Adamantyl Analogues of Paracetamol as Potent Analgesic Drugs via Inhibition of TRPA1

Paracetamol also known as acetaminophen, is a widely used analgesic and antipyretic agent. We report the synthesis and biological evaluation of adamantyl analogues of paracetamol with important analgesic properties. The mechanism of nociception of compound 6a/b, an analog of paracetamol, is not exerted through direct interaction with cannabinoid receptors, nor by inhibiting COX. It behaves as an interesting selective TRPA1 channel antagonist, which may be responsible for its analgesic properties, whereas it has no effect on the TRPM8 nor TRPV1 channels. The possibility of replacing a phenyl ring by an adamantyl ring opens new avenues in other fields of medicinal chemistry.


Introduction
Although paracetamol (4-hydroxyacetanilide, acetyl-p-aminophenol, acetaminophen, APAP) [1][2][3] was found to be an effective analgesic more than a century ago, its mechanism of action is complex and the subject of continuous research mainly due to its extensive metabolism in animals and humans [4].
Several reports have described pathways of APAP to exert its analgesic activity. Unlike non-steroidal anti-inflammatory drugs (NSAIDs), whose analgesic and anti-inflammatory effects are related to their inhibition of the cyclooxygenase enzymes (COX-1 and COX-2), paracetamol is a weak anti-inflammatory agent with an absence of COX-related adverse effects [5][6]. In the brain and spinal cord paracetamol is metabolized by fatty amide hydrolase (FAAH) to N-arachido- noylphenolamine (AM404). AM404 is a known activator of the capsaicin receptor (TRPV1) [7] and the cannabinoid CB 1 receptor system [8][9] both of which confer analgesia in the central nervous system. The main drawback found in the use of paracetamol is the mechanism-based inactivation (MBI) of cytochrome P450 enzymes (CYPs) [10]. APAP metabolites (e.g. N-acetyl-p-benzoquinoneimine (NAPQ1) have centered the attention because of their toxic actions causing hepatotoxicity [11][12]. Many articles have been devoted to this topic [13][14][15][16][17] and different attempts have been made to circumvent this problem. For instance, 3hydroxyacetanilide is non hepatotoxic because the oxygen atoms of quinones can only be at 1,2-,ortho, or, 4-positions, para [18]. Other authors have designed new analgesics derived from the paracetamol metabolite (AM404) having an anandamide chain instead of the acetamido group [19]. The replacement of the methyl group by saccharin or an open form of it (SCP-1 and SCP-123) improves considerably its water solubility [20]. A proline prodrug was developed with this purpose [21] and a paracetamol analogue known as propacetamol is currently in the market. Note that paracetamol presents polymorphism [22][23]. Whereas researchers have been trying to avoid the formation of the metabolite responsible for APAP toxicity designing non-hepatotoxic paracetamol analogues [24], a new antinociceptive mechanism driven by activation of TRPA1 in the spinal cord by APAP metabolites such as NAPQ1 has been reported [25].
In August 2013, the FDA warned of rare but serious skin reaction associated with paracetamol. Despite the fact that it has been on the market for decades, paracetamol can be considered a ''standalone'' drug of which no effective analogues are known [26]. Some compounds which have been reported include the N-methyl derivative, and N-(1H-indazolo-5-yl), N-(4-hydroxy)acetamides and recently pyridinol-fused ring derivatives, but in all of them, the aromatic ring is present [27][28][29].
The approach proposed in the present paper is totally different being based on the hypothesis that the replacement of the phenyl ring of paracetamol by an adamantane ring with the same substituents in 1,4-positions, will yield analogues of paracetamol. Adamantyl derivatives are found in medicinal chemistry in very different domains [30][31][32]. Adamantane is one of the few singular hydrocarbon moieties that have been successfully employed in pharmaceutical industry. The most popular ''hit'' is amantadine used for the treatment of influenza and Parkinson's disease [33]. Related to analgesia and/or inflammation some analogues of D 8 -THC such as compound 1 have been reported (Fig. 1) [31][32]. Quinolone 2 binds selectively to the cannabinoid CB 2 receptor being endowed with analgesic activity in vivo [34]. Some derivatives from 3 reported as P2X 7 receptor antagonists acted as agents against rheumatoid arthritis [35][36]. Amongst the carbohydrazides, the most interesting is 4 developed by Abbott [37]. As can be seen, none of the structures represented in Fig. 1 is related to our compounds.
In this work we report the synthesis and biological evaluation of compounds 5, 6a and 6b (Fig. 2) in which the phenyl ring has been substituted by an adamantyl.
Compounds 6a and 6b were synthesized following the route summarized in Fig. 4 [40]. These compounds were obtained after a reductive amination and in three steps from the commercially available 5-hydroxy-2-adamantanone (8). The synthesis was carried out using the pure enantiomer of the S-a-methylbenzylamine and the 5-hydroxy-2-adamantanone refluxing in absolute ethanol. The intermediate compound 9 was hydrogenated. This hydrogenation with the catalyst Pd/C 10% could have been performed in two steps, isolating the benzylamino derivatives, but it was preferred an overnight hydrogenation to yield compounds 10a and 10b. The last synthetic step consisted of the acetylation in water of the mixture using the same conditions as for compound 5 yielding an E/Z mixture of compounds 6a and 6b in a 7:3 ratio, with the E isomer being the major product. Compounds 6a and 6b were found in a paper by Gonzáles-Nuñez et al. [41] but with a different enriched mixture (Z:E 75/25) and not separated for fully   characterization. In our case, compounds 6a and 6b were purified by automated semipreparative LC-MS for their characterization.
The 13 C and 1 H NMR spectroscopy in DMSO-d 6 of the mixture of 6a and 6b is interesting. The signals were assigned taking into account that 6a was the major isomer (about 70%) and by means of 2D experiments. To check the assignments as well as the isomerism, we calculated the absolute shielding [42][43] (s, ppm) at the GIAO/B3LYP/6-311+G(d,p) level and transform the absolute shielding into chemical shifts (d, ppm, see Table S1).

Pharmacology
The analgesic activity of compounds 5 and the 6a/b mixture was compared to that of paracetamol and morphine in a model of visceral pain (Fig. 5). The intraperitoneal (i.p.) administration of acetic acid (2%) induced a typical writhing reaction, characterized by a wave of contraction of the abdominal musculature followed by extension of the hind limbs. After acetic acid administration, mice were placed in individual transparent containers and, after 5 min, the number of writhes was counted during a 10 min period. In naïve animals, the number of stretches induced by acetic acid was 25.4¡0.8. Morphine (5 mg/kg, i.p.), an opioid accepted as the gold standard of analgesic drugs, notably decreased the number of writhes by 75%. Paracetamol (100 mg/kg), administered i.p.  30 minutes before acetic acid, was less effective than morphine reducing only by 40% the nociceptic behavior. Furthermore, a 2-fold increment in the paracetamol dose did not result in an increase of its analgesic effect in this pain model (Fig. 5).
Akin to the reference analgesic agents, compounds 5 and 6a/b were administered i.p. 30 minutes before the injection of acetic acid (Fig. 5). The solubility of these derivatives was significantly higher than paracetamol allowing their testing at higher doses, which is a desirable property. Both compounds (5 and 6a/b) were able to induce a significant antinociceptive effect that was more potent for compound 6a/b. As illustrated in Fig. 5, compound 5 displays a similar 40% antinociceptive activity to that recorded with paracetamol. This analgesic effect of compound 5 displayed some dose-dependency and saturated at 200 mg/ kg. In contrast, compound 6a/b exhibited stronger analgesic activity than compound 5, and a clear dose-dependent analgesic activity that saturated at 750 mg/kg. At saturation, compound 6a/b displays an analgesic efficacy significantly higher than paracetamol and similar to morphine. Taken together, these results suggest that compound 6a/b is a more potent analgesic than paracetamol.
We next investigated the mechanism of action underlying the analgesic activity of analogue 6a/b. First, we evaluated a possible action on cannabinoid receptors. For this purpose, compound 6a/b was administered 30 minutes before the selective CB1 AM251 (3 mg/kg i.p.) and CB2 AM630 (1 mg/kg, i.p) antagonists, and the writhing test was carried out 30 minutes after the administration of the antagonists. Neither the CB1 nor the CB2 antagonist was able to block the antinociceptive activity of compound 6a/b (number of stretches 13¡1 (AM251) and 15.5¡1.7 (AM630). Thus, the cannabinoid system does not appear to play a significant role in the antinociception induced by the paracetamol derivatives 5 and 6a/b.
An alternative mechanism that we evaluated was the potential inhibition of cyclooxygenase enzymes COX-1 or COX-2. Our findings indicate that, while paracetamol inhibited COX-2 with an IC 50 of 7.08¡1.62 mM, compound 6a/b did not affect the enzymatic activity at 10 mM. This result implies that inhibition of COX enzymes does not underlie the in vivo analgesic activity of compound 6a/b.
Because of the pivotal role of thermoTRP channels in pain transduction [44][45][46], we hypothesized that the analgesic activity of compound 6a/b may be due to a direct inhibition of some these channels. Among the thermoTRPs, TRPM8 (''melastatin''), TRPV1 (''vanilloid'') and TRPA1 (''ankyrin'') are the most validated in pain signaling. TRPV1 is considered a molecular integrator of noxious heat stimuli in nociceptors [47], TRPM8 is a pivotal sensor for cold stimuli, and TRPA1 is a unique sensor of noxious environmental stimuli [45][46][47][48]. Therefore, we selected these channels to evaluate if any of them was the target of compound 6a/b (Fig. 6).
The channels were stably expressed in eukaryotic cells, and a Ca 2+ fluorographic assay used to monitor their activity upon instillation of their respective agonists in the absence and presence of compound 6a/b at 50 mM. As illustrated in Fig. 7a, only the activity of the TRPA1 channels was selectively blocked up to 85% at this concentration of 6a/b. No significative effect was recorded for the other thermoTRPs. A dose-response curve reveals that compound 6a/b displayed an IC 50 of 2.6 mM, indicating that compound 6a/b is a moderate antagonist of TRPA1 channels (Fig. 7b). The inhibitory activity of compound 6a/b was further demonstrated electrophysiologically as evidenced by the blockade of the AITCevoked ionic currents ( Figures S1 and S2 in the presence and in the absence of extracellular Ca 2+ ). Collectively, these findings indicate that TRPA1 is a molecular target of compound 6a/b, and could participate in the antinociceptive effect showed in the writing test.

Conclusions
We have described the synthesis and pharmacological evaluation of new paracetamol analogs derived from an adamantane scaffold. Compounds 5 and 6a/ b represent attractive leads to be developed since they show an improved antinociceptive effect compared to paracetamol. In addition, adamantane derivatives have proved to be very biocompatible, so possible toxic effects due to chronic treatment should not be expected.
The main result of the present communication is that phenyl ring, ubiquitous in medicinal chemistry, in some cases could be replaced by an adamantyl ring without loss but improvement of the biological properties. To the ortho, meta and para positions of a phenyl ring correspond to 1,2, 1,3 and 1,4-substituents on an adamantyl ring. This is in agreement with our findings since the 1,4-derivative, with a similar substitution to that of paracetamol, has shown greater potency than the 1,3-derivative.
Even though more pharmacological research is needed (using TRPA1 agonists for example), compound 6a/b, an analogue of paracetamol, able to block TRPA1 channel, is an interesting, new, antinociceptive drug.  recorded on Bruker Advance 300 spectrometer operating at 300.13 MHz and 75.47 MHz respectively, in CDCl 3 or DMSO-d 6 as solvents. Chemical shifts are reported in ppm on the d scale. In the case of multiplets, the signals are reported as intervals. Signals were abbreviated as s, singlet; d, doublet; t, triplet; and m, multiplet. Coupling constants are expressed in hertz. Hydrogenation reactions were carried out in a Shaker type hydrogenation apparatus (Parr). Elemental analysis was determined with a LECO Elemental Analyzer CHNS-932. LC-MS analyses were performed using an Alliance 2695 (Waters) with a diode array UV/ Vis detector Waters 2996 and interfaced to a Micromass ZQ mass spectrometer. Analyses were performed using reversed phase HPLC silica based columns: column Bridge C18 3.5 mm (2.16100 mm). Using an injection volume of 3 mL, a flow rate of 0.25 mL/min and gradient elution (2 to 30% over 15 min) of acetonitrile in water. Acetonitrile contains 0.08% v/v formic acid and water contains 0.1% v/v formic acid. Analyses were monitored at 254 nm wavelength.

Pharmacological evaluation
The animals, CD-1 male mice (25-27 g), were redistributed in groups, two of which were used as the control group. The remaining twelve groups were treated with paracetamol 100 and 200 mg/kg (two different groups), compound 5 (100, 200, 500 mg/kg) and compound 6a/b (250, 375, 500, 750 mg/kg). An additional group received morphine 5 mg/kg. All drugs were intraperitoneally (i.p.) administered, 30 minutes before the algogen agent. Control mice were treated with 10 ml/kg of saline solution. Each group included 10-12 animals; those animals showing behavioral alterations were previously discarded. Each animal was used only once. An observer who was unaware of the different treatments carried out the collection of data.
The experimental test used was the acetic acid writhing test. The effect of the i.p. administration of 10 ml/kg of acetic acid (2%) was measured by quantifying the number of writhing responses in the three groups of animals over a 10 minutes period. Responses were measured starting 5 minutes after the algogen agent was administered. The animals used were supplied by the animal house of Universidad Rey Juan Carlos. Room temperature was kept at 22¡1˚C. A 12 hours light -dark cycle was started at 8:00 a.m., at constant humidity. All animals were given free access to food and water.
Experimental protocols used in this investigation were approved by the Ethical Committee of Rey Juan Carlos University and were conducted in accordance with the guidelines of the International Association for the Study of Pain [49].

Inhibition of Isolated Cyclooxygenases COX-1 and COX-2
A COX inhibitor screening assay (kit 700100) was used to determine the activity of isolated ovine COX-1 and human recombinant COX-2 as described by the manufacturer (Cayman Chemical Company, USA).

Calcium fluorography
For fluorescence assays, cells expressing TRP channels (TRPV1-SH-SY5Y, TRPM8-HEK and TRPA1-IMR90) were seeded in 96-well plates (Corning Incorporated, Corning, NY) at a cell density of 40,000 cells 2 days before treatment. The day of treatment the medium was replaced with 100 mL of the dye loading solution Fluo-4 NW supplemented with probenecid 2.5 mM. Then the compounds dissolved in DMSO were added at the desired concentrations and the plate(s) were incubated at 37˚C in a humidified atmosphere of 5% CO 2 for 60 minutes.
The fluorescence was measured using instrument settings appropriate for excitation at 485 nm and emission at 535 nm. (POLARstar Omega BMG LAB tech). A baseline recording of 7 cycles was recorded prior to stimulation with the agonist (10 mM capsaicin for TRPV1, 100 mM menthol for TRPM8, and 100 mM AITC for TRPA1). The corresponding antagonist (10 mM Ruthenium Red for TRPV1 and TRPA1, 100 mM AMTB for TRPM8) was added for the blockade. The changes in fluorescence intensity were recorded during 15 cycles more. DMSO, at the higher concentration used in the experiment, was added to the control wells.
The degree of blockage (%) of TRP channel activity was calculated by: Where: F 0 is the fluorescence after the addition of agonist in the presence of the compound, F I is the fluorescence before the addition of agonist in the presence of the compound, Fc 0 is the fluorescence after the addition of agonist in the absence of the compound, Fc I is the fluorescence before the addition of agonist in the absence of the compound.
The Z factor was calculated using the following equation: Where: Mean max is the mean of the maximum fluorescence in the presence of agonist, Mean min is the mean of the maximum fluorescence in the presence of agonist and antagonist.

Computational details
The geometries of compounds 5, 6a and 6b were optimized at the B3LYP/6-31G(d) level and the corresponding frequencies (all 0) proved that they are minima. These geometries were reoptimized at the B3LYP/6-311++G(d,p) level. Absolute shieldings (s, ppm) and SSCC (Hz) were calculated on these geometries using the GIAO approximation for the s values. The corresponding chemical shifts (d, ppm) were obtained using the empirical transformation equations we have established for a large collection of compounds. The used equations transform calculated values in the gas phase to experimental values in solution (usually CDCl 3 or DMSO-d 6 ). See supplementary data. Figure S1. Compound 6a/b strongly inhibits TRPA1-mediated currents in the presence of extracellular Ca 2+ . A) Representative whole-cell voltage clamp recording from hTRPA1-expressing IMR-90. Currents were measured every second during a holding potential of -60 mV. Pre-application of 6a/b (20 s) was followed by co-application with 100 mM AITC for 60s in presence of 2.0 mM extracellular Ca 2+ . Current traces of different colours denote the different concentrations of compound 6a/b tested. B) Dose response of compound 6a/b blockade activity. Solid line depicts the fitting to a Michaelis isotherm. The estimated IC 50 value was 10.6¡0.7 mM. Data are given as mean ¡ sem, with n>4 cells) per data point. doi:10.1371/journal.pone.0113841.s001 (TIF) Figure S2. Compound 6a/b weakly inhibits TRPA1 currents in the absence of extracellular Ca 2+ . A) Representative whole-cell voltage clamp recording from hTRPA1-expressing IMR-90 cells in the absence of extracellular Ca 2+ . Ionic currents, at positive and negative potentials were measured every 2s during a 350 ms voltage ramp from 260 mV to +60 mV and evoked with 100 mM AITC followed by the addition of 100 mM compound 6a/b for 20s. AITC was present with drug application and after application to evaluated current reversibility B) Current-voltage relationships of TRPA1 in the absence and presence of 100 mM compound 6a/b in Ca 2+ free medium. The inferred IC 50 value was >80 mM, as saturation of blockade was not reached in the absence of Ca 2+ . Data were obtained from n>4 cells. doi:10.1371/journal.pone.0113841.s002 (TIF)