Aminopyridines in the development of drug candidates against protozoan neglected tropical diseases

Abstract

Neglected tropical diseases (NTDs) pose a major threat in tropical zones for impoverished populations. Difficulty of access, adverse effects or low efficacy limit the use of current therapeutic options. Therefore, development of new drugs against NTDs is a necessity. Compounds containing an aminopyridine (AP) moiety are of great interest for the design of new anti-NTD drugs due to their intrinsic properties compared with their closest chemical structures. Currently, over 40 compounds with an AP moiety are on the market, but none is used against NTDs despite active research on APs. The aim of this review is to present the medicinal chemistry work carried out with these scaffolds, against protozoan NTDs: Trypanosoma cruzi, Trypanosoma brucei or Leishmania spp.

GRAPHICAL ABSTRACT

Plain language summary

Article highlights.

Introduction

• Neglected Tropical Diseases (NTDs) are major threat to impoverished populations in tropical zones due to the lack of adverse-free efficient oral treatments.

Aminopyridines compounds active against Trypanosoma cruzi, Trypanosoma brucei & Leishmania spp.

• Because of their intrinsic properties and ease of introduction into molecules, aminopyridines (APs) are interesting scaffolds for the development of new anti-NTD drugs.

• The work presented focuses on the development of AP drugs targeting Trypanosoma cruzi, Trypanosoma brucei and Leishmania spp.

• APs got three main contributions to the compounds in which they are included: lipophilicity reduction, additional(s) target interaction(s) and intermediate structure allowing access to more complex aza–heterocycles (such as pyrimidines or imidazoles).

Conclusion

• APs should be further included in drug discovery work focused on anti-NTD drugs because of their possible positive outcome on aqueous solubility and metabolic stability.

1. Background

Neglected tropical diseases (NTDs) are a group of infectious diseases and conditions mainly prevalent in tropical zones, mostly affecting impoverished populations [1], while being ignored by private research companies. This definition was introduced, in 2003 [2], by the World Health Organization (WHO) which estimated, in 2021, that 1.65 billion people affected by these pathologies required treatment and care [3]. NTDs involves 20 pathologies, 17 of which are due to infectious diseases caused by four types of pathogens: protozoa (Chagas disease, human African trypanosomiasis, leishmaniasis), bacteria (Buruli ulcer, leprosy, trachoma, yaws), viruses (dengue, chikungunya, rabies) and helminths (cysticercosis/taeniasis, dracunculiasis, echinococcosis, foodborne trematodiases, lymphatic filariasis, onchocerciasis, schistosomiasis, soil-transmitted helminthiasis). Each NTD has defined health targets to be reached by 2030 with important measures to be implemented such as chemoprevention, improved case management, vector control measures and improved hygiene and safe water supply [4]. The development of new drugs is therefore a key element in the fight against NTDs that is supervised, in part, by Drug for Neglected Disease Initiative (DNDi), a non-profit organization founded in 2003. As protozoan parasites are responsible for most of the morbidity and mortality worldwide NTDs, specific treatments are particularly important [5].

Aminopyridine (AP) is a compound consisting of a pyridine ring substituted by one (in position 2 [2-AP], 3 [3-AP] or 4 [4-AP]) or multiple amine groups. In this review, we have considered AP for which the nitrogen atom(s) of the amine function(s) can contain multiple substitutions. It (they) can be included in a saturated ring (such as piperazine) but not in an aromatic cycle (Supplementary Figure S1). Moreover, no other cycle attached to the pyridine was accepted. Among monosubstituted APs (2, 3 or 4-AP), only 4-AP has shown pharmacological properties. As a potassium channel blocker, it is used to improve walking speed in multiple sclerosis patients [6]. It is also a widely used compound in neurological studies due to its convulsive properties.

When designing new compounds of interest in medicinal chemistry, different molecular descriptors are used to guide the drug design process. There are various rules indicating the desired ranges for these descriptors such as Lipinski's Rule of 5 (or Ro5, established in 1997) [7] and its variants (Veber's rule, Ghose filter) for design of orally active drugs and the Rule of 3, defined in 2003 by Congreve et al. [8], to identify molecular fragments of interest (see descriptors and their rules values in Supplementary Table S1). The lipophilicity of a molecule, measured by its logP, has a direct influence on its metabolic stability and permeability [9]. The polarity of a molecule is reflected by its total polar surface area (TPSA) which is considered detrimental for oral bioavaibility if it exceeds 140 Ų [10,11], and limits blood–brain barrier (BBB) permeability if it exceeds 90 Ų [12]. In addition, the number of H-bond donors (HBD) and acceptors (HBA) is important for compound-target interactions and aqueous solubility but too many of them would result in a compound with too much polarity. Lipinski's rule predicts poor absorption or permeation when there are more than 5 H-bond donors, 10 H-bond acceptors, the molecular weight (MW) is greater than 500 and the calculated Log P (cLogP) is greater than 5. As mentioned beforehand with TPSA, the 140 Ų limit is one of the additional parameters included in a Ro5 variant, the Veber Ro5 [10]. Supplementary Table S1 summarizes the calculated values of the main descriptors (MW, HBA, HBD, TPSA, clogP) influencing the pharmacokinetics and pharmacodynamics of APs. To study the influence of the nitrogen atom, calculations were carried out for 2-AP, 3-AP, 4-AP and their closest chemical structures: toluene, aniline and methylpyridines. Compared with these structures, APs, with their two nitrogen atoms, have two advantages: a low clogP value (from 2.51 for toluene to around 0.5 for APs) and one HBA and one HBD whereas the other molecules have one or the other (or none for toluene). However, the two nitrogen atoms come increase the TPSA value compared with the structures with only one (from 12.89 Ų and 26.02 Ų for methylpyridines and aniline respectively to 38.91 Ų for APs). Thus, by replacing a tolyl moiety with an AP in a compound, we will observe a reduction of the clogP, with a potential improvement in metabolic stability and permeability. Additional HBA and HBD could also improve potency against the target by adding new interactions and improving the aqueous solubility. Moreover, from a synthesis point of view, compounds with an AP group are easier to obtain than those with a tolyl or methylpyridinyl group, since the amine function allows a number of simple chemical reactions such as nucleophilic substitution.

There are around 40 compounds on the market with an AP moiety, none of which have activity against NTDs, and only two with anti-infective activity such as the antifungal isavuconazonium and the antibiotic ozenoxacin (Supplementary Figure S2A). Moreover, as of June 2023, DNDi still had no drug candidate in its portfolio with an AP moiety [13]. However, for non-NTD protozoan diseases, two compounds with an AP moiety are currently in clinical trials for the treatment of malaria: MMV688533 and ZY19489 (Supplementary Figure S2A & B). Apart from protozoan diseases, AP compounds can also be found in antitubercular molecules, including the 2-aminothiazole lead developed by Azzali et al. (Supplementary Figure S2B), which has recently led to another potential antitubercular 2-AP-oxazole scaffold. [14–16] The 2-AP moiety on MMV688533 replaced a guanidine function on former lead MMV669851 but, in the absence of data available to compare the two, no conclusion can be given as to the impact of the 2-AP moiety on MMV688533 [17]. For ZY19489 [18], its development started with the HTS hit I (Supplementary Figure S2B) which displayed a good potency on Plasmodium falciparum (150 nM) but a low solubility value (10 μM) and potential cardiac toxicity due to hERG activity (3.7 μM). Both these problems were resolved by replacing the p-ethylaniline moiety with a 2-methyl-4-aminopyridine (see compound II in Supplementary Figure S2B) which significantly improved solubility (830 μM) and reduced hERG potency by more than 20-fold (85 μM). Other pharmacomodulations modified the side chain from a 4-AP to a 2-AP ring.

APs are of particular interest in the development of NTD drug candidates due to their intrinsic physiochemical (PC) properties and consequently their potential positive impact on pharmacokinetic (PK) and pharmacodynamic properties. Thus, this review aims to present medicinal chemistry works carried out with these scaffolds of interest against protozoan NTDs including early study to identify an active molecule up to the preclinical stage. Trypanosoma cruzi, Trypanosoma brucei and Leishmania spp are responsible for Chagas disease, human African trypanosomiasis and leishmaniasis respectively. These three NTDs, known as the tritryps, are the cause of the highest rates of NTDs mortality [19]. We will focus our review on APs that have been developed against these three protozoan parasites.

2. Aminopyridines compounds active against Trypanosoma cruzi

Trypanosoma cruzi is responsible of the Chagas disease, which affects an estimated number of 6 to 7 million people, mostly in Latin America [20]. It is mainly transmitted to humans through contact with the urine or faeces of triatomine bugs. Congenital transmission is also an important method of infection. The disease has two phases: an initial acute phase during the first two months, mostly asymptomatic or with unspecific symptoms, with blood-circulating parasites (extracellular trypomastigote form); a chronic phase that can last decades where the parasite is hidden in the heart or digestives muscles (intracellular amastigote form) leading to cardiac, digestive or neurologic disorders [21]. Two drugs, available for treatment during the initial phase: benznidazole and nifurtimox, do not have an aminopyridine core (Supplementary Figure S3) [22]. Both drugs are effective if given quickly. However, they have side effects including major ones such as dermatitis and neuropathies. In addition, treatment duration is long (up to 90 days) thus, resulting in a high drop-out rate. They are also contraindicated for pregnant women and nursing mothers, making them unusable to prevent congenital transmission. Furthermore, these two molecules have a limited effect on the chronic stage. There are currently no drugs on the market to treat the chronic stage. Therefore, efforts are made to develop drugs effective on both stages (thus on both promastigotes and amastigotes) for a sterile cure (elimination of all parasites in host's cells) with limited side effects. Among the drugs that have been developed, several AP have been studied. The AP parts of the anti-Chagas compounds described below are either required for activity, based on an interaction between the pyridine's nitrogen and heme iron of TcCYP51; or regarded as chemical tool for improving PK parameters.

In 2013, Choi et al. [23] introduced a 4-AP based series targeting the sterol 14-demethylase enzyme (CYP51) of T. cruzi. CYP51, a member of the CYP450 superfamily, is a key enzyme in the synthesis of ergosterol, an important membrane building block for fungi and protozoa. A preliminary library screening on Mycobacterium tuberculosis CYP51, identified the N-(4-pyridyl)-formamide as an interesting moiety to inhibit the active site of MtCYP51 [24]. Later, Chen et al. co-crystallized compounds bearing a N-(4-pyridyl)-formamide moiety to study their interactions with the active site of MtCYP51 and tested their binding affinities with different CYP51s [25]. The hit compound LP10 has been identified. It displayed a K_D TcCYP51 of 42 nM, a Tc amastigotes EC₅₀ of 0.65 μM and moderate in vivo activity on mouse model (Figure 1). For all co-crystallized APs, pyridine's nitrogen coordinates with heme iron in MtCYP51 active site. However, LP10 shown an additional interaction thanks to its tryptophan-like scaffold where the indole filled up a hydrophobic pocket, enhancing its affinity for the active site. Thus, LP10 was engaged in the structure-based drug design work carried out by Choi et al. Minor modulations on the 4-AP and the indole moieties were detrimental to the K_D and EC₅₀ values, confirming the in silico information of the importance of the 4-AP and indole moieties. Therefore, modulations focused on the amidocyclohexyl moiety. 40 compounds were synthesized from benzamide 1 which displayed improved in vitro activity compared with LP10. Among them, four compounds exhibited a Tc amastigotes EC₅₀ below 60 nM, 3 being the most potent (Figure 1). The introduction of a fluorine atom at position 2 and a substituted phenyl ring at position 4 of the benzamide improved both in vitro and metabolic stability (Figure 1). Authors also improved their in silico model by successfully co-crystallizing the hit 2 with Trypanosoma brucei CYP51 (85% homology with TcCYP51). This model highlighted the importance of the fluorine atom on the benzamide, allowing new hydrophobic interactions.

Figure 1.

Evolution of the structure and in vitro/in vivo properties of the 4-AP-indole series targeting TcCYP51.

Observing differences between (R) and (S) enantiomers binding poses in silico, Choi et al. [26] reevaluated the best enantiomer for this series. Contrary to their initial expectations, the (R)-enantiomers are the most active: (R)-4 is 1000-fold more potent than its enantiomer (S)-2. Among the synthesized compounds, (R)-5 is the more active with an in vitro activity close to picomolar (Figure 1). Calvet et al. [27]. synthesized and evaluated in a mouse model a series of compounds using bioluminescent T. cruzi strain (expressing a luciferase). In vitro and in vivo potency were not correlated since 5, the most potent in vitro, was not the most effective in vivo (Figure 1). The introduction of a piperazine moiety between the two phenyl rings proved essential. A co-crystal structure of TcCYP51 with 6 gave additional structure–activity relationship: the piperazine moiety elongates and gives more flexibility to the molecule, allowing additional hydrophobic interactions between the difluorophenyl ring and surrounding aminoacids (not observed with TbCYP51 crystal). In vivo potencies were greatly improved, from 40 to 70% of inhibition at 25 mg/kg per os during 4 days for the biphenyl series (e. g. compound 5) to 90% inhibition with the introduction of a piperazine moiety (e.g. compound 6). Moreover, compared with 5, 6 displayed an improved in vivo half-time (from 3.29 to 4.39 h) and a reduced observed clearance (from 53.78 to 38.84 ml/min/kg) (Figure 1). Vieira et al. [28] pursued the structure-based drug design work, starting from 6, by making modifications on the terminal N-phenyl ring. In silico studies indicated that substitution of the phenyl ring with hydrophobic substituents (Cl or CH₃) at position 2, 3 or 5 should promote interactions with the enzyme because they allowed the tail of the compounds to fit better into hydrophobic cavity of CYP51. Four compounds out of 10 showed an in vivo percentage of inhibition above 98%, in an infected mouse model, with 25 mg/kg/day per os during 4 days. Compound 7 was the best with improved in vitro metabolic stability and observed clearance compared with 6 although its survival rate is lower than 7 (Figure 1).

6 and 7 were selected for further in vivo studies [29]. Both compounds reduced the parasitemia by more than 99% in a mouse model of a severe acute infection with 25 mg/kg/day per os during 28 days. Although the survival rate at day 30 post-infection was better for 7, no sterile cure was achieved (Figure 1). 6 was selected for a test on a chronic infection mouse model and proved effective up to 27 days after treatment (25 mg/kg/day per os during 28 days) with bioluminescence levels close to uninfected mouse levels. However, following immunosuppression, parasitemia relapse was observed indicating that no sterile cure was achieved. Along all these works, interactions with humans CYP was closely monitored to limit possible interactions with these off-targets [23,25–28].

Another AP compound potentially targeting TcCYP51 was designed by Keenan et al. [30]. in 2013. They based their work on fenarimol, a plant fungicide targeting fungi CYP51 enzyme, which had an in vitro potency of 350 nM against T. cruzi amastigotes and cytotoxicity of 77 μM on L6 cell-line [31]. A first round of optimization led to analog 8 with a Tc amastigotes EC₅₀ of 12 nM and a promising in vivo potency reducing parasitemia levels by over 99% in a mouse model at 20 mg/kg/day for 20 days but failing to achieve a sterile cure (Supplementary Figure S4). From this work, authors designed an achiral series changing the chiral carbon by a nitrogen leading to a 3-AP series of fenarimol analogs. 20 3-AP were synthesized, all (except one) possessing a Tc amastigotes EC₅₀ below 40 nM without in vitro cytotoxicity on L6 cell-line. 9 was identified as the best compound of the 3-AP series due to its good in vitro potency and metabolic stability with good PK parameters such as an in vivo t_1/2 per os of 12.5 h and an oral bioavailability of 95%. In an in vivo mouse model, 9 displayed major improvement over 8 with a sterile cure rate of 60% (other mouse did not survive) at the dose of 20 mg/kg/day during 20 days (Supplementary Figure S4).

From their compound library, Friggeri et al. [32] selected molecules for their structural similarity with antifungal azoles targeting CYP51. They identified the 4-AP compound 10 with a Tc amastigotes EC₅₀ of 36 nM and a moderate cytotoxicity at 27.2 μM (Supplementary Figure S5) on L6 cell-line. Unlike previously described compounds, the 4-AP moiety of 10 is located in the ‘tail’ of the molecule and the heme iron coordination is provided by the imidazole nitrogen (like antifungal azole compounds). 10 was also tested on other parasites: it is potent on the K1 strain of P. falciparum with an EC₅₀ of 1.5 μM.

Clinical studies carried out with posaconazole [33] and fosravuconazole [34], two known CYP51 inhibitors, showed a high relapse rate against T. cruzi. For both drugs, the chelating moiety is a triazole ring and not an AP like previously described compounds. However, following this discovery, CYP51 is now considered an undesirable target for the development of new anti-Chagas drugs and all compounds targeting this protein have had their development stopped. Finally, the activity of new anti T. cruzi compounds against this protein should be evaluated early, in order to eliminate those that inhibit it [35].

In 2017, Brand et al. [36] discussed the optimization of a 5-amino-1,2,3-triazole-4-carboxamide series active against T. cruzi. Starting from hit compound 11, the authors synthesized compounds 12–15 (Supplementary Figure S6). Compared with the phenyl analog 12, 13 and 14 showed a minor loss of activity on T. cruzi but, for 13, the 2-AP moiety decreased intrinsic clearance and hERG channel potency compared with 14. Observations on these AP-compounds combined with the importance of meta-substitutions led to 15 as the best compound of this series. This work highlights the possibility offered by AP-moiety to pursue modulations into others aza-heterocycles such as pyrimidine.

The latest AP-compound targeting T. cruzi was published by McGonagle et al. [37]. The authors identified hit compound 16 by high-throughput screening (HTS) from GSK library. Although, it has an interesting activity (EC₅₀=0.16 μM), it showed metabolic instability, reflected by its high intrinsic clearance. In order to reduce this instability while maintaining potency against T. cruzi, modulations at different parts of 16 were performed leading in the synthesis of 45 compounds. Replacing the central phenyl ring of 17 with pyridine 18 improved the metabolic stability and gained solubility but unfortunately resulted in a 16-fold loss in potency (Supplementary Figure S7). None of the 45 analogs reached the levels of metabolic stability and potency desired by the authors.

3. Aminopyridines compounds active against Trypanosoma Brucei

Trypanosoma brucei is the parasite responsible for sleeping sickness, mainly transmitted to humans by the infected bite of the tsetse fly. Two subspecies are able to infect humans, T.b. gambiense (western and central Africa, 97% of the reported cases) and T.b. rhodesiense (eastern and southern Africa) [38]. The number of cases is continuously decreasing since 1998, dropping from 40,000 to 663 in 2020. Sleeping sickness is divided in two phases: a hemolymphatic phase with non-specific symptoms where the parasite slowly develops in body fluids (blood, lymph, etc.) followed by a phase with neurologic symptoms (including sleeping disorders) due to the fact that the parasite has crossed the BBB. In both stages, parasites are found in the trypomastigote form. If left untreated, sleeping sickness is lethal. Mainline treatment of T.b. gambiense is fexinidazole (introduced in 2019, Supplementary Figure S8) for adult cases and, pentamidine or the combination nifurtimox-eflornithine for children under 6 [39]. For T.b. rhodesiense, which causes a more acute form of the disease, first-line treatment is suramin [40]. A major advance was the introduction of fexinidazole, the first active oral treatment for sleeping sickness with fewer side effects than other drugs [41]. Fexinidazole is currently being studied against T.b. rhodesiense and the European Medicine Agency recently gave a positive opinion for its use [42]. Another oral treatment, acoziborole, displayed a 98% success rate in a phase II/III clinical trial against T.b. gambiense [43]. Efforts must be maintained to develop other oral treatments against T. brucei. AP-containing compounds targeting T. brucei described in the literature primarily targeted parasite kinases or were derived from known human kinases inhibitors.

In 2013, Patel et al. [44] initiated pharmacomodulations on lapatinib, a marketed human tyrosine kinase inhibitor (Figure 2), which proved potent (EC₅₀ = 1.54 μM) against T. b. brucei (another subspecies only found in animals and used as a laboratory model). This work focused on modulation on position 6 of quinazoline core of lapatinib. Compound 19, harboring a morpholino–phenyl substituent, was more active against lapatinib with a T.b. brucei EC₅₀ of 42 nM. Starting with 19, Devine et al. [45]. continued the pharmacodulations of the quinazoline moiety, in position 6 or 7, while modifiying this heterocycle (to quinoline, phtalazine or cinnoline). The resulting compounds were cross-test on different parasites (T. cruzi, T.b. brucei, L. major and P. falciparum) but, for T.b. brucei no potency gain was observed, 20 being the best compound with a T.b. brucei EC₅₀ of 79 nM.

Figure 2.

Chemical structures and in vitro properties of lapatinib analogs [44–48].

From these two initials works, three additional pharmacomodulations studies were published using 19 or 20 as a starting point (Figure 2). However, compounds like 20 displayed an interesting potency but had poor calculated drug-like properties: high molecular weight and cLogP combined with low predicted solubility and metabolic stability. The authors previously observed that deletion of the 3-fluorobenzyl group on compound 19 can have a beneficial effect on potency. For instance, compound 21 had a moderate amelioration of activity against T.b. brucei (EC₅₀ = 0.82 μM, Figure 2) compared with 22 (EC₅₀ = 1.3 μM, Figure 2). Devine et al. [46]. decided to work on a series of quinoline derivatives because this scaffold is widely represented in medicinal chemistry. 26 compounds with size-limited substituent at position 4 were synthesized, four of them bearing an AP moiety: 4-AP, 3-AP, 2-AP and 6-nitro-3-AP. The compound 23 with a 4-AP moiety was one of the best compounds of this series with a T.b. brucei EC₅₀ of 0.11 μM and an improved ligand–lipophilicity efficiency (LLE) at 4.11 (Figure 2). But, the metabolic instability of 23 was worsened compared with 19. Introduction of the nitrogen atom on position 4 improved activity value and LLE score compared with aniline 24 but, PC and PK properties were better for 24.

In the meantime, Woodring et al. [47] applied the same strategy on the quinazoline series. 26 compounds modified at position 4 were synthesized but none was as active as 19. Of the six compounds with an AP-moiety, quinazoline 25 displayed the best potency with an EC₅₀ against T.b. brucei of 0.29 μM (Figure 2). Compounds with 3-AP or 2-AP moiety had lower activities (around 3 and tenfold respectively), which were moderately improved for 2-AP (2 or threefold) by introducing a substituent on the pyridine ring (amine or acetamide). PC and in vitro PK parameters of 25 were measured. Aqueous solubility was improved compared with 19 (2.9 μM vs <1 μM), while the plasma protein binding was maintained around 99%. However, the introduction of the 4-AP moiety increased metabolic liability for 25 compared with 19, with the Cl_int on human microsomes rising from 63.0 to 80.2 μl/min/mg.

In 2018, Ferrins et al. [48]. described the screening of 20 quinolines, based on structure of 20 by modulations at position 4 and/or 7. Three compounds in the ‘position 4 modified’ set included a 2-AP moiety, substituted at position 4. Of these three, compound 26 was the best one with a T. b. brucei EC₅₀ of 0.27 μM associated with a HepG2 CC₅₀ of 10 μM (Figure 2). The best compound of the series, 27, had a highly potent against T.b. brucei (EC₅₀ = 6 nM). However, 27 was inactive on an in vivo mouse model infected by T.b. brucei, even at dose up to 70 mg/kg per os.

While screening a set of compounds against T.b. brucei in 2018, Veale and Hoppe identified 2-AP 28 [49], from which the antimalarial 2-AP MMV390048 (Supplementary Figure S9) was derived. MMV390048 targets plasmodial phosphatidylinositol-4-kinase (PfPI4K) and has reached stage II of clinical trials. Compound 28 displayed an EC₅₀ of 0.33 μM against T.b. brucei with a CC₅₀ on the HeLa cell-line above 25 μM and was used as starting point for pharmacomodulations by Veale et al. in 2019 as potential TbPI4K inhibitor [50]. The authors synthesized over 90 compounds, belonging to the 2-AP and 2-aminopyrazine series, and evaluated their antiparasitic activity on T.b. brucei. The majority of compounds had an EC₅₀ below 10 μM and only 2-AP 29 and 30 showed activities below 1 μM (Supplementary Figure S9). Among the few compounds evaluated for their pharmacochemical properties, the 2-aminopyrazine 31, bearing a 2-(4-methylpiperazine)-pyridine moiety, displayed the best solubility (182 μM) and microsomal stability (<7 μl/min/mg).

Also in the field of kinase inhibitor design, Tear et al. developed a series of pyrazolo[1,5-b]pyridazine based on compound 32 (Figure 3) [51]. Pyrazolo[1,5-b]pyridazines have been described as being able to target S. aureus serine/threonine kinase, human glycogen synthase kinase 3β and cyclin dependent kinase 2 and 4. The 2-AP 32 displayed a EC₅₀ against T.b. brucei of 6 nM and a cytotoxicity above 50 μM on MRC5 cell-line, but poor PC and PK parameters, notably a clearance of 210 μl/min/mg. Around 70 compounds were synthesized with modulations focusing on: introduction of substituents on the pyrazolopyridazine core, change of the pyrimidine core to other aromatic cycles and modification of the cyanoaniline moiety to other amine substituents including AP moieties. For the AP compounds 33–36, EC₅₀ against T. b. brucei varied from 5 to 0.02 μM. Compared with 32 or the aniline 37, their PK parameters were improved, with both better aqueous solubility and metabolic stability with reduced plasma protein binding. Based on the interesting results obtained with 3-AP 34, the authors worked on positions 2 and 6 of the pyrazolopyridazine core, leading to the identification of 3-AP 38. It displayed an EC₅₀ at 12 nM, good metabolic stability but a poor aqueous solubility of 0.8 μM and was not selective at 1μM against selected human kinases (inhibition >80%). In order to gather in vivo PK data, 34 was studied and displayed good BBB penetration but, as anticipated by the authors, its poor selectivity was reflected by an observed toxicity.

Figure 3.

Chemical structures and in vitro properties of compounds 32–38.

Another HTS of kinase inhibitors led to the identification of 39, a diaminopurine, as a potent compound against T. b. brucei [52]. Its good antiparasitic activity (EC₅₀ = 32 nM) was unfortunately counterbalanced by a poor metabolic stability with a clearance of 64 μl/min/mg (Figure 4). Moreover, the presence of the thiophene ring can potentially give highly reactive metabolite. Thus, 39 was investigated in a medicinal chemistry work by Singh et al. in 2020 [53]. Out of 60 synthesized compounds, the two compounds with an AP moiety (40 and 41) did not show a better activity compared with 39 or the aniline 42. However, 4-AP 40 displayed very good PC and PK parameters with an aqueous solubility close to 1mM and a very good metabolic stability with a clearance inferior of 3 μl/min/mg. Among the 60 obtained compounds, only pyrazole 43 was the most interesting both in terms of activity and studied PC/PK parameters (see target values in Figure 4). An in vivo study on mouse model showed that 43 improved the mean survival time of infected animals by 18 days when administered at 30 mg/kg for 5 days.

Figure 4.

Chemical structures and in vitro properties of compounds 39–43.

N-myristoyltransferase (NMT) is an essential protein catalysing the transfer of myristate from myristoyl–CoA to other proteins. NMT was found to be an essential protein for many organisms including T. brucei [54]. From the hit 44, Brand et al. [55]. developed the 4-AP 45 which displayed very good activity (EC₅₀ = 2 nM) on T. b. brucei and good metabolic stability. However, a high cytoxicity (CC₅₀ = 0.4 μM) against the MRC5 cell-line as well as lack of selectivity (TbNMT IC₅₀ = 2 nM vs HsNMT IC₅₀ = 4 nM) were observed (Figure 5). In an in vivo mouse model of the hemolymphatic stage, 45 displayed a fully curative effect at a dose of 12.5 mg/kg/day for 4 days. Unfortunately, in an in vivo mouse model mimicking the central nervous system stage, it was inactive due to its inability to cross the BBB (B:B ratio <0.1, Figure 5). The co-crystal of 45-L. major NMT (used as surrogate model) identified multiple interactions between 45 and the essentials aminoacids of the active site: hydrogen bond formed by basic nitrogen atom from the pyrazole ring and Ser330, hydrogen bonds formed between the sulfone/piperazine ring and water and the dichlorophenyl ring filling up a hydrophobic pocket.

Figure 5.

Chemical structures and in vitro properties of compounds 44–49.

From compound 45, two different lead optimization approaches were carried out. Brand et al. [56]. focused on substitution of the sulfonamide nitrogen atom and replacement of the piperazinopyridine moiety. This work led to the synthesis of 46 with a difluoromethyl substitution on the sulfonamide and a N-methylpiperidinepropyl instead of the piperazinopyridine in 45. The most notable improvement of 46 over 45 is its ability to cross the BBB barrier (B:B ratio = 1.6) due to its increased logP (4.7 vs 3.1) and decreased PSA (58 vs 101), key parameters for BBB penetration, while maintaining the same activity against T.b. brucei. In 2017, a structure-based drug design work was carried out by Bayliss et al. [57]. using former crystallographic data (LmNMT) and a new Aspergillus fulminus NMT model (with up to 92% homology on the binding site compared with TbNMT). Thus, the scaffold of 45 was simplified to facilitate modulation of structural elements involved in interaction with the active site and keep PSA and logP in acceptable range. Compound 47 was the best of this simplified series with an IC₅₀ value of 0.5 μM on TbNMT, an EC₅₀ value of 2 μM against T.b. brucei and good metabolic stability. Structural elements from 45 (dichlorophenyl and pyridine ring) were reintroduced in 47 structure, resulting in 48 with an improved T.b. brucei potency (EC₅₀ = 0.7 μM) and maintained metabolic stability compared with 47. Additional modulations confirmed that the dichlorophenyl ring was a crucial structural moiety while extensive modulation of the pyridine ring led to 49 as the lead compound of the series. Compared with 45, the most notable improvement for 49 is the B:B ratio measured at 1.9 instead of less than 0.1 for 45.

In 2014, Ferrins et al. [58] identified, through an HTS against T. brucei, the pyridyl benzamide 50 (Supplementary Figure S10). Its moderate activity (EC₅₀ = 3 μM) was counter-balanced by its good PC properties (low molecular weight, PSA and logP), making it a good starting point for a pharmacomodulation work. Around 80 2-AP analogs were synthesized with a methodical substituent exploration (small size substituents on the different positions on both cycles). Compound 51 and three compounds (52–54) exhibited an EC₅₀ close to or below 0.1 μM against T. b. brucei and T. b. rhodesiense respectively. Regarding PC parameters, this series displayed a good aqueous solubility on some compounds and a medium to high plasma protein binding. However, some results revealed that this series could be sensible to metabolism.

4. Aminopyridines compounds active against Leishmania spp.

Leishmania spp. are parasites responsible for leishmaniasis, transmitted to humans by the infected bites of female phlebotomine sandflies. Leishmaniasis can take three forms: visceral, cutaneous or mucosal leishmaniasis [59]. Visceral leishmaniasis (VL) is a slowly progressing disease that affects several organs (mostly spleen and liver) and is fatal without treatment. Cutaneous leishmaniasis (CL) is a disfiguring disease caused by skin infection beginning with papules or nodules that can lead to ulcers. Mucosal leishmaniasis (ML) is also a disfiguring, but less frequent form, in which the mucous membranes of the nose, mouth or throat are destroyed. CL and ML are not fatal but cause important social stigma to affected populations. The WHO estimates the number of VL new cases between 50,000 and 90,000 annually (5710 deaths in 2019) [60] and the number of CL new cases between 600,000 and 1 million [61]. The most important parasite stage of leishmaniasis is the intracellular amastigote stage, which develops inside macrophages, and is therefore the main target for drugs. For VL, liposomal amphotericin B (Supplementary Figure S11) is highly efficient but its price and storage conditions (refrigerator) make it difficult to use in poor countries [62]. Amphotericin B is also used but causes severe and potentially lethal side effects, mostly renal. Pentavalent antimonials, miltefosine and paromomycin are other available treatments that have good efficacy but are long-term treatment with possible serious adverse effects. Efforts are being made to obtain new oral [63] and topical treatments for CL and ML which are affordable, safer and more efficient than the current ones.

Starting with an HTS campaign, Bhuniya et al. identified a series of aminothiazoles as compounds of interest against L. donovani, the most common species in VL [64]. Hit compound 55 had a medium activity (IC₅₀ = 3.6 μM) on intramacrophagic amastigotes, better than miltefosine (IC₅₀ = 8.2 μM), a weak cytotoxicity value (CC₅₀ = 54.4 μM) on KB cell line and a good aqueous solubility (Figure 6). However, 55 is quickly metabolized by hamster and mouse liver microsomes. The authors started with eighteen N-(phenylsubstituted)-4-(pyridin-2-yl)thiazol-2-amines with different substitution patterns on the phenyl ring. Most compounds were less active than 55 with IC₅₀ above 40 μM except with 56 and its 2,4,6-trimethylphenyl N-substituent (IC₅₀ = 0.37 μM). Modulations of the N-substituent was shifted from aniline to 2-AP and proved to be of major interest. Compound 57 was much more potent than its phenyl analog 58 (IC₅₀ = 2.1 μM vs >100 μM), associated with a low toxicity value (CC₅₀ = 75.6 μM) on KB cell-line. 57 also displayed an improved aqueous solubility, compared with 55, at 60 μM. Among seven 2-AP compounds, 59 displayed the best selectivity index with a good activity (IC₅₀ = 90 nM) and a CC₅₀ value close to 100 μM. However, the introduction of the chlorine atom in 59 compared with 57 had a major impact on the aqueous solubility (S <10 μM vs 60 μM). Moreover, like 55, all of these 2-AP were metabolically unstable. To solve this problem, the 2-pyridine moiety was substituted, resulting in five new compounds including 60 and 61. While 61 was the most potent with IC₅₀ on the nanomolar range (IC₅₀ = 3 nM), it suffered from metabolic instability despite its substitution. 61 was selected for in vivo PK study because of its greater in vitro metabolic stability.

Figure 6.

Chemical structures and in vitro properties of compounds 55–61.

In 2018, Coimbra et al. designed two series of 2-pyrazyl and 2-pyridylhydrazone compounds for their potential against L. braziliensis promastigotes and L. amanozensis promastigotes and amastigotes (both species found in CL and ML) [65]. Among the various 2-pyridylhydrazones (thus harboring a 2-AP moiety), hydrazones 62 and 63 (Supplementary Figure S12) displayed the best activity on intracellular L. amazonensis amastigotes with IC₅₀ of 15.10 and 11.98 μM respectively, close to the miltefosine value. Among the nine 2-pyrazylhydrazones synthesized, 64 was the best compound with IC₅₀ values close to those of miltefosine. However, its 2-pyridyl analog, compound 65, showed a loss of potency on L. braziliensis promastigotes and L. amanozensis amastigotes (IC₅₀ values >100 μM) compared with 64. Interestingly, most compounds exhibited no significant cytotoxicity with CC₅₀ above 150 μM on macrophages compared with 138.62 μM for miltefosine.

Clofazimine, an antimycobacterial used against M. tuberculosis and M. leprosis, showed in vitro and in vivo antileishmanial effects [66]. However, clinical trials have shown it to be ineffective in the treatment of VL [67]. Thus, Barteselli envisaged modulation on the phenazine core by changing the phenyl rings in position 2 and 4 and the isopropyl grafted on the imine (Figure 7) [68]. Among the 21 synthesized analogs (including four compounds with 3-APs moieties), 66 and 67 displayed the greatest potencies on the promastigote form of L. infantum (IC₅₀ = 0.23 and 0.37 μM) and L. tropica (IC₅₀ = 0.12 and 0.22 μM). Replacing the aniline on position 2 of 67 with a 3-AP (68) resulted in lower potency. However, replacement of the bulky quinolizine moiety by a dimethylamine (69) restored the antileishmanial activity. Unfortunately, all clofazimine analogs studied in this work displayed a significant increase in cytotoxicity, with often a CC₅₀ of less 4 μM compared with the references clofazimine (CC₅₀ = 18.6 μM) or amphotericin B (CC₅₀ = 25.7 μM) on the Human Mammary Epithelial Cells 1 (HMEC-1) cell line. Compound 69 was identified as a new hit and used as a starting point for further modifications by Bassanini et al. [69]. 14 riminophenazines were synthesized, including 3-APs compounds, and tested on the promastigote form of L. infantum and L. tropica. Best compounds were then evaluated on the amastigote form of L. infantum. Removal of the AP moiety like in compound 70, the phenyl analog of 69, led to an eightfold decrease in antileishmanial potencies and increased cytotoxicity value. For the 12 3-AP compounds, the dimethylamine side chain was modified into other amine moieties. Two compounds, 71 and 72 with a piperidine and a N-methylpiperazine respectively, displayed the best results on the amastigote form of L. infantum with a percentage inhibition of 44% at 1 μM, which is far from the value of amphotericin B value, which is 99.8% (Figure 7). 71 and 72 also had an improved cytotoxicity value on the Bone Marrow-Derived Macrophage (BMDM) cell line compared with 71. However, they are still more cytotoxic than amphotericin B on this model. The clofazimine analogs derived from these two studies [68,69] were also tested on P. falciparum and a majority displayed an antiplasmodial activity below 0.5 μM on the W2 strain.

Figure 7.

Chemical structures and in vitro properties of clofazidine analogs 66–72.

After developing a new assay to screen compounds against Leischmania amastigotes, Paape et al. [70]. tested seven compounds previously described as L. major [56,57] or L. donovani [71] NMT inhibitors. Thus, 2-APs 45, 73 and 74 were tested on LdNMT and on L. donovani amastigotes. All three compounds showed nanomolar values for inhibition of LdNMT but only 73 could give an IC₅₀ value, at 190 nM, against L. donovani amastigotes (Supplementary Figure S13). The absence of activity for 45 can be explained by a lack of permeability attributable to the non-substituted nitrogen atom in the sulfonamide function. For 74, the low activity could be due to the loss of a key H-bond between the piperazine nitrogen and LdNMT, the piperazine ring nitrogen being substituted by methyl for 73 unlike 74.

5. Conclusion

In this review, we presented APs developed as drug candidates against Trypanosoma cruzi, Trypanosoma brucei or Leishmania spp, parasites responsible of the three protozoan NTDs: Chagas disease, human African trypanosiomasis and leishmaniasis. APs showed interesting PK and pharmacodynamics properties due to their intrinsic PC properties. Indeed, the medicinal chemistry works presented here, allowed the identification of three different uses for AP moieties to obtain new drug candidates against NTDs.

The three main interests of AP due to the nitrogen atom are to allow a reduction in lipophilicity, new interactions with the target such as H-bonds and to demonstrate the interest of possibly introducing other heterocycles. Firstly, the majority of studies in this review used APs as a tool to reduce the lipophilicity of compounds and consequently, potentially improve the biological and/or the PC and PK profile of the AP-containing compounds. The work of Tear clearly illustrates the effects of replacing an aniline with an AP on aqueous solubility, intrinsic clearance and plasma protein binding (Figure 3) [51]. It also shows the importance of the position of the nitrogen atom in the ring for these properties. Secondly, APs are of interest when a series of compounds has a defined molecular target with a known 3D structure. They can provide a useful element for structure-based drug design thanks to their possible additional interactions: H-bond or acceptor and π-stacking. Because of the difficulty of identifying molecular targets in parasites, only the extensive works on CYP51 inhibitors give us an example of AP-containing compounds where the AP moieties is crucial for the activity (Figure 1 & Supplementary Figure S4). Thirdly, an AP can be an intermediate moiety leading to its subsequent replacement by more complex or more difficult to introduce heterocycles (pyrimidines, triazines, imidazoles, etc.). Indeed, an AP group can be easily introduced thanks to its amine function and these are inexpensive commercially available reagents. The work by Brand displays the progressive stages of modification, starting with a methoxyaniline, then different APs, before finally deciding to introduce a dimethoxyaminopyrimidine, based on data obtained from AP compounds (Supplementary Figure S6). [36]

Based on the data collected, Table 1 summarizes the impact of 2-AP, 3-AP and 4-AP on antiparasitic activity or pharmacokinetics parameters. Although these data remain limited due to the lack of available points of comparison (anilines, methylpyridines and methylphenyls versus APs) we can highlight a few important points. The first observation is that none of 2-AP, 3-AP and 4-AP appears to have a greater impact than the others on antiparasitic properties or pharmacokinetic parameters. Secondly, with regard to biological properties, APs are considered as a key structural element for antiparasitic activity. Biological targets are rarely identified for APs, with the exception of 1–7 and 9 that are CYP51 inhibitors [35]. The final point concerns the impact on APs on pharmacokinetic parameters, which very often appears favorable. In the four examples (13, 18, 24 and 40) in Table 1, the replacement of an aniline by an AP was followed by an improvement in metabolic stability (reduced intrinsic clearance) and, for three of them, an improvement in solubility.

Table 1.

Positive effects of representative aminopyridines compared with their anilines analogs.

Effects on antiparasitic activity	Aminopyridines	Effects on pharmacokinetic parameters
Key structural element: chelation of the heme in TcCYP51 (1 – 7 and 9) Increased potency on T. b. brucei: 0.11 μM for 23 vs 0.98 μM for aniline analog 24 – worse potency for 2-AP and 3-AP analogs		Improvement of 40 vs aniline analog 42	Solubility: 955 vs 15 μM. Intrinsic clearance: <3 vs 26 μl/min/mg. Plasma protein binding: 27 vs 96%
Increased potency on Leishmania spp.: 0.34 vs 3.20 μM on L. infantum promastigotes and 0.34 vs 2.51 μM on L. tropica promastigotes for 69 and aniline analog 70 respectively		Improvement of 18 vs aniline analog 17	Solubility: ≥373 vs 43 μM. Intrinsic clearance: 39 vs >50 ml/min/g.
		Improvement of 34 vs aniline analog 37	Solubility: 6 vs 0.8 μM. Intrinsic clearance: 69 vs >300 μl/min/mg. Plasma protein binding: 89 vs 96%.
Important structural element: from structure–activity relationship data (28 – 30) Increased potency on L. donovani: 2.1 vs >100 μM for 57 and aniline analog 58 respectively		Improvement of 13 vs aniline analog 12	Intrinsic clearance: 2.1 vs 33 ml/min/g.

6. Future perspective

APs are an interesting and versatile option for the pharmacomodulation of compounds targeting protozoan NTDs. Thanks to their intrinsic properties, they often improve the aqueous solubility and metabolic stability of the compounds into which they have been introduced. However, the lack of comparison between the AP moiety and other aromatic core in a number of the studied works does not always make it possible to highlight the benefit of its introduction on pharmacodynamic and PK parameters. Given the importance of the improvements observed in the published work, we hope that in the next decade APs will be used as much as possible during hit-to-lead chemistry and lead optimization steps, mainly to improve aqueous solubility and metabolic stability, key parameters in the development of oral anti-NTD drugs.

Supplementary material

Supplementary data for this article can be accessed at https://doi.org/10.1080/17568919.2024.2359361

Author contributions

Writing – review: R Mustière, A Dassonville-Klimpt, P Sonnet. Project administration & editing: P Sonnet.

Financial disclosure

The authors have no financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Competing interests disclosure

The authors have no competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.