Nitric oxide (NO) is an endogenic product from plants, bacteria, and animal cells that has many important effects in those organisms. It is produced by nitric oxide synthase (NOS), which is found in main three isoforms, namely endothelial NOS (eNOS), inducible NOS (iNOS), and neuronal NOS (nNOS). It has an important role in homeostasis in different physiological systems, such as micro- and macro-vascularization, inhibition of platelet aggregation, and neurotransmission regulation in the central nervous, gastrointestinal, respiratory, and genitourinary systems. However, its overproduction has been associated with diseases such as arthritis, asthma, cerebral ischemia, Parkinson’s disease, neurodegeneration, and seizures. For this reason, and due to better understanding of the molecular mechanisms by which NO provokes those diseases, the interest on the design of NOS inhibitors with therapeutic purposes has highly increased. Based on the foregoing considerations, the proposal of this chapter is to show an overview about the design strategies, mechanism of action at the molecular level, and the main advances toward the search for selective NOS inhibitors available in the literature.
- nitric oxide synthase isoforms
- structure-based drug design
- enzymatic inhibition
- heterocyclic compounds
Nitric oxide (NO) is a diatomic neutral molecule, produced by bacteria, plants, and animals. Having one unpaired electron, its effect in biological system is related to the stabilization of this electron. It acts as dissolved nonelectrolyte in the organisms, except for the lungs, where it is found in gaseous state [1–3].
NO has gain importance mainly in the 1990s, and from then on, it has been studied to obtain interesting pharmacological effects. A review by Serafim and collaborators describes the state of the art of this compound use in drug design .
The basal NO production has an important contribution to homeostasis in different physiological systems, such as micro- and macro-vascularization, inhibition of platelet aggregation, and neurotransmission regulation in central nervous, gastrointestinal, respiratory, and genitourinary systems. However, NO overproduction has been strongly associated with some diseases such as arthritis, asthma, cerebral ischemia, Parkinson’s disease, neurodegeneration, and seizures [5–9].
Nitric oxide synthase (NOS) is the enzyme responsible for NO biosynthesis, and there are three main kinds of NOS isoforms [endothelial NOS (eNOS), inducible NOS (iNOS), and neuronal NOS (nNOS)]. They are tetramers, constituted of two NOS monomers associated with two calmodulin monomers (CaM), and contain relatively tightly bound cofactors, BH4, FAD, FMN, and iron protoporphyrin IX (heme group). Their chemical function is to catalyze the reaction of
Exacerbated induction of iNOS is associated with septic shock, inflammatory, and noninflammatory impairment processes in different tissues/organs, and, likewise, the nNOS is triggered in neurotoxicity, neurodegeneration process, and proliferation increase of some neoplastic cell lines. Depending on the clinical condition, decreasing NO levels is necessary, and excellent benefits might be achieved using NOS inhibitors. However, it is much important not to inhibit eNOS, because of its central role in smooth muscle relax, controlling vascular tone and blood pressure [12–14].
The first inhibitors designed (during the1980s and early 1990s) were based on
NOS isoforms were validated as target for new drugs soon after their X-ray crystallography was available. From then on the design of effective and selective inhibitors has been an important approach in modern drug discovery involving NO biochemical pathways related to many dysfunctions of the human organism [12, 17–19].
2. Experimental studies
2.1. Inducible nitric oxide synthase (iNOS) inhibitors
Garvey and collaborators (1994) were the first to report highly selective iNOS inhibitors. The compounds were isothiourea derivatives (Figure 3—1) designed as
The selective iNOS inhibition by aminoguanidine (Figure 3—AG) showed that NO can mediate the disruption of hematopoiesis during acute graft-versus-host disease (GVHD), also decreasing the endogenous bacterial infections in the spleen and liver in mice receiving the inhibitor . The deleterious neuro-inflammation effects of iNOS/NO system stimulated by lipopolysaccharide (LPS) on learning and memory were evaluated in rats. Aminoguanidine decreases TNFα levels, oxidative stress indicators, and NO metabolites . In addition, this compound seems to significantly relieve periapical inflammation in the canine teeth of cats and to reduce histological multiple organ damage in rats [24, 25].
In 2000, Hagmann and collaborators explored the structure-activity relationships of a series of substituted 2-aminopyridines. Compounds 4,6-dialkyl substituted (Figure 3—2 and 3) were found to be the most potent inhibitors of iNOS, presenting a significant degree of selectivity for this isoform . Exploring the same scaffold, the synthesis of the derivatives N-4-piperidinyl-2-aminopyridine led to compound 4 (Figure 3), a 4-methoxy substituted derivative, 4-fold more potent in iNOS inhibition compared to the 4-methyl compound. Moreover, 4-cyanobenzamide derivative (Figure 3—AR-C133057XX) presented IC50 = 0.071 μM, being 1400-fold and around 100-fold selective for eNOS and nNOS, respectively. X-ray crystallography of AR-C133057XX showed that pyridine moiety binds deeply to the heme pocket, while the exocyclic ring interacts with another binding pocket. This difference in interaction could explain the good selectivity of this molecule .
On the other hand, the 1,2-dihydro-4-quinazolinamine compound AR-C102222 (Figure 3) showed a dose-dependent inhibition on NO production induced by lipopolysaccharide (LPS). At the highest dose tested (100 μmol/kg) in rats, this compound completely abolished the chronic inflammatory arthritis development all over the 20-day experiment, thus confirming the in vivo efficacy of the class .
Structural-based approach using crystal structure and mutagenesis have identified specific induced-fit binding mode, which can generate some conformational changes toward a new specific cavity. Garcin and coworkers showed the cis-amidine moiety of quinazoline and aminopyridine chemotypes in the compounds AR-C133057XX and AR-C102222, respectively, promoted interactions of hydrogen with Glu (Glu371 and Glu377, mouse and human, respectively) at the binding site and with the heme group (mimicking the
Other compounds, as acetamidine derivatives (Figure 5—5 and 6), designed to inhibit iNOS, showed submicromolar activities (IC50 = 0.428 and 0.165 μM, respectively) and excellent selectivity over eNOS (>2300 and 550-fold more selective, respectively). In silico findings revealed that the activity drastically changes when ending amino groups are located instead of carboxylic function in the acceptor H-bond region, which is adjacent to the lipophilic region. This occurs because the charged amino group and its alkyl chain are not able to be stabilized inside the pocket interaction region of the enzyme, decreasing the binding efficiency .
A protein called SPSB2 plays an important role in modulating the activity of iNOS through its proteasomal degradation in defense cells. Since this complex is blocked, the NO production from iNOS is prolonged, increasing the killing activity against pathogen microorganisms, making it an interesting anti-infective target . Some cyclic peptidomimetic compounds were designed using this strategy, and the most potent compound 7 (Figure 5) showed strong inhibition of SPSB2-iNOS complex in macrophage cell lysates and potent affinity value (KD = 29 nM) .
High-throughput screening (HTS) strategy has been used to identify new iNOS inhibitors hits such as the compound 8 (Figure 5). After structural optimization process, the quinolone amide derivative 9 (Figure 5) was generated. This derivative has 2000-fold selectivity over human eNOS, besides being very potent (iNOS IC50 = 11 nM) and presenting oral bioavailability in mouse although its clearance (Clp > 100 mL/min/kg) has shown to be uninteresting . The efforts to overcome this effect was to improve the pharmacokinetic properties, leading to the compound 10 (Figure 5), a 4,7-imidazopyrazine derivative. This is a dual iNOS/nNOS inhibitor, showing high potency in human iNOS (IC50 = 0.091 μM) and activity over nNOS (0.30 μM) while maintaining the desired selectivity over eNOS (180-fold). Its mechanism of action at molecular level is based on the inhibition of the iNOS dimerization process. Furthermore, compound 10 was effective using in vivo models of neuropathic pain, presenting desired clearance value (4–9 mL/min/kg), good oral bioavailability, and no tolerance after repeated doses .
Phenylpyrroles, pyrazoles, urea kynurenamines, ethynylcyanodienones, and amidine derivatives (Figure 6—11, 12, 13, 14, and 15) have also been interesting scaffolds to generate iNOS inhibitors [36–40]. The first derivative reduced significantly the iNOS activity to control values in MPTP-Parkinson`s disease model, showing a potential to act in central nervous system (CNS) disorders .
Natural products have been a rich source of new bioactive molecules. Some examples in the NOS inhibition are a sesquiterpenoid, isolated from Curcuma wenyujin (Figure 6—16) and its isomer. They were strong inhibitors of NO production by LPS in iNOS (IC50 = 7.6 and 8.5 μM, respectively). Anmindenols A and B (Figure 6—17 and 18), from marine-derived bacterium Streptomyces sp., also demonstrated a relevant inhibitory activity in macrophage cells NO production (IC50 = 23 and 19 μM, respectively) [41, 42].
2.2. Neuronal nitric oxide synthase (nNOS) inhibitors
In the beginning of the 1990s, efforts to design selective nNOS inhibitor compounds were addressed, using the substrate
The non-arginine-based compound 7-nitroindazole (Figure 7—NI) showed little nNOS in vitro selectivity but high in vivo selectivity. Its mechanism of action involves competitive inhibition of H4B cofactor, and series of related structures have been designed . This compound has suppressed open-field behavior expressed as distance moved, exploratory rearing and grooming, suggesting that this compound can increase cortical excitability and interfere with some physiological and behavioral parameters . Nevertheless, its anticonvulsant activity should be better understood, since studies in rodents reveal a beneficial activity although proconvulsant effect can be found in kainite-, nicotine- and soman-induced convulsions .
Entrena and collaborators, by using kynurenamine scaffold (Figure 7—19) as a template, carried out the synthesis of a series of new candidates to neuroprotective compounds, showing a pharmacophore model to interact with nNOS catalytic site .
N-Substituted acetamidines (Figure 7—20 and 21) showed nNOS inhibition activity (IC50 = 0.2 and 0.3 μM) with good selectivity index (500 and 1166-fold selectivity over eNOS, respectively, and 50 and 100-fold, over iNOS, respectively). In silico studies were useful to understand the fit of these scaffolds inside the catalytic site. nNOS contains a bigger heme group cavity compared with other isoforms, which could explain why bulky groups better accommodate in the neuronal isoform .
Aminopyridine is an attractive pharmacophoric group to bind in different regions of nNOS through H-bond. Using this moiety, compound 22 (Figure 7) and its optical isomer showed to be selective against nNOS, although unable to penetrate the blood-brain-barrier (BBB) in in vivo studies [15, 47, 48]. Trying to increase the CNS permeability, prodrug design approach was used in primary and secondary amines. However, this strategy did not increase the BBB penetration, even masking the charge by carbamate and azide functions . On the other hand, using these compounds containing basic nitrogen, Xue and coworkers attached electron-withdrawing groups (Figure 7—23) close to these amine functions, decreasing their pKa values and improving the membrane permeability in cell-based assays .
Concerning 4,5-dihydro-1-H-pyrazole derivatives, they were confirmed as selective nNOS inhibitors. Compound 24 (Figure 7), the most active of the derivatives (82% of inhibition), showed that its methoxy electron-donating group is important to improve potency and selectivity. By molecular modeling, it was possible to identify that the phenyl moiety can fit below the heme group, establishing π-π interaction. The methoxy groups adopt a conformation that allows them to interact with Arg481 by H-bonds. Moreover, the amine group interacts by H-bond with one of the carboxylate moieties of the heme group. On the contrary, electron-withdrawing groups are better to generate inhibitors for iNOS .
Other interesting structures such as 2-aminoquinolines are effective scaffold to be included in the structure of nNOS inhibitors. Crystallography studies showed that those compounds act as competitive arginine mimics. This scaffold makes important H-bonds with the active-site Glu residue, and the non-coordinating aryl rings are stabilized in a hydrophobic pocket in the extremity of the substrate access channel. Moreover, this structural class showed good pharmacokinetic properties (Figure 8—25), such as brain penetration and oral bioavailability according to the permeability results in Caco-2 cell assay . Trying to optimize this class of compounds, chlorine was added on the phenyl ether central aryl ring (Figure 8—26). This substitution was found to be selective and highly potent in the design of nNOS inhibitors while retaining CNS penetration and showing a diminished off-target interaction. In complex with human nNOS, this compound showed a phenyl ring orientation where the alkyl amine makes an H-bond with the H4B site .
Exploring the heme-coordinating potential of imidazole group, a series of 2,4-disubtituted pyrimidine compounds (Figure 8—27) was designed. They presented a nanomolar affinity to both rat and human nNOS (>200-fold and >100-fold selectivity over eNOS and iNOS, respectively), exhibiting a minimal off-target binding to 50 CNS receptors. Crystal structures of the complex (nNOS-27) indicate that heme Fe coordinates by the 2-imidazolyl group, and the non-coordinating aryl rings are stabilized in a hydrophobic pocket at the far end of the substrate access channel. The fluorine atom in compound 27 also interacts into the hydrophobic pocket, and the secondary amine of this derivative establishes dual ionic interaction with both heme propionates (Figure 9). This molecular interaction profile is important to obtain potency and selectivity. In addition, nitrogen from pyrimidine ring performs an H-bond with the His342 side chain. The imidazole ring of the most active compound acts as a weak CYP3A4 inhibitor, suggesting that modulating hydrophobicity and bulkiness can be useful to attenuate the effects in CYP isoforms .
Studies using aminopyridine-based scaffold with pyridine linker (Figure 8—28) showed that difference in the position of an amino acid, Asp597 of nNOS versus Asn368 of eNOS, controls the affinity and binding mode of this class of nNOS inhibitors. While the central pyridine is at least partially protonated to points up toward Tyr562 for optimal electrostatic interactions, the Asp597 provides additional and important electrostatic stabilization to the other part of the inhibitor [55–57].
Rational strategy for identifying new nNOS inhibitors using a combination of virtual screening approach based on 3D pharmacophore model and molecular docking was able to identify a hit compound structurally different from the available inhibitors (Figure 8—29). This strategy can be useful to design novel optimized analogs .
2.2.1. Double-headed nNOS inhibitors
Double-headed compounds have been explored by researches with the aim of obtaining high affinity binding in nNOS. Attaching a double-headed aminopyridine moiety in a compound led to a very potent (Ki = 25 nM) and selective (107-fold selective over nNOS and eNOS) compound (Figure 10—30). Both aminopyridine moieties interact in different positions with the enzyme—Glu592 in the active site and the heme group. According to the X-ray crystals, there is a second 30 molecule binding also in the H4B site, specifically, and the pyridine moiety coordinates with the Zn atom. This interaction does not take place in eNOS. Cellular permeability studies confirmed compound 30 as an interesting lead [59, 60].
Other symmetric double-headed aminopyridine series without charge groups were designed to contain a tail on the central aromatic ring. The objective is to achieve an interaction in the electronegative region in the catalytic site, since only the neuronal isoform has Asp597 in this region. Derivative 31 (Figure 10) was very potent (Ki = 56 nM) and highly selective over other isoforms (472-fold selective for eNOS and 239-fold for iNOS). This occurs because an electropositive functional group (Ciano group) is preferred near Asp597 in nNOS, explaining the selectivity of this compound .
Furthermore, double-headed inhibitors containing chiral linkers derived from natural amino acids were designed and synthesized. The best compound (Figure 10—32) showed high potency (Ki = 32 nM) against nNOS and a good selectivity profile (475-fold selective and 244-fold over eNOS and iNOS, respectively). The aminomethyl moiety was crucial in this compound, allowing it to bind to the heme propionates in nNOS and leading to a high selectivity level .
2-Amino-4-methylpyridine groups with a chiral linker derived from proline were designed as selective nNOS inhibitors. They showed to be interesting as they can interact in a unique orientation, what led to selectivity toward neuronal isoform. The aminopyridine groups interact with a Glu592 residue and the heme propionate in nNOS active site. In addition, the nitrogen from pyrrolidine linker is important to contribute to additional hydrogen bonds to the heme propionate, resulting in the most potent compound (Ki = 9.7 nM) (Figure 10—33). The finding that the isomer activities are different also reinforces the importance of chirality control of this kind of inhibitors and shows the dynamism of the target .
In addition, using chiral double-headed inhibitors, the α-amino-functionalized aminopyridine derivative 34 (Figure 10) was more potent than other chiral compounds (Ki value of 24 nM for nNOS, with 273-fold and 2822-fold selectivity against iNOS and eNOS, respectively). Structure-activity relationship studies reveal that the α-amino group close to the center phenyl ring is crucial to stabilize the double-headed binding. Those studies also showed that by changing to aminomethyl group the potency is improved. The inhibitor is able to make H-bonds with both the H4B binding site and the propionate of the heme A-ring, which is essential to obtain selectivity over other isoforms. It is also important to note that the distinct electrostatic environments in different isoforms resulted in lower binding free energy in nNOS, which also can explain the potency difference .
The non-chiral double-headed thiophene-2-carboximidamide compound (Figure 10—35) exhibited an excellent inhibitory potency and selectivity (Ki = 5 nM; 540-fold and 340-fold selective over eNOS and iNOS, respectively). This compound also showed to be active in metastatic melanoma A375 cells, exhibiting EC50 values of 1.3 μM, better that of the drug cisplatin (EC50 = 4.2 μM) .
2.3. Bacterial nitric oxide synthase (bNOS) inhibitors
Bacterial nitric oxide synthase (bNOS) is present in many Gram-positive microorganisms and has been described as part of their defense system against other species and the oxidative stress provoked by antibiotics through NO releasing. Therefore, bNOS inhibition can increase the antibiotic potential and be harmful to bacterial cell .
A screening showed that some known nNOS inhibitors can decrease significantly the percent survival of Bacillus subtilis WT treated with the antimicrobial acriflavine. This potential is consistent with NO production inhibition, as it decreases the bacterial resistance against that compound .
Exploring the potential of bNOS as a drug target, high selectivity levels are necessary to its inhibitors. In this context, the design of compounds that target the active and pterin-binding site has been considered an important strategy (Figure 11—36). These compounds are able to carry out an unexpected rotameric position of residue Arg247 in the active site, besides interacting with the important residue of Glu243 from the same site .
In addition, with the goal to identify the differences among bNOS and other isoforms, crystallography studies were performed using different inhibitor chemotypes. Researchers observed that Tyr706 from nNOS is conserved in bNOS (Tyr 357) and both have the same rotameric behavior, which is very different, compared with eNOS. This molecular feature can be useful to design new selective bNOS over eNOS inhibitor. Since the pharmacokinetic properties are very different between bNOS and nNOS, selectivity over the latter is not a trouble. However, due to steric hindrance in the tail end of thiophenecarboximidamide analogs, this scaffold can bind differently to bNOS comparatively to nNOS .
3. Clinical studies
A nonselective compound
Evaluating another inhibitor, N(G)-nitro-
Recent phase I study in advanced solid tumors with the iNOS inhibitor ASP9853 (Figure 12) showed that the efficacy dose predicted in preclinical studies was not achieved due to overall toxicity limitations. In summary all these clinical information showed that the manipulation of the NOS pathway, with or without chemotherapy, appears to be more challenging than expected . While designing new selective NOS inhibitors which should be highlighted, deeply studies to evaluate clinical benefits are also required.
Many scaffolds have been found to inhibit nitric oxide synthases. Some of them were presented in this chapter as promising for important therapeutic activity. It must be emphasized that the research about nitric oxide synthase inhibitors has expressively advanced thanks to the X-ray crystallographic studies of this enzyme. This helps the structure-based design approach toward the search for selective inhibitors of this enzyme and the comprehension of their mechanism of action. Notwithstanding, efforts have been made for imparting them with a drug-like profile.
The authors thank CAPES, for RAM Serafim scholarship, and CNPq, for EI Ferreira fellowship.