All the solvents used were analytically pure. The reagents 5-isopropyl-2-methylphenol (carvacrol), chlorosulfonic acid, morpholine, 4-fluoroaniline, pyridin-2-yl methanamine, 2-hydroxyaniline, 2,4-dichloroaniline were obtained from Sigma Aldrich.

The synthesis sulfonamides S1–S5, as already described in the literature (de Oliveira et al., 2016) was performed in two steps: firstly, the synthesis of 4-hydroxy-2-isopropyl-5-methylbenzene-1-sulfonyl chloride (ChS) was performed, subsequently, the ChS was used in reactions with different amines (Scheme 1). ChS was obtained from the reaction of carvacrol to six equivalents of chlorosulfonic acid. The sulfonamides obtained in this study were prepared from ChS with two equivalents of amine added slowly. Reactions were followed by thin layer chromatography (TLC). All sulfonamides were purified by acid-base extraction and the compounds were duly characterized by spectroscopic and spectrometric techniques.

Results are reported as average values ± standard deviation (SD) with the exception of ID₅₀ and EC₅₀ values, which were calculated from single experiments using nonlinear regression implemented in GraphPad 7.0 (GraphPad software, San Diego, CA, United States). The glutamate test with paw edema measurement and the open-field test showed a normal data distribution in line with the Shapiro–Wilk threshold (p = 0.05) and, thus, were submitted to one-way ANOVA analysis and to Dunnett test for multiple analyses. Only p-values below 0.05 were taken as significant (p < 0.05).

The antioxidant activity of the sulfonamides derived from carvacrol (Table 1) was analyzed by the NO and H₂O₂ scavenging activity assays.

The electronic properties were directly correlated with the antioxidant activity of the molecules. The E_HOMO and E_LUMO indicate the molecule’s ability to donate and receive electron density, respectively. The difference between the two energy levels is termed the band gap and gives an estimate of the reactivity of a molecule. The distance between the HOMO and LUMO energy levels is inversely proportional to the reactivity the compound. The HOMO and LUMO potentials and band gap of the carvacrol derivatives are shown in Figure 2.

Figure 3 shows the correlation between HOMO energy and experimental EC₅₀ ^NO and EC₅₀ ^H2O2. The correlation coefficients r ² and Person's coefficient (r) of the EC₅₀ ^NO versus E^HOMO were 0.87 and 0.93, respectively. For EC₅₀ ^H2O2 versus E^HOMO , r ² and r were 0.88 and 0.94, respectively. The angular coefficient values of the equations EC₅₀ ^NO = EC₅₀ ^NO (E^HOMO) and EC₅₀ ^H2O2 = EC₅₀ ^H2O2 (E^HOMO) were, respectively, 10.28 ± 2.03 and 11.30 ± 2.10 µmol. (L.eV)⁻¹. In addition, the linear coefficients were 87.90 ± 14.49 and 96.63 ± 14.96 µmol. (L.eV)⁻¹, respectively. From these equations, the minimal values of E^HOMO (i.e., EC₅₀ ^NO = EC₅₀ ^H2O2 = 0) can find the maximal activity. Thus, with the HOMO energy tending to −8.55 eV for both equations, the maximal elimination of NO and H₂O₂ is reached for both experiments.

For a better understanding of the antinociceptive effect of sulfonamides derived from carvacrol (S1–S5), we used the model of glutamate-induced (i.pl.) nociception. This method allowed us to investigate the possible interaction of peripheral antinociceptive action of the analyzed compounds with the glutamatergic system. The results are shown in Figure 4.

Figure 5 shows the results of treatment with carvacrol and its derivatives on paw edema induced by glutamate (i.pl.). Our results show that only S1 and S5 were able to significantly reduce edema. However, S1 inhibited edema more effectively and dose-dependently. The percent inhibition values were: 36 ± 10%, 47 ± 6%, and 73 ± 12% for S1 at 0.0003, 0.003 and 0.03 mg/kg i.g., respectively; 19 ± 9%, 33 ± 6%, and 28 ± 7% for S5 at 0.0003, 0.003, and 0.03 mg/kg i.g., respectively. The value of ID₅₀ for compound S1 was 0.002 (0.0009–0.005) mg/kg. Furthermore, the calculated values for the ID₅₀ antiedematogenic effect of S1 (0.002 mg/kg) agree with the dose found in the glutamate test, showing homogeneity of the data in this group. Thus, we suggest that S1 may be an interesting target for the reduction of edema in inflammatory conditions.

Figure 6 shows that intragastric administration of carvacrol and compounds S1, S2, S3, S4, and S5 at doses ranging from 0.0003 to 0.03 mg/kg had no effect on the locomotion of animals in comparison with the animals in the control group, suggesting that the compounds do not induce impairment of motor function in the animals. These results exclude the possibility that the antinociceptive action of carvacrol and its derivatives is nonspecifically associated with activity on the peripheral or central levels of locomotion control, such as sedation or motor dysfunction.

As previously shown (Fundytus, 2001), the administration of glutamate receptor (GluR) antagonists has an analgesic effect on peripheral pain. To assess whether the mechanism of action of S1–S5 is likely to involve these receptors, molecular docking simulations were performed over different GluR structures of Rattus norvegicus bound to antagonists (Supplementary Table S1).

For the predicted binding modes of S1–S5, the main interactions involving the common scaffold are hydrogen bonds with Gln405, Arg523, Thr518 and Ser572 and a displaced π-stacking interaction with Phe484. Of these, the interactions with Arg523, Thr518, and Phe484 are also observed for the crystallographic antagonist TK40 (Ravn et al., 2013). The main interactions observed for carvacrol are only hydrogen bonds with Pro516 and Thr518 and the displaced π-stacking interaction with Phe484 (Figure 7). For the different R groups, mainly van der Waals interactions were established. Only for the R groups of S3 and S5, -CH···π interactions with Leu538 and Ser572, respectively, were observed. Among all five molecules, S4 established the lowest number of contacts. The scores of each analyzed pose are presented in Supplementary Table S2.

The three levels of perception of pain—the cerebral (Dickenson, 1995), spinal and peripheral (Gordh et al., 1995)—appear to be affected by NO. This compound is an essential regulator of various immune and inflammatory functions (Moncada et al., 1991). In this work, we investigated, besides the NO scavenging activity, the possible intermolecular interactions between the sulfonamides and NO synthase. First, to validate the molecular docking protocol, redocking analysis (Figure 8) of 5,6,7,8-tetrahydrobiopterin (the crystallographic ligand, PDB ID 6NGJ) (Do et al., 2019) was carried out with GOLD. The ligand occupied the same interaction site in molecular docking when compared to the crystallographic structure, with emphasis on hydrogen bond interactions with Ser334, Val677, and Arg 596 and a π interaction with Trp678.

The molecular docking results agree with the results obtained in the in vitro (NO scavenging activity) and in vivo tests. All compounds showed an inhibitory profile against NO synthase, except S4, which was not effective in all performed assays. The two most active compounds, S1 and S5 presented lower ID₅₀ values and higher values for the scoring function, which demonstrate the high correlation between the in vivo and in silico results. The higher activity of these compounds was probably due to π stacking interactions and a hydrogen bond between compounds S1 and S5 and Trp 678 (Figure 9), which were also observed for the co-crystallized ligand, but was not found for the other sulfonamides. The scores of each analyzed pose are presented in Supplementary Table S3.

Physicochemical and topological parameters of compounds S1–S5 were estimated to evaluate their pharmacokinetics profile. The octanol–water partition coefficient (miLogP), topological polar surface area (TPSA), molecular weight (MW), number of atoms, hydrogen-bond acceptors (HBA) and hydrogen-bond donors (HBD), number of rotatable bonds (NRB), Lipinski RO5 violations, and molecular volume are presented in Table 2. The silico-derived descriptor values were compared with the solubility and permeability filters for drug candidates reported by Lipinski (Barret, 2018), Oprea and Veber (Veber et al., 2002).

The SwissADME web tool used to calculate the parameters is available at http://www.swissadme.ch and allows straightforward submission and analysis. It allows different input methods, multi-molecule computation, and offers the possibility to view and save results for each molecule, in addition to an interactive and intuitive visualization tool. To study the ADME parameters of the most active sulfonamide in the in vitro and phenotypic tests (S1), the Bioavailability Radar (Figure 10), lipophilicity, drug likeness (Figure 11), medicinal chemistry and pharmacokinetics (Figure 12) parameters were analyzed.

The Bioavailability Radar (Figure 10) provides a graphical output for the drug-likeness of a compound. The central shaded surface is the optimal domain for lipophilicity (XLOGP3 from −0.7 to +5.0), size (MW from 150 to 500 g/mol), polarity (TPSA from 20 to 130 Å²), aqueous solubility (logS ≤ 6), saturation (fraction of sp3 carbons ≥ 0.25), and flexibility (rotatable bonds ≤ 9). Compound S1 falls within the optimal range for all parameters.

In addition, S1 has a good medicinal chemistry and synthetic accessibility profile, which is very important in obtaining a drug that can be commercially distributed at a more affordable price. Moreover, S1 has high gastrointestinal absorption (GI) and blood-brain barrier permeability (Figure 12).

The evaluation of the antioxidant activity of a series of compounds should be performed in more than one experiment, allowing for the reliability of the results (Sens et al., 2018). Diverse in vitro antioxidant assays have been published. Herein, the antioxidant ability of derivatives S1–S5 was determined in two in vitro tests. Subsequently, the results of these tests were correlated with the findings from the HOMO and LUMO studies.

Compounds S1, S3, and S5 were more active than ascorbic acid (AA), which was used as the reference compound. Compound S5 showed the highest activity, and S4 demonstrated to be the least active. A linear correlation was found between both experimental results (EC₅₀ ^H2O2 = 1.085EC₅₀ ^NO + 0.2250; r ² = 0.99).

NO plays a critical part in the control of multiple physiological responses. Also, the NO cascade is associated with many conditions, including Alzheimer’s disease (Di Meo et al., 2016). H₂O₂ readily decomposes into water and oxygen, resulting in the production of hydroxyl radicals (OH•), lipid peroxidation and DNA injury, which makes it a target for research of new compounds with antioxidant properties (Phaniendra et al., 2015).

Extensive research has revealed that NO plays an essential role in several biological processes, such as neurotransmission, immune defense, and regulation of cell death (Snider and McMahon, 1998). The early 20th century witnessed the discovery of the role played by NO in nociception in both the central and peripheral levels (Zhuo and Gebhart, 1997). One of the physiological functions of NO was initially found in the vasculature; it was shown that the role of endothelium-derived relaxation factor (EDRF) could be quantitatively explained by the formation of NO by endothelial cells (Moncada and Higgs, 2006).

Treatment of pain with NO donors began with the use of nitroglycerin (NTG), which figures among the oldest treatments for ischemic heart disease (Boden et al., 2015). Discovered in 1847, NTG was used for the therapy of pain in angina pectoris for 100 years. However, its mechanism of action was not revealed until EDRF was identified as NO (Marsh and Marsh, 2000). Independently, NO was found to be an endogenous activator of soluble guanylate cyclase, resulting in the formation of cyclic GMP (cGMP), which acts as a second messenger in many cells, including the sensory neurons (Pereira et al., 2011).

NO is a highly reactive chemical messenger diffusible through the cytoplasmic membranes that is critical for the control of neuronal transmission, inflammation, cytotoxicity, and neural plasticity (Pacher et al., 2007). NO modulates the excitability of spinal sensory neurons and contributes to pain in different ways. The control of NO biosynthesis is regulated by NO synthase (NOS) enzymes. Three NO synthase isoforms (NOS; EC 1.14.13.39) catalyze the production of NO (Förstermann and Sessa, 2012). They use O₂ and L-arginine as substrates and flavin mononucleotide (FMN), flavin adenine dinucleotide (FAD), reduced nicotinamide-adenine-dinucleotide phosphate (NADPH), and tetrahydrobiopterin (BH4) as cofactors (Förstermann and Sessa, 2012). In this work, molecular docking was used to investigate NOS inhibition by the carvacrol derivatives.

The HOMO profile showed a variation of the charge density among the carvacrol derivatives. The HOMO and LUMO energies of compound S1 is −7.40 and −0.62 eV, respectively. The electronic density is concentrated in the phenol group for HOMO and LUMO. Compound S2, however, differs regarding the position of the charge density for these orbitals. In HOMO (−6.86 eV), the orbitals are concentrated on the fluoro-phenyl group. This is because fluorine tends to attract electron density (electronegative atom). In LUMO (−0.75 eV), the electronic density tends to be favorable in the phenol group. The band gap in this compound is −6.11 eV. Compound S3 has HOMO and LUMO energies of −7.35 and −0.86 eV, respectively. The electronic density of HOMO tends to be located at the phenol. In LUMO, however, the electronic density concentrates in the region of the pyridinic group. In compound S4, the HOMO charge density surrounds the phenolic substituent (−6.83 eV). In LUMO, however, the charge density concentrates in the carvacrol fragment (−0.59 eV). Differently from the other compounds, the charge distribution in S5 distributes throughout the structure in HOMO (−7.38 eV) and LUMO (−1.01 eV). In HOMO, the positive density concentrates on the sulfonamide group and p-fluorine atom. In LUMO, however, the same region is predominantly negative throughout the structure. In ascorbic acid, the dihydroxyfuran has the HOMO electron density (−6.91 eV) close to the hydroxyl groups in the resonant region. In LUMO (−0.49 eV), the signal of electronic density changes and concentrates close to the oxygen atom of the furan group.

Injection of glutamate (i.pl.) in the mouse paw causes significant paw edema and nociception (Beirith et al., 2002; Meotti et al., 2010). Figure 4 shows that systemic administration of carvacrol, S1, S2, S3, and S5 significantly inhibits nociception induced by injection of 20 µmol/paw glutamate, suggesting that these compounds have an important therapeutic effect for the treatment of acute pain of inflammatory origin, probably due to a decrease in peripheral glutamatergic signaling. Treatment with the compounds significantly reduced pain behavior induced by glutamate (i.pl.), characterized by spontaneous licking/biting of the injected hind paw. Carvacrol was able to reduce nociceptive behavior by 19 ± 6 and 44 ± 8% at 0.0003 and 0.003 mg/kg, respectively. Moreover, the sulfonamides derived from carvacrol showed the following percent inhibitions: 16 ± 5%, 62 ± 5%, and 50 ± 7% for S1 at 0.0003, 0.003, and 0.03 mg/kg, respectively; 26 ± 5% and 34 ± 6% for S2 at 0.003 and 0.03 mg/kg, respectively; 25 ± 9% and 22 ± 5% for S3 at 0.003 and 0.03 mg/kg, respectively; 39 ± 8% and 27 ± 13% for S5 at 0.003 and 0.03 mg/kg, respectively.

The calculated mean ID₅₀ value for sulfonamides derived from carvacrol was 0.002 (0.001–0.002) mg/kg for S1, 0.442 (0.063–0.387) mg/kg. Thus, the results of the present study demonstrate that carvacrol and S1, S2, S3, and S5 reduce nociception induced by glutamate (i.pl.), suggesting that inhibition of the stimulatory mechanism via peripheral glutamatergic neurotransmission may contribute, at least in part, to the antinociceptive effect of these compounds. In addition, we would like to highlight that carvacrol and compounds S1, S2, S3, and S5 may be interesting lead compounds for acute pain, especially S1 (0.003 mg/kg) since it presented the highest efficacy among the analyzed compounds.

Importantly, the compounds derived from carvacrol, selected to carry out the in vivo experiments, were chosen from the results presented in the molecular docking, quantum studies, and the in vitro antioxidant activity. Our results corroborate previous results (Arigesavan and Sudhandiran, 2015) which also found antioxidant and anti-inflammatory effects after treatment with carvacrol, using a carcinogenicity model in the colon of rats. Moreover, previous studies demonstrated that carvacrol attenuates mechanical hypernociception induced by carrageenan (Guimarães et al., 2012) and the acute pain acetic acid-induced abdominal constriction and formalin (Cavalcante Melo et al., 2012). Also, it was shown (Barnwal et al., 2018) that carvacrol increased the activities of antioxidant enzymes and downregulated expression by reducing the inflammation marker in positively dyed cells (iNOS, NF-κB, and COX-2) in a pulmonary toxicity model. These data from the literature reinforce the antinociceptive, anti-inflammatory, and antioxidant potential of carvacrol observed in our study.

Findings from the literature (Pacher et al., 2007; Förstermann and Sessa, 2012) indicate that superoxide (SO, O^(2)•(−)) and peroxynitrite (PN, ONOO⁽⁻⁾, the product of its reaction) are essential for the emergence of pain caused by different etiologies. These findings reinforce the concept that ROS play an essential part in NMDA activation, which is a critical ionotropic glutamatergic receptor, which contributes to central and peripheral pain. Therefore, this study supports previous results (Wang et al., 2004) that stated that superoxide mediates hyperalgesia (increased sensitivity to painful stimulation) through M40403, a manganese(II) complex with a bis(cyclo-hexylpyridine-substituted) macrocyclic ligand, which is a superoxide dismutase mimetic. These findings disclosed the central role played by superoxide in the peripheral signaling of nociception. In addition, it was shown that the M40403 antihyperalgesic activity could not be reverted by naloxone, which excludes the participation of opioid signaling cascades. Moreover, so far, few studies have investigated the effect of carvacrol on neurotransmitter modulation. The studies by Zotti et al. (2013) demonstrated that carvacrol, when ingested regularly in low concentrations, influences brain activity by increasing the levels of neurotransmitters such as serotonin and dopamine, which can determine feelings of well-being and reinforcing positive effects. Thus, our interest in investigating the glutamatergic system has arisen, considering that glutamate is a major mediator in the CNS, mediating excitatory neurotransmission in mammals, including in sensory neurons that convey pain, being strongly involved in the stimulation of peripheral and central pain. Therefore, our findings are unprecedented and relevant as they demonstrate the inhibitory capacity of carvacrol on the peripheral glutamatergic pathway.

It was shown (Kuo et al., 2017) that carvacrol mitigated injury in tissues and inflammation derived from periodontitis induced by ligation. Besides that, carvacrol proved to attenuate inflammatory response induced by carrageenan, decreasing mouse paw edema (Guimarães et al., 2012). These data from the literature support the anti-inflammatory, antinociceptive and antiedematogenic effects of carvacrol observed in our study. Importantly, paw edema and pain induced by glutamate are essentially associated with non-NMDA ionotropic glutamate receptors and NO production, a vasodilator, and an important neurotransmitter (Beirith et al., 2002). When in excess, it may be involved in the production of oxidative lesions in proteins. These findings reinforce the importance of studying glutamate-induced paw edema and nociception, as well as the beneficial effects of carvacrol and its derivatives found in this study.

Our results agree with literature data which demonstrated that carvacrol had no effect on the spontaneous locomotion in mice (Cavalcante Melo et al., 2012; Guimarães et al., 2012). However, these studies used a curve of carvacrol doses ranging from 25 to 100 mg/kg in the open-field test and we are the first group to test a much lower dose curve for carvacrol (0.0003, 0.003, and 0.03 mg/kg) in pain, edema, and spontaneous locomotion. In addition, Guimarães et al. (2012) demonstrated that carvacrol at a dose of 100 mg/kg reduced the animals’ ambulation in the open-field test, 30 min after intraperitoneal administration, showing that this dose is not safe as it causes nonspecific effects on locomotor activity and should be excluded in future pain studies. It is already well described that some drugs can cause motor slowness (bradykinesia) or even act as a muscle relaxant, causing non-specific changes in the locomotor activity of animals (Cartmell et al., 1991). In addition, drugs like benzodiazepines and other anxiolytics decrease the exploratory behavior of animals (Hazim et al., 2014). In this regard, it was demonstrated (Coderre and van Empel, 1994) that many glutamate antagonists, primarily via ionotropic NMDA receptor, such as the receptor channel block MK-801, produce significant antinociceptive effects, but decrease exploratory behavior of animals. In contrast, our results demonstrate that the intragastric treatment with the tested compounds can induce a significant antinociceptive effect via inhibition of peripheral glutamate, without causing any detectable motor dysfunction. Thus, carvacrol and its derivatives S1, S2, S3, and S4 at doses up to 0.03 mg/kg have an attractive analgesic potential to treat acute pain without causing CNS sedation.

In general, no significant binding modes were obtained concerning poses matching the available structural criteria of known antagonists (Ramírez and Caballero, 2018). Only the docking simulations in the NMDA-GluN₁ glycine binding site (LBD-GluN₁) excelled, which agrees with previous observations for selective ligands of this site, such as HA-966, “which barely interacts with other ionotropic glutamate receptors” (Planells-Cases et al., 2005).

Considering the docking results and the non-ataxic effects of the compounds at the administered doses, the compounds are likely to be partial agonists, instead of agonists of the NMDA-GluN1 glycine binding site, such as rapastinel (Wood et al., 2008) (GLYX-13 or BV-102), (+)-HA-966 (Millan and Seguin, 1993) and the recently reported 1-amino-1-cyclobutanecarboxylic acid (Fung et al., 2019).

The Lipinski RO5 applies to compounds that are active after oral administration. The RO5 includes four physicochemical property ranges (logP ≤ 5, MW ≤ 500, HBD ≤ 5 and HBA ≤ 10) that are present in 90% of the drugs that are active after oral administration and have reached phase II clinical development (Barret, 2018). The sulfonamides investigated in this work are within the RO5 desirable range, except for the miLogP of sulfonamide S5 (miLogP = 5.32), which is slightly higher than expected.

TPSA correlates with a compound’s ability to permeate biological membranes through passive transport. Medicinal chemists use TPSA as an important parameter to optimize drug permeation through membranes. Molecules having TPSA values higher than 140 Å² are likely to permeate poorly into cell membranes (Pajouhesh and Lenz, 2005). For molecules that are required to act in the CNS, penetration into the blood-brain barrier is needed, which requires a TPSA lower than 90Ų (Hitchcock and Pennington, 2006). All investigated sulfonamides are in accordance with these parameters. A molecule that has a higher number of rotatable bonds becomes more flexible and have a good binding affinity with the binding pocket. For a potential drug candidate, Veber proposed that NRB should be ≤10. All investigated sulfonamides are following this parameter.

The molecular volume assesses the transport properties of molecules such as blood-brain barrier penetration. The calculated values for this property are in line with the values expected for drug candidates.

During the discovery of novel drugs, molecules with useful therapeutic properties and low levels of toxicity are highly desirable. In this process, knowledge of the absorption, distribution, metabolism, and excretion profiles (ADME) is essential. It is well-known that the early evaluation of ADME during the drug discovery process reduces the attrition rates during clinical development.