Discovery of Novel Orally Active Anti-Inflammatory N-Phenylpyrazolyl-N-Glycinyl-Hydrazone Derivatives That Inhibit TNF-α Production

Herein, we describe the synthesis and pharmacological evaluation of novel N-phenylpyrazolyl-N-glycinyl-hydrazone derivatives that were designed as novel prototypes of p38 mitogen-activated protein kinase (MAPK) inhibitors. All of the novel synthesized compounds described in this study were evaluated for their in vitro capacity to inhibit tumor necrosis factor α (TNF-α production in cultured macrophages) and in vitro MAPK p38α inhibition. The two most active anti-TNF-α derivatives, (E)-2-(3-tert-butyl-1-phenyl-1H-pyrazol-5-ylamino)-N’-((4-(2-morpholinoethoxy)naphthalen-1-yl)methylene)acetohydrazide (4a) and (E)-2-(3-tert-butyl-1-phenyl-1H-pyrazol-5-ylamino)-N’-(4-chlorobenzylidene)acetohydrazide (4f), were evaluated to determine their in vivo anti-hyperalgesic profiles in carrageenan-induced thermal hypernociception model in rats. Both compounds showed anti-inflammatory and antinociceptive properties comparable to SB-203580 used as a standard drug, by oral route at a dose of 100 µmol/kg. This bioprofile is correlated with the ability of NAH derivatives (4a) and (4f) suppressing TNF-α levels in vivo by 57.3 and 55.8%, respectively.


Introduction
The production of proinflammatory cytokines, e.g., TNF-a, IL-1b and IL-6, is a key factor in chronic inflammatory diseases, such as rheumatoid arthritis, Crohn's disease, psoriasis and asthma [1,2]. Moreover, evidence exists that supports the involvement of cytokines in other diseases, including cardiac heart failure, ischemic retinopathy [3] and the development of insulin resistance in diabetes [4]. Due to the role of cytokines in various inflammatory diseases, many pharmaceutical companies have made efforts to develop new orally active substances that can modulate the production of proinflammatory cytokines.
Tumor necrosis factor-alpha (TNF-a) is a pleiotropic cytokine that possesses proinflammatory and osmoregulator actions [5]. It is the major cytokine mediator of acute inflammation, it activates platelets, and it is also involved in the genesis of fever and anemia. TNF-a also mediates many inflammatory events in rheumatoid arthritis, including immune cell activation, proliferation, apoptosis and regulation of leukocyte movement [6], which has led to the development of strategies to block TNF-a-mediated effects. The currently available anti-TNF-a strategies involve either adminis-tration of anti-TNF-a antibodies or soluble TNF receptors to remove circulating TNF-a [7]. These inhibitors act by binding to TNF-a and preventing it from binding to its receptors on nearby cells, thus preventing the initiation of apoptosis or an inflammatory response [8].
Despite the approval of anti-TNF-a drugs, e.g., infliximab, etanercept and adalimumab, which demonstrated the effectiveness of therapeutic strategies based on the depletion of TNF-a, the appearance of side effects resulting from the debilitating actions of these drugs on the immune system highlights the necessity of identifying new alternative mechanisms to modulate the actions of pro-inflammatory cytokines [9,10].
One of the most promising targets involved in modulating the production of pro-inflammatory cytokines is the mitogen-activated protein kinase (MAPK) pathway, particularly p38 MAPK, a serine-threonine protein kinase that has been identified as a molecular target of the pyridinyl-imidazole derivatives SB-203580 (1) and SB-202190 (2) (Figure 1) [11,12]. These terphenylheterocyclic derivatives, which have been widely used to study p38 MAPK function, competitively bind at the ATP-binding pocket of p38 MAPK and inhibit TNF-a and IL-1b production.
Over the years, a large number of structurally diverse p38a and p38b MAPK inhibitors have been developed with both enhanced potency and specificity. Most of the p38 MAPK inhibitors are ATP competitors [13], but a new class of allosteric inhibitors has also been reported [14]. For example, BIRB-796 [15] (3) produces a conformational reorganization of the kinase that prevents ATP binding and activation.
In this context, the present work describes the synthesis of novel N-phenylpyrazolyl-N-glycinyl-hydrazone derivatives 4a-g, which were designed as structural analogues of the p38 MAP kinase inhibitor BIRB-796 (3), and the investigation of their anti-cytokine and anti-inflammatory properties. For the proposed derivatives (4a-g), we investigated the replacement of the urea subunit of BIRB-796 (3) by a N-acylhydrazone unit [16] (A', Figure 2), which was attached to the N-phenyl-pyrazole nucleus through an NHCH 2 spacer (B, Figure 2). Furthermore, we performed a series of molecular simplifications in the functionalized naphthyl framework attached to the imine unit of the NAH group of compound 4a to better understand the structure-activity relationships ( Figure 2).

Results and Discussion
The first step to obtain the N-phenylpyrazolyl-N-glycinylhydrazone derivatives 4a-g consisted of preparing the derivative 3-tert-butyl-1-phenyl-1H-pyrazol-5-amine (5) [15] from the condensation reaction between 4,4-dimethyl-3-oxopentanenitrile and phenylhydrazine (6) in refluxing toluene. The amino-pyrazole derivative 5 was subjected to alkylation with ethyl 2-bromoacetate in toluene and triethylamine under reflux, which gave rise to the corresponding amino ester 7 with a 60% yield. Next, the hydrazinolysis of the ester 7 with hydrazine hydrate in ethanol under reflux produced the corresponding hydrazide intermediate 8 with an 80% yield. The novel N-phenyl-pyrazolyl-N-glycinyl hydrazone derivatives 4a-g (Table 1) were then prepared in satisfactory yields through the acid catalyzed condensation of hydrazide 8 with aromatic aldehydes at room temperature ( Figure 3).
The structures of the N-phenyl-pyrazolyl-N-glycinyl-hydrazones 4a-g were completely characterized by common spectroscopic methods and the analytical results for C, H and N were within 60.4% of the calculated values.
According to the literature, N-acylhydrazones (NAHs) may exist as Z/E geometrical isomers about the C = N double bond and syn/anti amide conformers [17]. For most NAH derivatives described herein, the 1H-NMR spectra were recorded at room temperature, and they indicated the presence of two isomers, whereas only one species was detected by reversed-phase HPLC ( Figure S22). In a study involving compound 4 g, the 1H-NMR spectrum in DMSO-d6 at 90uC showed that the two isomers were in rapid equilibrium ( Figure 4A and Figure S13) [18]. Interestingly, complete coalescence of the signals was reached at 90C, and the reversibility of the changes was verified, indicating the presence of conformational isomers ( Figure 5). Moreover, the 1D NOESY showed spatial relationships of amide and imine hydrogens of compound 4 g that were compatible with the relative configuration (E) at the imine double bond (Figure S14, Figure S15 and Figure S23).
Another approach that was used to evaluate the presence of mixtures of conformers in our series of NAH derivatives 4a-g was based on the work of Wyrzykiewicz and Palla [17,18]. A 1 H-NMR spectrum of the compound 2-(3-tert-butyl-1-phenyl-1H-pyrazol-5ylamino)-N'-(propan-2-ylidene)acetohydrazide (9), which was obtained by a reaction of the previously obtained hydrazide 8 with acetone ( Figure 3), was performed because compound 9 cannot exist as E/Z geometrical isomers about the imine double bond. Nevertheless, the 1 H-NMR spectrum of compound 9 displayed duplicate signals for amide, methylene and pyrazole hydrogens, which completely coalesced at 90uC ( Figure 4B and Figure S18).
To evaluate whether the amino spacer exerts some influence on the stabilization of the conformational isomers in solution, we inserted a methyl group into the amino spacer, as described in Figure 6. The protection of the primary amine group [19] of compound 5 by treatment with acetic anhydride in acetic acid and sodium acetate resulted in the acetamide compound 10 with an 80% yield. Subsequent N-methylation was performed by deprotonation of compound 10 with NaH in THF followed by the addition of CH 3 I, which resulted in a 90% yield of Nmethylacetamide 11. The next step consisted of the removal of the protecting acetyl group to obtain the 3-tert-butyl-N-methyl-1phenyl-1H-pyrazol-5-amine (12), which showed a 90% yield. The alkylation of the monomethylamine derivative 12 with ethyl 2bromoacetate in ethanol and sodium carbonate under reflux provided the corresponding ethyl ester 13 with a 60% yield. Hydrazinolysis of the ester 13 followed by condensation of the corresponding hydrazide 14 with benzaldehyde under acid catalysis resulted in the desired N-acylhydrazone derivative 15 with a 60% yield.
The 1 H-NMR spectrum of the N-methyl derivative 15 showed the same pattern of duplicity that was observed for the other synthesized N-acylhydrazones 4a-g. We were also able to observe conformational isomers of the amide unit of compound 15, suggesting that the amino spacer does not participate as a hydrogen bond donor in the stabilization of the conformational isomers in solution ( Figure S20 and Figure S21).
We also performed the chemoselective N-alkylation of the Nacylhydrazone derivative 4g ( Figure 3) to investigate the influence of an alkyl group on the observation of conformational isomers in solution.
The pattern of duplication observed in the 1 H NMR spectrum of the N-acylhydrazone derivative 4g disappeared after methylation of the NAH framework, i.e., we did not observe conformational isomers for the corresponding N-methyl N-acylhydrazone derivative 4h. These results suggest that the insertion of the methyl group at the amide nitrogen leads to a steric or electronic effect that does not allow the distinction of conformational isomers in solution by 1 H-NMR ( Figure S19), as previously reported by Kummerle and co-workers [20].
Because the novel N-acylhydrazone derivatives 4a-g were designed based on the p38a MAPK inhibitor BIRB-796 (3), they were all evaluated for their in vitro capacity to inhibit p38a MAPK activity [23] at a concentration of 10 mM. Interestingly, only compounds 4b and 4e were active, and they inhibited approximately 30% of p38a activity (Table S1).
To evaluate the in vivo anti-inflammatory and antinociceptive profile of the NAH derivatives 4a, 4b, 4c and 4f, we employed the carrageenan-induced thermal hypernociception model [24]. Compounds were orally administered at a dose of 100 mmol/kg. SB-203580 (1) (100 mmol/kg, p.o.) was used as a standard. Figure 7 shows that compounds 4a and 4f were effective anti-hypernociceptive agents. Although these two compounds have shown similar capacities to inhibit TFN-a production in vitro (Table 2), compound 4a was more effective in vivo. In addition, compound 4a was able to completely inhibit the hypernociceptive response, whereas compound 4f was only able to partially inhibit this response.
We then investigated whether the inhibition of carrageenaninduced thermal hypenociception by 4a and 4f occurs through the inhibition of TNF-a. Four hours after carrageenan injection, the TNF-a level in the paw was elevated by more than two times that of the saline control. Interestingly, pretreatment with 4a and 4f (100 mmol/kg) suppressed the elevation of tissue TNF-a level by 57.3 and 55.8%, respectively ( Figure 8).
About the best anti-hypernociceptive profile of the compound 4a in comparison to derivative 4f, we decided to investigate the molecular reasons associated with a probable distinction in the  The analytical results for C, H and N were within 0.4% of the calculated values. [b] Yields obtained for the condensation step of hydrazide (8) with the corresponding aromatic aldehydes. doi:10.1371/journal.pone.0046925.t001 respective pharmacokinetic behaviors. The physicochemical property cLog P doesn't seems to explain the better in vivo profile of derivative 4a since both compounds, 4a and 4f, have the same theoretical lipophilicity, i.e. cLogP = 6.0 and 6.1, respectively. Considering that an adequate balance between the lipophilicity and aqueous solubility is essential for a good oral  absorption of a drug candidate, we decided to determine experimentally the solubility of compounds 4a and 4f in buffer solutions of pH 6.4 and 7.4 ( Figure 9). The derivative 4a, which contains the ethoxymorpholine-naphthyl group, exhibited an improvement in solubility at both pH values when compared with para-chlorophenyl derivative 4f, i.e. ca. 5 times at pH 7.4 and ca. 12 times at pH 6.4. As expected, at pH 6.4 only compound 4a showed to present an improvement in aqueous solubility (ca. three times), due to the partial ionization of its basic morpholine subunit. These solubility results allow us to rationalize that the improved in vivo activity of compound 4a is due to its better water solubility, which could favor its gastrointestinal absorption.
Moreover, we also evaluated the in vitro metabolic stability of derivatives 4a and 4f when placed in contact with preparations of liver and plasma of rats. The two NAH derivatives were resistant to oxidative microsomal metabolism, but the derivative 4a was about four times more resistant than derived 4f to plasma degradation, as described in Table 3. Taken together, these results indicate that the plasma stability associated to the better aqueous solubility are responsible for the better in vivo pharmacological profile shown by the NAH derivative 4a when given orally.
This study describes the synthesis and pharmacological evaluation of novel N-phenylpyrazolyl-N-glycinyl-hydrazone derivatives that were designed as novel prototypes of p38 MAPK inhibitors. All novel synthesized compounds described were evaluated for their in vitro capacity to inhibit TNF-a production in cultured macrophages and their in vitro p38a MAPK inhibition. The two most active anti-TNF-a derivatives were . These two compounds were evaluated for their in vivo anti-hypenociceptive profiles. Both compounds showed anti-inflammatory and anti-hypenociceptive properties that were comparable to SB-203580 (1), which was used as a standard.

Materials and Methods
Reactions were routinely monitored by thin-layer chromatography (TLC) in silica gel (F245 Merck plates) and the products visualized with ultraviolet lamp (254 and 365 nm). NMR spectra were recorded on a 200/50 MHz Bruker DPX-200, 250/ 62.5 MHz Bruker DPX-250, 400/100 MHz Varian 400-Mr, 300/75 MHz Varian Unity-300 spectrometer at room temperature. Peak positions are given in parts per million (d) from tetramethylsilane as internal standard, and coupling constant values (J) are given in Hz. Infrared (IR) spectra were obtained using a Nicolet Magna IR 760 spectrometer. Samples were examined as potassium bromide (KBr) disks. Elemental analyses were carried out on a Thermo Scientific Flash EA 1112 Series CHN-Analyzer. Melting points were determined using a Quimis instrument and are uncorrected and the compounds 4a-f had their melting points determined using a differential scanning calorimeter (Shimadzu DSC-60). Column chromatography purifications were performed using silica gel Merck 230-400 mesh. All described products showed 1 H and 13 C NMR spectra according to the assigned structures.
All organic solutions were dried over anhydrous sodium sulfate and all organic solvents were removed under reduced pressure in rotatory evaporator.
HPLC for purity determinations were conducted using Shimadzu LC-20AD with a SHIM-PACK CLC-ODS analytical column (4.6 mm 6 250 mm) or Kromasil 100-5C18 (4.6 mm 6 250 mm) and a Shimadzu SPD-M20A detector at 254 nm wavelength. The solvent systems for HPLC purity analyses was acetonitrile:phosphate buffer solution pH7 = 70:30. The isocratic HPLC mode was used, and the flow rate was 1.0 ml/min.

Procedure for Preparation of 3-tert-butyl-1-phenyl-1Hpyrazol-5-amine (5)
A round-bottomed flask charged with phenylhydrazine (0.83 mL; 8.39 mmol), 4,4-dimethyl-3-oxo-pentanenitrile (2.0g; 8.0 mmol) and toluene (3 ml) was stirred and heated at reflux for 24 hours. The resulting mixture was concentrated on a rotary evaporator and the residue was purified by column chromatography on silica gel (hexane/ethyl acetate, gradient), to yield the title compound (1.38g, 80%) as a white solid (mp: 50-52uC). 1  Procedure for the Preparation of Ethyl 2-(3-tert-butyl-1phenyl-1H-pyrazol-5-ylamino)acetate (7) To a solution of amine 5 (100 mg; 0.465 mmol) in toluene (3.0 mL) and trietylamine (0.1 mL), was added ethyl 2-bromoacetate (1.5 eq, 0.697 mmol, 0.077 mL). The resulting mixture was stirred and heated at reflux for 4 hours. The residue was partitioned between water and ethyl acetate. The combined organic phases were dried over Na 2 SO 4 , filtered, and concentrated. The brown residue was purified by silica gel chromatography hexane/ethyl acetate (gradient) to give the title compound (84 mg, 60%) as a brown oil. 1  Procedure for the Preparation of 2-(3-tert-butyl-1-phenyl-1H-pyrazol-5-ylamino)acetohydrazide (8) A round-bottomed flask charged with 600 mg (2 mmol) of ester (7), hydrazine hydrate 100% (20 eq) and ethanol (5 mL) was stirred and heated at reflux for 2 hours. To the resulting mixture was added water and the aqueous phase was extracted with ethyl acetate to give the title compound (430 mg, 80%) as a yellow oil. General Procedure for the Preparation of N-phenylpyrazole N-glycinyl-N-acylhydrazones (4a-g) In a round flask containing hydrazide 8 (1.6 mmol) in ethanol (10 mL), was added aldehyde (1.68 mmol; 1.05 eq) and catalytic concentrated hydrochloric acid. The mixture was stirred for about 2 hours at room temperature. At the end of the reaction the volume of ethanol was reduced, saturated solution of sodium bicarbonate and ice were added to the reaction. The precipitate formed was filtered, or the mixture was extracted with dichloromethane.  Figure S1]. 13 Table 2. Effects of N-phenylpyrazolyl-N-glycinyl-hydrazone derivatives 4a-g on TNF-a production and cell viability in murine peritoneal macrophages.

LPS-induced TNF-a Production in Culture of Mice Peritoneal Macrophage
BALBc mice were stimulated with thyoglicollate 3% (1 mL/ mice; i.p.) and 3 days later the peritoneal cavity were washed with RPMI 1640 and the peritoneal macrophages were plated onto 96wells plate (30.000 cells/well) for 1 hour at 37uC in an humidified 5% CO 2 atmophere. Then, macrophages were incubated with the vehicle or compounds and 1 hour later stimulated with LPS (100 ng/mL) for 24 hour when the supernatants were collected to evaluate TNF-a production by ELISA kit (BD Bioscience).

Cell viability by MTT Assay
The peritoneal macrophage were obtained and plated as described above. The cells were incubated with the vehicle or compounds for 20 hours when was added 20 ml of MTT (3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) solution (5 mg/mL), followed by 4 hour of incubation. Then, the culture mediums were collected and the precipitates were solubilized in 200 ml of DMSO. The optical density was measured at 490 nm.

Carragenaan-induced Hypernociception Assay
Wistar rats of both sexes (150-200g) were used. The compounds were administered orally (100 mmol.5mL-1.kg-1) as a suspension in 5% arabic gum in saline (vehicle). Control animals received an equal volume of vehicle. One hour later, the animals were injected with either 0.1 ml of 1% carrageenan solution in saline or sterile saline (NaCl 0.9%), into the subplantar surface of one of the hind paws. The thermal hypernociception was determined using the modified hot-plate test. Rats are placed individually on a hot plate with the temperature adjusted to 51uC. The latency of the withdrawal response of the left hind paw is determined at 0, 30, 60, 120, 180, and 240 min post-challenge. The time of maximum permanence permitted on the hot surface is 20 s. Hypernociception to heat is defined as a decrease in withdrawal latency and calculated as follows: D paw withdrawal latency (s) = (left paw withdrawal latency at time 0) -(left paw withdrawal latency at the others times).
The paw was homogenized 4 h after intraplantar injection of carrageenan, and the level of TNF-a in the supernatant was determined by ELISA.

Solubility
The solubility was evaluated after twenty-four hour agitation of 1 mg of test compound in 1 mL aqueous buffer (pH 6.4 and pH 7.4) at 37 uC followed by centrifugation and filtration for HPLC-UV analysis.

Rat liver Microsomal Stability Studies
The incubation was conducted at 37uC for 60 min. The experiments contains MgCl 2 (1.3 mM), NADP+ (0.4 mM), glucose-6-phosphate (3.5 mM), 0.5 U/mL glucose-6-phosphate dehydrogenase in a phosphate buffer (0.1 M, pH 7.4) containing EDTA (1.5 mM) and the test compounds were added at final concentration of 50 mM with 0.25 mL of final volume. After the pre-warming of the mixture at 37 uC, the microsomal proteins were added to give a final protein concentration of 1 mg/mL. At the end of the incubation time the reaction was stopped by the addition of 375 mL of MeOH and 375 mL of CH 3 CN. The experiments were performed in duplicate. The samples were centrifuged and filtered for HPLC-UV analysis.

Rat Plasma Stability Studies
The rat plasma was obtained from blood by centrifugation and diluted in phosphate buffer (pH 7.4). The test compounds were added at final concentration of 50 mM with 0.25 mL of final volume and incubated at 37uC for 60 min under agitation. At the end of the incubation time the reaction was stopped by the addition of 375 mL of MeOH and 375 mL of CH 3 CN. The experiments were performed in duplicate. The samples were centrifuged and filtered for HPLC-UV analysis.
The detection was carried out at 285 nm and 330 nm wavelength for compound 4f and 4a, respectively.

Statistical Analysis
Data obtained from experiments were expressed as mean 6 S.E.M., compared with vehicle control groups and statistically analyzed by the ANOVA one-way (Bonferroni post hoc test) for carrageenan-induced thermal hyperalgesia and Student's t test or for the others experiments. In all cases p,0.05 was considered significant (*p,0.05; **p,0.01; ***p,0.001). When appropriate, the IC 50 values (i.e. the concentration able to inhibit 50% of the maximum effect observed) were determined by non-linear regression using GraphPad Prism software v. 5.0.