid (23c)
Step 1: To a solution of 22c (3.74 mL, 26.75 mmol) and diethyl methylmalonate (4.13 mL, 25 mmol) in DMF (30 mL), NaOH (1.11 g, 27.75 mmol) was added. The mixture was stirred at rt for 15 h. The mixture was stirred at rt for 15 h. The dark red solution was poured into ice, acidified with concentrated HCl (4 mL) and extracted with TBME. The organic solvent was washed with brine, dried over Na 2 SO 4 and evaporated under reduced pressure to give orange oil (9.1 g) which was used for the next step without any further purification.
Step 2: H 2 O (25 mL), AcOH (37 mL) and H 2 SO 4 (12 mL) were added to the orange oil and the mixture was refluxed for 24 h. AcOH was removed under reduced pressure, and the mixture was extracted with DCM. The organic layer was then extracted with a saturated aqueous Na 2 CO 3 solution and the aqueous layer was acidified with 1N HCl, and extracted with DCM. The organic layer was dried over Na 2 SO 4 and evaporated to give 23c as an orange oil (5.59 g, 85%).

(±)-Methyl 2-[4-nitro-3-(trifluoromethyl)phenyl] propanoate (24c)
23c (5.59 g, 21.24 mmol) was dissolved in MeOH (28 mL), H 2 SO 4 (58 mL, 1.06 mmol) was added and the mixture was stirred for 19 h. The solvent was removed under reduced pressure and the residue taken up in TBME, activated carbon was added and then the mixture passed through an allumina pad. The solvent was removed under reduced pressure to give 24c as an orange oil (5.70 g, 97%).

(±)-Methyl 2-[4-iodo-3-(trifluoromethyl)phenyl]propanoate (26c)
A solution of NaNO 2 (1.48 g, 21.38 mmol) in H 2 O (4 mL) was added slowly to a solution of 25c (4.94 g, 19.98 mmol) in 2N HCl (90 mL) at 0 C. After stirring for 30 min a solution of NaI (4.49 g, 29.97 mmol) was added dropwise and the mixture was allowed to reach rt and stirred for 2 h, then warmed to 60 C for other 2 h Na 2 SO 3 was added and the product was extracted with TBME, dried over Na 2 SO 4 and evaporated. The residue was purified by column chromatography (Cy: EtOAc, 95: 5) to give 26c as clear oil (5.51 g, 77%).

(±)-2-[4-[3-(hexylcarbamoyloxy)phenyl]-3-hydroxyphenyl]propanoic acid (29d)
To a solution of 28d (0.72 g, 1.49 mmol) in EtOH (29 mL), Pd/C (78 mg, 74 mmol) and cyclohexene (9 mL, 88 mmol) were added and the mixture was stirred at 80 C for 2 h. The catalyst was filtered through a pad of Celite and the solvent was removed under reduced pressure. The residue oil was taken up in dioxane (15 mL), 2 M HCl (15 mL) was added and the solution was stirred at 80 C for 15 h. The solvent was removed under reduced pressure and the residue was purified by column chromatography (DCM/MeOH, 98: 2) to obtain 29d as a white solid (414 mg, 73%). Mp: 61e62 C. 1  To a solution of 23e (1.95 g, 10 mmol) in MeOH (20 mL), concentrated H 2 SO 4 (0.1 mL) was added and the resulting solution was stirred overnight at rt. After solvent evaporation, the crude oil was diluted with Et 2 O (15 mL) and filtered through a pad of SiO 2 to afford 24e as yellow oil (2.10 g, quant.)

(±)-Methyl 2-(4-aminophenyl)propanoate (25e)
To a solution of 24e (1.05 g, 5 mmol) in MeOH (20 mL) was added 10% Pd/C (0.37 g, 0.35 mmol) followed by the addition of HCO 2 NH 4 (1.9 g, 30 mmol). The solution was stirred at rt for 1 h. The solution was filtered through a pad of Celite and the filtrate was concentrated under reduced pressure. The residue was dissolved in EtOAc and filtered through a pad of SiO 2 to afford 25e as an offwhite solid (0.89 g, quant.).

(±)-Methyl 2-(4-iodophenyl)propanoate (26e)
A solution of NaNO 2 (0.69 g, 10 mmol) in H 2 O (1.5 mL) was added slowly to a solution of 25e (1.75 g, 9.76 mmol) in 28 mL of 3N HCl at 0 C. After stirring for 1 h at 0 C, NaI (1.50 g, 10 mmol) was added. The resultant mixture was slowly warmed to rt for 5 min, and heated at 60 C for 2 h. After cooling down to rt, the mixture was extracted with Et 2 O and the organic phase was then washed with a 1 M solution of Na 2 SO 3 (20 mL), dried over Na 2 SO 4 . After evaporation, the residue was dissolved in EtOAc (40 mL) and treated with activated carbon and filtered through a pad of Celite. The solvent was removed under reduced pressure and the orange oil was purified by column chromatography (Cy/EtOAc, 95:5) to give 26e as a clear oil (2.05 g, 72%).

(±)-Methyl 2-(4-amino-3-nitro-phenyl)propanoate (25f)
Step 1: A solution of 25e (3.58 g, 20 mmol) in Ac 2 O (100 mL) was heated at 130 C for 1 h. The solution was poured into H 2 O, stirred for 3 h, then evaporated. The residual solid was taken up in H 2 O and filtered under vaccum to obtain a yellow solid. This solid was dissolved in MeOH (100 mL) and 37% HCl (5 mL) was added. The solution was stirred for 2 h and the organic solvent was removed under reduced pressure. H 2 O was added and the precipitate was filtered under vacuum and washed with H 2 O to obtain methyl 2-(4acetamidophenyl)propanoate as a cream colored solid (2.21 g, 50%).
Step 3: To a solution of (±)-methyl 2-(4acetamido-3-nitro-phenyl)propanoate (2.60 g, 9.77 mmol) in MeOH (98 mL) H 2 SO 4 (10 mL, 183 mmol) was added and the mixture was stirred at reflux for 2 h. MeOH was evaporated under reduced pressure and the solution was carefully poured into a aqueous solution of Na 2 CO 3 (2 M, 120 mL), then extracted with EtOAc, dried over Na 2 SO 4 and evaporated to give 25f as a dark orange oil (2.20 g, quant.) which was used in the next step without further purification.

(±)-2-[3-amino-4-[3-(hexylcarbamoyloxy)phenyl]phenyl] propanoic acid hydrochloride (29g)
To a solution of 29g (0.87 g, 2.10 mmol) in MeOH (21 mL), Pd/C (223 mg, 0.21 mmol), cyclohexene (5.32 mL, 52.48 mmol) were added and the solution was stirred at 80 C for 2 h. The mixture was filtered through a pad of Celite and the solvent removed under reduced pressure to give a residue which was purified by column chromatography (DCM/MeOH, 96: 4) to obtain a glassy oil, which was dissolved in dioxane (10 mL) and concentrated HCl (1 mL) was added. The solvent was removed under reduced pressure and the residue oil was suspended in DCM and Et 2 O. The solid was filtered under vacuum to obtain 29g as an off-white solid (488 mg, 55%).