Antimalarial Activity of Highly Coordinative Fused Heterocycles Targeting β-Hematin Crystallization

Abstract

The β-hematin formation is a unique process adopted by Plasmodium sp. to detoxify free heme and represents a validated target to design new effective antimalarials. Most of the β-hematin inhibitors are mainly based on 4-aminoquinolines, but the parasite has developed diverse defense mechanisms against this type of chemical system. Thus, the identification of other molecular chemical entities targeting the β-hematin formation pathway is highly needed to evade resistance mechanisms associated with 4-aminoquinolines. Herein, we showed that the highly coordinative character can be a useful tool for the rational design of antimalarial agents targeting β-hematin crystallization. From a small library consisting of five compound families with recognized antitrypanosomatid activity and coordinative abilities, a group of tetradentate 1,4-disubstituted phthalazin-aryl/heteroarylhydrazinyl derivatives were identified as potential antimalarials. They showed a remarkable curative response against Plasmodium berghei-infected mice with a significant reduction of the parasitemia, which was well correlated with their good inhibitory activities on β-hematin crystallization (IC₅₀ = 5–7 μM). Their in vitro inhibitory and in vivo responses were comparable to those found for a chloroquine reference. The active compounds showed moderate in vitro toxicity against peritoneal macrophages, a low hemolysis response, and a good in silico ADME profile, identifying compound 2f as a promising antimalarial agent for further experiments. Other less coordinative fused heterocycles exhibited moderate inhibitory responses toward β-hematin crystallization and modest efficacy against the in vivo model. The complexation ability of the ligands with iron(III) was experimentally and theoretically determined, finding, in general, a good correlation between the complexation ability of the ligand and the inhibitory activity toward β-hematin crystallization. These findings open new perspectives toward the rational design of antimalarial β-hematin inhibitors based on the coordinative character as an alternative to the conventional β-hematin inhibitors.

Introduction

Malaria is a life-threatening disease caused by parasites of the genus Plasmodium spp., which is transmitted to humans through the bites of infected female Anopheles mosquitoes. It is one of the largest tropical diseases worldwide, registering an estimate of 241 million cases of malaria and 627,000 deaths in 2020 according to World Health Organization (WHO). From that statistics, it is important to mention that most of malaria deaths are associated with children aged less than 5 years with 67% of all malaria deaths worldwide.¹

Talking about the Plasmodium parasite, it infects the red blood cells (RBCs) in mammals and it is able to consume more than 60–80% of hemoglobin during its erythrocytic stage.^2,3 From the hemoglobin digestion, the parasite releases a globin protein and heme group [ferriprotoporphyrin IX (FPIX)]. The first is useful as a nutrient source for the parasite, whereas the second species is highly toxic for the parasite with the capacity to induce membrane rupture. To overcome it, the parasite has developed a detoxification mechanism within its digestive vacuole where the heme group is converted into a nontoxic crystalline solid called hemozoin (β-hematin), which is insoluble and can be eliminated via excretion.⁴⁻⁶ The β-hematin solid is formed by noncovalent interaction between several molecules of the heme group, which are connected through hydrogen bonding interaction between the vicinal carboxylic moiety of two heme groups and the dimer unit (Figure 1A). The latter makes the inhibition of the β-hematin crystallization an important chemical target to design new antimalarial agents.⁶⁻⁸ Diverse investigations have put in evidence that 4-aminoquinoline is a convenient scaffold to design new antimalarial agents. Specifically, chloroquine (CQ) has been identified as one of the most important antimalarials targeting β-hematin crystallization. Other clinically 4-aminoquinolines including amodiaquine, quinine, and mefloquine have been recognized to inhibit the formation of β-hematin.⁶⁻⁸ Also, an extended group of 4-aminoquinolines have been developed as an inhibitor of the formation of β-hematin.⁷⁻¹² Although the mechanism of inhibition of the formation of β-hematin is not well understood, diverse experimental and theoretical evidence has suggested that 4-aminoquinolines are able to block the formation of a β-hematin dimer through a π–π stacking interaction between the porphyrin ring of β-hematin and the quinoline ring. It facilitates an effective accumulation of the heme/drug complex within the erythrocyte, causing parasite death by oxidative stress.^7,8

Figure 1.

Scheme for the dissociation of β-hematin by ligand interaction (A), complexation modes of phthalazine-hydrazones (B, C), benzo[g]phthalazine functionalized with a nitrogen-rich chain (D), and N¹-aryl-2-trifluoromethylbenzo[b]naphthyridin-4(1H)-ones (E).

However, the parasite has developed a defense mechanism against quinoline-based drugs, which consists of preventing the accumulation of this type of agent within the vacuole of the parasite through the development of a mutation known as PfCRT, a membrane protein placed into the digestive food vacuole.^6,13−15 It has favored the resistance and survival ability of the parasite by treatment with quinolone-based agents. To counteract it, some structurally complex 4-aminoquinolines bearing a CQ-reverser such as verapamil, promethazine, chlorpheniramine, or primaquine have been developed and proved against resistant Plasmodium strains,¹⁶ but that strategy remains insufficient. As a consequence, the discovered novel nonquinolinic systems targeting the inhibition of β-hematin crystallization with the ability to prevent the resistance mechanism of the parasite are highly needed as an alternative to 4-aminoquinolines.

To design a solid strategy, it is important to take into account the possible routes relative to the blocking of β-hematin crystallization. From experimental evidence, there are two interacting routes to block β-hematin crystallization: (i) π–π stacking between the aromatic ring of the inhibitor and the porphyrin ring of the heme group by using fused-aromatic systems and (ii) direct interaction of the iron(III) metal with the heme group complex. Justly, from Egan’s report, 4-aminoquinolines act through a combined interacting route, finding proximity between the iron metal of the heme group and quinolinic nitrogen from theoretical simulation.⁷ Then, the ability of the organic ligand to form a stable complex with iron(III) may be an interesting strategy to identify new nonquinoline antimalarials targeting β-hematin crystallization. The heme group contains an iron(III) metallic atom and its potential interaction with a coordinative agent may contribute to the degradation of the β-hematin dimer, releasing the desired toxic and soluble heme group within the parasite vacuole (Figure 1A).

From the literature, we found that diverse types of heterocyclic systems (e.g., pyridine, xanthones, imidazoles, phenothiazinium salt, quinolin-4(1H)-ones, chromones, lumefantrine, halofantrine, and benzimidazoles) featuring a coordinative N-alkyl chain have been demonstrated to be good inhibitors of the β-hematin crystallization with relevant antimalarial activity.¹⁷⁻²⁰ Recently, a number of heterocycles bearing a highly coordinative chain-like hydrazone moiety have shown good inhibitory activity toward the β-hematin crystallization pathway.²¹⁻²⁶ With all information in mind, herein, we proved the ability of a series of highly coordinative heterocycles from a small molecule library to inhibit the formation of β-hematin as well as its potential as an antimalarial by using an in vivo model of P. berghei. The library is composed of 66 active compounds including a hit of 1-arylhydrazinyl-phthalazine, a hit of 1,4-diarylhydrazinyl-phthalazine, a hit of 3-aryl-6-(N′-methylpiperazine)[1,2,4]triazolo[3,4-a]phthalazine, a hit of weak coordinative N¹-aryl-2-trifluoromethylbenzonaphthyridin-4(1H)-ones, and a hit of noncoordinative 2-arylquinazolin-4(3H)-ones as a negative control. All these compounds were previously synthesized and evaluated against in vitro models of Trypanosome cruzi and Leishmania spp. parasites.²⁷⁻³³ These parasites are responsible for Leishmaniasis and Chagas disease, respectively, and they, in conjunction with malaria, belong to the group of neglected tropical diseases.³⁴ Despite the genetic and host preference differences between the trypanosomatids and the Plasmodium parasites,³⁴ it is common to find antitrypanosomatid agents with good antimalarial activity.^35,36

Results and Discussion

Coordinative Character of the Tested Compounds

Our principle for the discovery of new inhibitors of β-hematin crystallization was based on the complexation ability of the molecular system with transition metals, in particular with the iron(III) ion. The latter is essential because the complexation ability of the ligand toward the iron(III) ion of the heme group may compromise the intermolecular interaction between multiple heme groups within the β-hematin crystal (e.g., π–π stacking between porphyry rings), affecting its molecular integrity and releasing the toxic heme group (Figure 1A).³⁷ In our study, we selected a group of five families of compounds (1a–1t, 2a–2h, 3a–3f, 4a–4u, and 5a–5l): three of them as flexible systems possessing coordinative nitrogen chains (1a–1t, 2a–2h, and 3a–3f), a semiflexible ligand like N¹-aryl-2-trifluoromethylbenzo[b]naphthyridin-4(1H)one, which may be able to form a half-sandwich complex (Figure 1E), and finally, a noncoordinative system of 2-arylquinazolin-4(3H)ones (Figure S1). The purpose of that selection is to analyze the relevance of the coordination character on the inhibition of β-hematin crystallization for an in vitro assay.

To have a preliminary knowledge relative to the coordination ability of the ligand with iron(III), we made an analysis from the HSAB (hard soft acid–base) theory. From the HSAB theory, the iron(III) metal is considered as a hard metal with a borderline character in some cases.³⁸ Then, analyzing the coordination character of our ligands, iron(III) should be able to form a stable complex with the ligands 1a–1t and 2a–2h because they contain a borderline basic moiety such as the hydrazinyl scaffold with a chelating environment that favors the stabilization of the metallic complex as bidentate and tetradentate ligands, respectively.³⁸ It is well documented that tetradentate ligands like 1,4-diarylhydrazinylphthalazines are able to form mononuclear and dinuclear complexes with hard transition metals like vanadium(IV), molybdenum(IV), and molybdenum(V) (Figure 1B).³⁹⁻⁴² Other complexes of Hg(II), Ni(II), Co(II), Pd(II), Cu(II), Ag(I), Zn(II), Cd(II), Mn(II), and Fe(III) have been reported for the monoarylhydrazinyl-phthalazines in a solid state or solution (Figure 1C).⁴³ Other phthalazine analogues featuring a nitrogen chain at the 1- or 4-position of the pyridazine core have been demonstrated to be able to form a complex with hard/borderline transition metals like manganese(II), cobalt(II), cobalt(III), iron(II), nickel(II), copper(II), zinc(II), and mercury(II).⁴⁴⁻⁵¹ All evidence highlights the potential of the phthalazines 1a–1t and 2a–2h to form a complex with iron^2+/3+ species.

Regarding 3-aryl-6-(N′-methylpiperazine)-1,2,4-triazolophthalazines 3a–3f, we expected that they may act as a bidentate ligand with a hard character, which may complex with iron(III) through N′-methyl terminal nitrogen and N-heterocycle at the 5-position of the 1,2,4-triazolophthalazine core. Some examples of dinuclear copper(II) complexes in solution have been reported for phthalazine systems featuring dialkylamine chains (Figure 1D).⁵² Recently, we reported that the 3-aryl-6-(N-methylpiperazine)-1,2,4-triazolophthalazine system acts as an inhibitor of the potassium channel for in vitro models of cancer cells, which suggests a tentative interaction of the potassium ion with the architecture of the mentioned heterocycle.⁵³

Meanwhile, the N¹-(4-chlorophenyl)-2-trifluoromethyl-8-methoxybenzo[b]naphthyridin-4(1H)one is a relatively novel chemical system, and no reports exist about its complexation ability with transition metals. From theoretical calculations herein, we found that the N¹-aryl-2-trifluoromethylbenzo[b]naphthyridin-4(1H)ones can interact selectively with the iron metal to form a complexation between quinolinic nitrogen-10 with iron(III) and iron(II), which is the metal atom that is close to the N-arene ring. It opens the door to a tentative half-sandwich complex as illustrated in Figure 1E. Recently, we found that 5k showed a selective quenching with the cesium cation, with an apparent dominant interaction on the methoxy moiety.⁵⁴ Finally, noncoordinative 2-arylquinazolin-4(3H)-ones were used as a negative control (see structures in Figure S1) and noncomplexation interaction with metals is expected. Then, the ability of these target chemical systems to coordinate with iron^2+/3+ is essential for its potential application as an inhibitor of β-hematin crystallization within our proposal. The coordination ability of all these heterocyclic systems in conjunction with their good antiparasitic activity against trypanosomatids makes them valuable systems for studying their inhibitory activity on β-hematin crystallization and antimalarial response against the Plasmodium parasite.

In Vitro Studies on the Heme Crystallization

Initially, we focused on the ability of the 66 compounds to inhibit β-hematin crystallization by using the reported method with a few modifications (Tables 1 and 2).⁵⁵⁻⁵⁷ All these compounds were previously prepared and characterized by our group.^{27−33,58,59} From the 4-chloro-1-phthalazine-aryl/heteroarylhydrazinyl derivatives, four of them (1a, 1d, 1e, and 1h) displayed moderate inhibitory responses with IC₅₀ values ranging from 11.34 to 15.93 μM. Phthalazines 1f and 1k exhibited modest inhibitory responses of IC₅₀ = 28.79 and 28.80 μM, respectively, whereas the rest of phthalazine-hydrazones including those constructed from acetophenone, i.e., 1n–1t, displayed poor inhibitory responses (IC₅₀ > 50 μM). In particular, the non-inhibitory activity of nitro-derivatives 1j, 1l, and 1m can be attributed to its poor solubility in aqueous solution, whereas the non-activity of the phthalazine featuring 2-hydroxyphenyl 1i may be associated with the Z-isometry of the ylidenic bond. It is important to mention that geometric configuration is modestly able to form stable complexes with transition metals. The rest of 1-arylhydrazinyl-phthalazines 1a–1m presented an E-isometry. Regarding monosubstituted phthalazines 1n–1t, their modest inhibitory responses can be attributed to their restricted ability to form a stable complex with metals. Details are discussed later.

Table 1. Inhibitory Activity on Heme Crystallization (IHC) for the 1-Arylhydrazinyl-phthalazines (1a–1t) with Corresponding Log P Values.

^aRed bold text indicates up-borderline Log P according to Lipinsky’s rule.

Table 2. Inhibitory Activity on Heme Crystallization (IHC) for the 1,4-Arylhydrazinyl-phthalazines 2a–2h, 1,2,4-Triazolophthalazines 3a–3f, 2-Arylquinazolin-4(3H)ones 4a–4u, N¹-Aryl-2-(trifluoromethyl)benzo[b][1,8] Naphthyridin-4(1H)-ones 5a–5l, and Chloroquine (CQ) with Corresponding Log P Values.

^aRed bold text indicates up-borderline Log P according to Lipinsky’s rule.

Talking about the disubstituted phthalazin-hydrazones (2a–2h), these compounds exhibited the most significant inhibitory activity on the heme crystallization, finding good inhibitory activities for six derivatives (2a, 2b, 2d, 2e, 2f, and 2h) with IC₅₀ values ranging from 5.30 to 8.51 μM. That inhibitory response is barely comparable to that found for the chloroquine reference (IC₅₀ = 1.22 μM) (Table 2). In particular, the hexadentate 1,4-disubstituted phthalazines 2e (IC₅₀ = 5.30 μM) and 2f (IC₅₀ = 5.50 μM) displayed the highest inhibitory response within the 1,4-disubstituted phthalazines (Figure 1A), followed by compounds 2d (IC₅₀ = 5.80 μM), 2b (IC₅₀ = 7.10 μM), 2a (IC₅₀ = 7.24 μM), and 2h (IC₅₀ = 8.51 μM). The non-inhibitory activity found for nitro-derivatives 2c and 2g seems to be associated with their poor solubility in aqueous solution.

Regarding the 1,2,4-triazolophthalazine system (3a–3f), only two of them, in particular, those bearing N-methylpiperazine substitution at the 6-position, 3b and 3c, exhibited moderate responses with IC₅₀ values of 15.60 and 17.66 μM, respectively (Table 2). The replacement of the N′-methylpiperazine substitution by a chlorine atom at the 6-position of phthalazine-1,2,4-triazolo led to a reduction of the inhibitory activity (3a in Table 2). These findings suggests that N′-methylpiperazine seems to play an important role in the inhibitory response of the phthalazine-1,2,4-triazolo system. As was mentioned earlier, we demonstrated that this type of system bearing N-methylpiperazine acts as a potassium channel blocker, suggesting its ability to interact with metals.⁵² Meanwhile, N¹-aryl-benzo[b]naphthyridin-4(1H)-ones showed a moderate inhibitory response toward heme crystallization, identifying two compounds, 5a and 5k, with good IC₅₀ values of 9.7 and 11.8 μM, respectively (Table 2). The rest of studied N¹-aryl-benzo[b]naphthyridin-4(1H)-ones showed comparable inhibitory responses to each other with IHP values from 26 to 39% at a 50 μM drug concentration. It seems clear that the 8- or 7-methoxy substitution on the benzo-core plays an important role in the inhibitory activity of these interesting compounds because the majority of their derivatives (5a, 5b, 5j, or 5 k) presented IHP values higher than 35% (with a few exceptions). The nonsubstituted derivatives (5g) or methyl-substituted derivatives (5d, 5f, and 5i) showed more discrete IHP responses from 9 to 26%. It is important to mention that the 7-methoxy derivatives (3j, 3k, and 3l) were consistently more active than 7-methoxy derivatives, which suggested that the disposition of the methoxy group at the 7-position of the benzo-core seems to be relevant in the inhibition of β-hematin crystallization. Recently, we found that the 7-methoxy derivatives showed more efficient π–π stacking between two molecules of benzo[b]naphthyridin-4(1H)-one than 8-methoxy derivatives from X-ray crystal structure packing (Figure S2), which was associated with a homogeneous distribution of Mulliken charge along the benzo[b]naphthyridin-4(1H)-one ring (Figure S3). It may play an important role in the inhibition of β-hematin through π–π stacking between benzo[b]naphthyridin-4(1H)-one and one porphyrin molecule as has been described for 4-aminoquinolines. It is important to mention that the direct interaction of the quinolinic nitrogen-10 of the ligand with the iron(III) metal is drastically compromised by the presence of vicinal N¹-arene, which is more remarkable by the hindrance environment of the iron(III) metal into the heme group. As a comparative reference, some analogues of our N¹-aryl-benzo[b]naphthyridin-4(1H)one such as the floxarcrine or chromones containing a trifluoromethyl moiety have demonstrated good inhibitory responses toward the heme crystallization, which put in evidence of the role of the quinolin-4(1H)one ring for the design of β-hematin inhibitors.²⁵

Finally, as we expected, none of the noncoordinative 2-arylquinazolin-4(3H)-ones 4a–4u presented inhibitory responses at a drug concentration of 50 μM. All these findings reflect that the coordinative character of the tested molecules seems to play a pivotal role in the inhibitory response toward heme crystallization, finding the highest response for highly coordinative 1,4-disubstituted phthalazine derivatives 2a–2h, followed by a moderate coordinating system like monosubstituted phthalazines 1a–1t, 1,2,4-triazolophthalazines 3b and 3c, and N¹-aryl-benzo[b]naphthyridin-4(1H)ones 5a–5l, and a non-inhibitory response for the noncoordinative 2-arylquinazolin-4(3H)-ones 4a–4u.

In Vitro Toxicity Assay

The toxicity assay was performed for the most prominent inhibitors 1d, 1e, 1h, 2a, 2b, 2d, 2e, 2f, 2h, 3b, 3c, 5a, and 5k by using peritoneal murine macrophages. The latter were obtained from the mouse peritoneal cavity.³⁰⁻³³ Toxicity results (expressed in terms of CC₅₀) are summarized in Table 3. Among the studied compounds, 1d, 3b, 3c, and 2f exhibited the lowest toxicity against macrophage cells with CC₅₀ values higher than 100 μM, followed by the N¹-aryl-benzo[b]naphthyridin-4(1H)ones 5a and 5k with CC₅₀ values of 67.54 and 40.53 μM, respectively, the 1,4-disubstituted phthalazines 2a–2f with CC₅₀ values from 31.56 to 56.32 μM, and the monosubstituted phthalazines 1e and 1h with CC₅₀ values of 29.43 and 45.54 μM, respectively. Finally, the nitro-derivative 2h showed the highest toxicity with a CC₅₀ value of 25.16 μM. Toxicity results were matched with their corresponding IHC (inhibition of β-hematin crystallization) response to obtain a relative selectivity index (S.I.), which was used as a criterion to select the best candidate for in vivo efficacy studies. Consequently, only the best candidate with the highest S.I. from each compound group was selected for further assays. Then, compound 1d (S.I. > 10.6) was chosen from group 1 over compounds 1e (S.I. = 2.1) and 1h (S.I. = 3.2). From group 2, compounds 2f (S.I. = 20.3), 2d (S.I. = 9.7), 2e (S.I. = 8.0), and 2a (S.I. = 5.5) were selected as the best candidates. Derivative 2b, despite its modest S.I. (4.5), was chosen for further assays, taking advantage of its potential as a leishmanicidal agent against infective strains of the Leishmania parasite.³⁰ Within group 3, despite the similar inhibitory or cytotoxicity response found between compounds 3b and 3c, 3c was chosen for further assay based on its recognized antiparasitic activity. Previously, we demonstrated the potential of 3c as a leishmanicidal agent with significant in vitro activity against the infective and resistant strain of the Leishmania parasite, which was higher than that found for derivative 3b.³¹ Finally, compound 5a (S.I. = 7.0) was chosen over 5k (S.I. = 3.4) from group 5.

Table 3. Inhibitory Response against β-Hematin Crystallization (IHC), Toxicity against Murine Peritoneal Macrophages, Selectivity Index (S.I.), Hemolysis Activity, Antimalarial Response for the Infected Model of P. berghei, and Partition Coefficient (Log P) for the Most Prominent Inhibitors 1d, 1e, 1h, 2a, 2b, 2d, 2e, 2f, 2h, 3b, 3c, 5a, and 5k^a.

^aResults are expressed as the media ± standard deviation (SD).

^bIHC: inhibition heme crystallization, expressed as IC₅₀, for the most active derivatives.

^cRelative toxicity was determined on murine peritoneal macrophages.

^dSelectivity index (S.I.) from IC₅₀ of inhibition toward β-hematin and CC₅₀ (S.I. = IC₅₀/CC₅₀).

^eRelative hemolysis percent response was determined on red blood cells at 100 mM. Further hemolysis response data at 1 and 10 mM drug compound concentrations are listed in Table S1.

^fIn vivo % P: percentage of parasitemia (compound dosage for in vivo assay was 20 mg kg^–1).

^gPost-infection days determined by using infected mice with P. berghei, n = 6 (number of treated mice).

^hCalculated from the SwissADME software website, ref (61). *P < 0.05 compared to the control-treated group. **P < 0.01 compared to the control-treated group.

ⁱControl negative [C(−)] represents the experiment in the absence of any compound or reference drug.

^jPositive control treated with chloroquine (CQ) (20 mg kg^–1).

In Vitro Hemolysis Assay

To discard any potential toxic effect on red blood cells (RBCs) by the highly coordinative ability of the tested compounds, we performed hemolysis assays using human red blood cells (RBCs) for the most active compounds 1d, 2a, 2b, 2d–2f, and 2h. It is significantly relevant for the most active disubstituted phthalazines (2a–2h) because they justly can easily form a complex with metallic species³⁹⁻⁴² including the iron present in red blood cells (RBCs) and other metabolic enzymes like cytochrome family enzymes. An assay was performed at three compound concentrations (1, 10, and 100 mM). The hemolysis data are listed in Table 3, and further information details can be found in Table S1. From the results, it should be noted that all active disubstituted phthalazines 2a, 2b, 2d, 2e, 2f, and 2h displayed a discrete hemolysis (19–35%) response at a high compound concentration (100 mM), which was in general comparable to that response found for the CQ (28.73%). In particular, derivative 2b disclosed the lowest hemolysis activity, whereas the nitro-derivative 2h displayed the highest response, which was in good correlation with its relative in vitro toxicity against macrophages. The latter compromises the applicability of derivative 2h as an antimalarial agent. At a 1 mM compound concentration, low hemolysis activity from 5.20 to 12.40% was found for the 1,4-disubstituted derivatives, having the lowest hemolysis for compound 2b and the highest for compound 2h. Candidate derivatives 2d and 2f displayed hemolysis response values of 8.70 and 9.2% at a 1 mM compound concentration, respectively, which were comparable to that found for the chloroquine reference (7.40%) (Table S1). Meanwhile, derivative 1d showed a lower hemolysis (11.1%) response than 1,4-disubstituted aryl-hydrazinyl-phthalazines 2a–2h (19–35%) at a 100 mM compound concentration. From the hemolysis assay, most of tested 1,4-disubstituted aryl-hydrazinyl-phthalazines 2a–2h as well as compound 1d were considered safe against RBCs for further biological assays.

In Vivo Efficacy

A selected group of compounds with low toxicity and good inhibitory response against β-hematin crystallization were tested in vivo in the P. berghei-infected model. The selected compounds consist of a series of compounds with S.I. higher than 5 and low hemolysis response including (i) the best β-hematin inhibitors of 1,4-disubstituted phthalazines (2a, 2b, 2d, 2e, and 2f) and (ii) moderate and nontoxic inhibitors of 1-substituted phthalazine-hydrazone (1d), 1,2,4-triazolophthalazine (3c), and N¹-aryl 2-trifluoromethyl-benzo[b]naphthyridin-4(1H)-ones (5a). The in vivo model consisted of mice infected with P. berghei ANKA, a chloroquine-susceptible strain of murine malaria (Table 1). Efficacy in the P. berghei assay was determined following oral administration of 20 mg kg^–1 day^–1 for 4 consecutive days (4 day Peters’ test). The parasitemia was determined at 4 day post-infection, and subsequently, the survival days for treated mice with compounds were monitored and compared with negative untreated mice for symptoms for up to 30 days.⁶⁰ Experimental details can be found in the Experimental Section and Supporting Information. The efficacy results are summarized in Table 1. The untreated mice died between 7 and 8 days post infection, whereas treated mice displayed a superior survival time in conjunction with a diminution of parasitemia. In general, all tested compounds were able to increase the survival time post infection in conjunction with a reduction of the parasitemia compared to the untreated control (9 survival days and parasitemia index by about 79%). Within the tested compounds, 1,4-disubstituted phthalazine-hydrazones (2) were able to enhance the survival time over 20 days post infection (average mice survival days from 24 to 28 days) with a significant reduction of parasitemia, having the lowest parasitemia percentages in treated mice from 0.23 to 1.14% (parasitemia reduction percentage from 99.77 to 98.86%). From the 1,4-disubstituted phthalazines, two of them (2d and 2f) afforded a complete cure (3/3 mice cured and survival days of 30) with parasitemia reduction by about 99.77 and 99.68%, respectively, at day 4. The efficacy of the most active 1,4-disubstituted phthalazines (2d, 2e, and 2f) was comparable to that found for the chloroquine reference with a parasitemia index of 0.2% (3/3 mice with parasitemia reduction of 99.80%) and a survival time of >30 days, making these 1,4-disubtituted phthalazines potential antimalarial agents targeting the inhibition of β-hematin crystallization. Meanwhile, the rest of derivatives (no 1,4-disusbtituted phthalazines) displayed modest antimalarial activity with a survival time from 14 to 16 days and a parasitemia percentage in infected mice of about 27–29% post infection. Interestingly, neither of the structural analogues of the 1,4-disubstituted phthalazines, the monosubstituted phthalazines 1d and the cyclic version 3c, were efficacious at 20 mg/kg drug dosage, which suggests that the 1,4-disusbtitution is pivotal in the antimalarial activity of the studied phthalazines.

In general, a good correlation between the activity against the P. berghei model and the inhibitory activity against β-hematin crystallization was found for the studied derivatives, suggesting that the inhibition of heme crystallization is the main mode of action of these most active compounds, in particular, for the 1,4-disubstituted phthalazine-hydrazones. A general correlation between the inhibitory activity and the parasitemia index and survival times is shown in Figure 2, which illustrates clearly the important role of the inhibition of β-hematin in promoting a high curative response with reduction of parasitemia (Figure 2A,B). From Figure 2, the best inhibitors 2d, 2e, and 2f displayed a high curative response and reduction of parasitemia, close to that found for CQ, whereas modest inhibitors 1d, 3c, and 5k exhibited the lowest curative and antiparasitic response.

Figure 2.

Correlation between the inhibitory response toward β-hematin and survival days post infection (A) and parasitemia index (B) for the most active compounds.

In definition, a comparison between in vitro β-hematin inhibition, antimalarial activity (survival time and parasitemia), hemolysis, toxicity, and Log P reflects that derivatives 2d and 2f emerged as the most promising agents with a good inhibitory response toward heme crystallization (IC₅₀’s of 5.80 and 5.50 μM for compounds 2d and 2f, respectively), moderate toxicities (CC₅₀’s of 56.32 and 111.45 μM for compounds 2d and 2f, respectively), hemolysis (less than 22%), post-infection day survival (28–29 days), parasitemia indexes (0.230 and 0.320 for compounds 2d and 2f, respectively), and partition coefficient (Log P < 3). Compound 2e showed an interesting profile, but it showed higher toxicity (CC₅₀ = 42.51 μM), barely higher parasitemia index for treated mice (0.410), and partition coefficient (Log P = 3.13).

In Silico ADME Parameters⁶²

In an attempt to rationalize the varying in vivo efficacies, an in silico analysis of the physicochemical, pharmacokinetic, and druglikeness parameters through the SwissADME website⁶¹ was performed for the best candidates of each compound family (Table 2 and Tables S1–S4 in the Supporting Information). Focusing on three more active derivatives 2d, 2e, and 2f, compound 2f showed the best profile with (i) the best water solubility, about 10-fold more than those of derivatives 2d and 2e, (ii) higher gastrointestinal (G.I.) absorption, (iii) a lower inhibitory activity toward cytochrome proteins, exhibiting only a tentative activity on CYP1A2 and CYP2C19 proteins, (iv) no violations of druglikeness principles, and (v) the absence of a PAINS (pan assay interference structure) moiety. Derivative 2e seems to be the least convenient candidate because it presents a PAINS moiety such as the hydrizinyl-phenol ring as well as the violation of two lead-likeness principles (MW > 350, XLOGP3 > 3.5). Meanwhile, despite the different drawbacks found for compound 2d including (i) the violation of three lead-likeness principles (MW > 350, Rotors > 7, XLOGP3 > 3.5) and (ii) two alerts within Brenk’s ruler (imine and nitro groups) and large TPSA (166.20 Ų), these compounds showed the best survival time for treated mice and a lower parasitemia index post infection, which are comparable to the CQ reference. The latter reflects that the tentative side effects and solubility limitation would not seem to be dominating for the in vivo model, thus being a candidate to be taken into account for further assays. From the results, it should be noted that compounds 2d and 2f with a better ADME profile showed the most significant in vivo antimalarial response (Table 4). The discrete antimalarial activity and curative properties of derivatives 2a, 2b, and 2h can be associated with the higher Log P of derivatives 2a and 2b and lower water solubility of 2h (details in Tables S1–S4). Meanwhile, despite the good ADME profile and good simulated pharmacokinetic properties and druglikeness found for derivatives 1d and 3c, their low in vivo efficacy may be mainly attributed to the lack of an intrinsic biological activity of them against the Plasmodium parasite (Table S4). Also, finally, the N¹-(aryl)-2-(trifluoromethyl)-8-methoxybenzo[b][1,8]naphthyridin-4(1H)-ones 5a and 5k showed a good ADME, pharmacokinetic, and druglikeness profile from the in silico study, although they exhibited a low aqueous solubility. It may be one of the main physicochemical factors that affect the in vivo efficacy of these compounds, although the lack of intrinsic antiparasitic activity cannot be discarded.

Table 4. In Silico Physicochemical, Pharmacokinetic, and Druglikeness Parameters of the Prominent Antimalarial Candidates 2d and 2f.

type of parameter	parameter	2d	2f
physicochemical properties	MW (g/mol)a	456.41	346.34
	no. of rotatable bonds	8	6
	no. of H-bond acceptors	8	6
	no. of H-bond donors	2	2
	molar refractivity	132.14	99.03
	TPSA (Å2)b	166.20	100.84
delipophilicity	Log Po/w (iLOGP)	1.49	1.56
	Log Po/w (XLOGP3)	4.65	3.79
	Log Po/w (WLOGP)	3.96	3.33
	Log Po/w (MLOGP)	2.90	1.88
	Log Po/w (SILICOS-IT)	–0.17	2.92
	Consensus Log Po/w	2.56	2.69
water solubility	Log S (ESOL)	–5.55	–4.55
	solubility	2.82 μM (moderately soluble)	28.3 μM (moderately soluble)
	Log S (Ali)	–7.87	–5.60
	solubility	13.6 nM (poorly soluble)	2.50 μM (moderately soluble)
pharmakinetic properties	G.I. absorptionc	low	high
	BBB permeantd	no	no
	P-gp substratee	no	no
	CYP1A2 inhibitor	no	yes
	CYP2C19 inhibitor	yes	yes
	CYP2C9 inhibitor	yes	no
	CYP2D6 inhibitor	no	no
	CYP3A4 inhibitor	yes	no
	Log Kp (skin permeation)	–5.78 cm/s	–5.72 cm/s
druglikeness	Lipinski	yes; 1 violation: NorO > 10	yes; 0 violation
	Ghose	no; 1 violation: MR > 130	yes
	Veber	no; 1 violation: TPSA > 140	yes
	Egan	no; 1 violation: TPSA > 132	yes
	Muegge	no; 1 violation: TPSA > 150	yes
	bioavailability score	0.55	0.55
medicinal chemistry	PAINSf	0 alert	0 alert
	Brenk	2 alerts: imine_1, nitro_group	1 alert: imine_1
	leadlikeness	no; 3 violations: MW > 350, Rotors > 7, XLOGP3 > 3.5	no; 1 violation: XLOGP3 > 3.5

^aMW: molecular weight.

^bTPSA: topological polar surface area.

^cG.I.: gastrointestinal.

^dBBB: blood–brain barrier.

^eP-gp: P-glycoprotein.

^fPAINS: pan assay interference structure. The rest of in silico data can be found in Tables S1–S5 of the Supporting Information. Fundamental definition of each parameter can be found in ref (62).

Role of the Complexation with the Iron Metal

To confirm the role of the complexation ability of the ligand in the inhibition of β-hematin crystallization, we studied the ability of the most active derivative to form a complex with the iron(III) ion from theoretical and experimental studies. Iron(III) is the main oxidation state present in the iron atom featuring the β-hematin crystal. Extra complexation calculations of the ligands were performed for iron(II) ions to explore the relative coordinative selectivity of the ligand because iron(II) is present in important human enzymes, cytochrome, hemoglobins, etc. Regarding theoretical calculations, first, the geometries of the studied derivatives were optimized by using the B3LYP method⁶³ in combination with the 6-31G(d,p) basis set.⁶⁴ The calculations were implemented in Gaussian09 package software.⁶⁵ Subsequently, the interaction between the most active ligand and a molecule of iron(III) or iron(II) was performed by using the BSSE model-implemented counterpoise command.⁶⁶ All optimized ligand and complex structures can be found in Section VI of the Supporting Information. Some relevant structures and complexation energies are summarized in Figures 3 and 4 and Table 5, respectively. From the results, the disubstituted phthalazines (group 2), in general, showed the highest complexation ability among the studied compounds with complexation energies from −827 to −920 kcal/mol. Within group 2, nitro-derivatives 2c, 2d, or 2h showed the lowest complexation ability with complexation energies of −850.33, −853.76, and −827.15 kcal/mol, respectively, whereas compound 2f displayed the highest complexation energy (−920.19 kcal/mol). Meanwhile, the monosubstituted phthalazine 1d showed a complexation energy of −800 kcal/mol, which was comparable to those found for the disubstituted 1,4-arylhydrazinyl-phthalazines (−827 to −927 kcal/mol). It is expected because the complexation calculations for disubstituted phthalazines were performed by using one molecule of iron(III), although the higher complexation energies of disubstituted phthalazines suppose that they present a higher complexation ability than monosubstituted phthalazines. The 3-aryl-1,2,4-triazolophthalazines 3c displayed a complexation energy of −686.1 kcal/mol, whereas the trifluoromethyl-benzo[b]naphthyridin-4(1H)-one 3k showed a complexation energy of −301.1 kcal/mol. The rest of derivatives 1p, 3a, and 3d did not show convergence of the geometry of their corresponding iron complexes, putting in evidence of the poor ability of these compounds to form stable complexes with iron(III). For derivative 1p, the incorporation of the methyl group on the ylidenic bond seems to affect the coordination character of the ligand because the aryl ring tends to place in the coordination region (see 1p in page S19 of the Supporting Information and Figure 3for its iron complex). On the other hand, the complexation tendency of the ligands is comparable between iron(III) and iron(II), although the ligand showed a preference toward iron(III) over iron(II) with lower complexation energies (e.g., −800 vs −350 kcal/mol for 1,4-diarylhydrazinylphthalazines). It has great relevance because a high complexation preference toward iron(II) could more selectively favor the inhibition of β-hematin formation than the interaction with any essential iron enzyme or protein rich in iron(III).

Figure 3.

Structure of ligand–iron(III) complexes for the disubstituted phthalazin-hydrazones 2a–2f and monosubstituted phthalazin-hydrazones 1d and 1p.

Figure 4.

Structure of ligand–iron(III) complexes for the 3-aryl-6-(N′-methylpiperazine)-1,2,4-triazolophthalazines 3c and 3d and N¹-aryl 2-trifluoromethyl-benzo[b]naphthyridin-4(1H)-one 5k.

Table 5. Comparison between the Complexation Ability and TPSA Data of the Most Active Derivatives with Its Inhibitory Response toward β-Hematin Crystallization.

^aComplexation energy between the ligand and iron(II) calculated from B3LYP/6.31G(d,p).

^bComplexation energy between the ligand and iron(III) calculated from B3LYP/6.31G(d,p).

^cTPSA (topological polar surface area) calculated from SwissADME.

^dIC₅₀ for the inhibition heme crystallization. Red color, critical case; bold black text, highlighted results. N.D., not determined.

Experimental determination of complexation confirmed the theoretical findings (Table 5). In general, the 1,4-arylhydrazinyl-phthalazines disclosed the highest binding constants with values from 550 × 10⁷ to 3100 × 10⁷ M^–2. The 1,4-diarylhydrazinyl-phthalazines showed a stoichiometry ratio of 2, whereas the rest of ligands showed a stoichiometry ratio of 1. Within the highly coordinative 1,4-arylhydrazinyl-phthalazines, the highly conjugated nitro-derivatives 2c and 2h showed the lowest binding constants of 150 × 10⁷ and 550 × 10⁷, respectively, whereas 2d and 2f exhibited the highest affinity complexation with values of 3100 × 10⁷ and 1510 × 10⁷ M^–2, respectively. Regarding the rest of ligands, the monosubstituted 1-arylhydrazinyl-phthalazine 1d displayed a lower binding constant of 3200 M^–1 and derivative 5k displayed a binding constant of 30 M^–1, whereas the 1,2,4-triazolophthalazines displayed the lowest complexation ability with a binding constant of less than 10 M^–1 for compounds 3c and 3d. The CQ showed a discrete binding constant of about 6300 M^–1, which suggests that the inhibitory response is not strictly dependent on its complexation ability toward the iron(III) metal, but also other chemical functionality may play an important role. It is documented that the terminal tertiary amine moiety is interacting with the terminal carboxylate group of the porphyry ring of the heme group, thus with the presence of the alkyl group relevant for a high inhibition of the β-hematin formation.⁷

Regarding the nature of complexes, as expected, the mono- or di-arylhydrazinyl-phthalazines (groups 1 and 2) are able to coordinate with the iron metal through imine nitrogen and nitrogen-2 of the pyrazine ring to form a five-membered cycle (Figure 3) with complexation energies of less than −800 kcal/mol and a binding constant of 3200 for monosubstituted compounds and in the 10⁸ range for disubstituted phthalazines. Meanwhile, the triazolophthalazine ligands bearing N′-methylpiperazine showed the iron(III) metal placed close to the piperazine ring, but it is not sufficient to form a stable complex (no convergence of complex geometry) (Figure 4), whereas the N¹-aryl 2-trifluoromethyl-benzo[b]naphthyridin-4(1H)-one coordinates through quinolinic nitrogen-10 and a vicinal arene ring to form a classical half-sandwich complex and that mode of coordination allows one to interpret the high complexation ability of the rigid benzo[b]naphthyridin-4(1H)-one (5k in Figure 3). It is important because the complexation character of the ligand plays an important role in the inhibition of β-hematin crystallization via retention of the iron metal. From calculation and experimental studies, the least coordinative ligands, the 1,2,4-triazolophthalazines 3a, 3c, and 3d, showed the lowest inhibition response, whereas the highest coordinative ligands, the phthalazine-hydrazones (groups 1 and 2), displayed the highest inhibition response from 5 to 10 μM. In particular, the double coordinative character of the disubstituted ligand to form dinuclear complexes may favor the ability of these compounds to interact more efficiently with the iron metal than the monosubstituted ligand, which can help interpret the highest inhibitory response of disubstituted phthalazine (5 μM) compared to monosubstituted phthalazine (>10 μM) (Tables 1 and 2). A weak coordinative ligand such as phthalazine-hydrazone 1p showed a non-inhibitory response. A weak ligand by incorporation of strong electron-deficient nitro-rings compromises an effective complexation, which was transduced in a weak inhibitory response (e.g., 2c, 2g, and 2h). Regarding the N¹-aryl 2-trifluoromethyl-benzo[b]naphthyridin-4(1H)-one, the proximity of the N-arene moiety to the structure may compromise the ability of the ligand to interact directly with the iron metal of the heme group, which is reflected in its lower binding constant of about 29 M^–1. That lower binding constant may be correlated with the moderate inhibitory response (IC₅₀ values of more than 9 μM) of the N¹-aryl 2-trifluoromethyl-benzo[b]naphthyridin-4(1H)-ones 5a and 5k. To establish a more consistent correlation between the inhibitory response and the complexation ability and surface extension of the ligand, we constructed K_avs IC₅₀. The graphical representation is illustrated in Figure 5A. From Figure 5A, the ligands with higher binding constants, 2d, 2e, and 2f, showed the most significant inhibitory response, whereas the less coordinative ligands 1p, 3a, and 3d did not a show inhibitory response. The inhibitory response showed a direct and consistent correlation with the complexation ability of the ligand to coordinate with iron(III), and the correlation K_avs IC₅₀ showed an exponential behavior, where a significantly high complexation character of the ligand guarantees a high inhibitory response toward β-hematin crystallization.

Figure 5.

Correlations between the inhibitory response toward β-hematin crystallization and the binding constant of the ligand–iron(III) complexation (A) and the TPSA of the ligand (B). Blue points remark the most promising compounds, red points indicate the less prominent correlation, and green points are taken as exceptions to the general correlations.

Beyond the role of the complexation ability of the ligand, it should be noted that the extension or surface area of the ligand may be relevant for the inhibition process to favor ligand–hematin interaction via an effective π–π stacking. To clarify that observation, we constructed a TPSA vs IC₅₀ plot, and the graphic is shown in Figure 5B. From Figure 5B, the most extended compounds with high TPSA (>100 Å) showed the highest inhibitory response, with an exception of the nitro-derivatives. Similar to the K_avs IC₅₀ tendency, the TPSA vs IC₅₀ exhibited a pseudo-exponential behavior, showing a high inhibitory response with an increase in TPSA magnitude. Less extended ligands such as 1p (49.11 Å) and 3a (40.32 Å) did not show an inhibitory response, whereas other ligands like 1d (50.17 Å), 3c (49.56 Å), and 5k (44.12 Å) showed a modest inhibitory response. Then, the location at the left position of the TPSA vs IC₅₀ plot for compounds 2d, 2e, and 2f puts in evidence of the important role of the extension of the molecule for a high inhibitory response. The planar molecular disposition and high extension of the surface of the 1,4-disubstituted phthalazine with TPSA values of more than 100 Å (Table 2 and Tables S2–S5) seem to be convenient features to favor the interaction with β-hematin through a π–π stacking with the porphyrin ring. It can favor the rupture of π–π stacking between porphyrin rings into β-hematin. All evidence suggests that the 1,4-disubstituted phthalazines act as an inhibitor of β-hematin, which are unique chemical systems among studied compounds that present high TPSA values and are able to form a stable complex with iron(III). Our results showed that these two features seem to be important to promote high inhibition and they can be essential for the rational design of β-hematin inhibitors. Probably, the inhibition of the β-hematin crystallization initially involves a π–π stacking between the ligand core and the porphyrin ring of the heme group, and the strength of the hematin–ligand interaction is increased by the interaction between the ligand and the iron(III) placed into the heme group. Both interacting events are favored by the high planarity, extension of the molecular system, and its highly coordinative character with two equivalents of possibility to form polynuclear complexes.

Conclusions

In summary, we evaluated the potential of a series of highly coordinative ligands with known leishmanicidal activity for the inhibition of the formation of β-hematin and antimalarial activity against an in vivo model of P. berghei. In particular, a group of highly coordinative phthalazines bearing coordinative moieties at 1- and 4-positions showed a high inhibitory response against the formation of β-hematin as well as low toxicity on macrophages and low hemolysis response on human RBCs. Two of the 1,4-disubstituted phthalazines (2d and 2f) disclosed a good curative response and significant reduction of parasitemia from the in vivo model of P. berghei, with all these findings comparable to those found for the CQ reference. Compound 2f displayed the best toxicity and ADME profiles, which is a promising candidate for further biological experiments including in vitro assay against CQ-sensitive and CQ-resistant Plasmodium falciparum models, genotoxicity evaluation, pharmacokinetic studies, and in vivo assays.

The significant inhibitory response for the formation of β-hematin seems to be associated with the complexation ability of the active ligands with iron(II) and iron(III) metals. A consistent correlation between the complexation ability of the ligand and the inhibitory response toward β-hematin crystallization was found for a group of active compounds, having the highest inhibitory response for the best ligands with the highest binding constants with iron(III). Alternatively, a good correlation between the TPSA and the inhibitory response of the ligand was found, finding a high inhibition response for ligands with high TPSA values. The present investigation opens new perspectives toward the rational design of antimalarial agents targeting β-hematin crystallization, which is the highly coordinative character of the ligand to form a stable complex with iron(III) and its high molecular surface with planar disposition, two relevant features to take into account for future design of β-hematin inhibitors.

Experimental Section

Materials and Instruments

The (E)-1-(2-benzylidenehydrazineyl)-4-chlorophthalazines 1a–1t (Z-isomer for 2i), 1,4-bis(2-((E)-benzylidene)hydrazinyl)phthalazines 2a–2h, 3-aryl-6-(N′-methylpiperazine)[1,2,4]triazolo[3,4-a]phthalazines 3a–3f, 2-arylquinazolin-4(3H)-ones 4a–4u, N¹-(aryl)-2-(trifluoromethyl)-8-methoxybenzo[b][1,8]naphthyridin-4(1H)-ones 5a–5l were previously prepared and characterized by our group.²⁷⁻³³ The chloroquine reference was purchased from Sigma-Aldrich and they were used without further purification for the biological analysis. Solvents (ethanol or DMSO) were purchased from commercial sources (Sigma-Aldrich or Merck). Absorption data for the β-hematin crystallization assay were obtained from a Thermo Scientific Varioskan Flash Multimode instrument for air-equilibrated solutions at 25 °C.

Biological Experimental Information

Peritoneal macrophages were obtained from the mouse peritoneal cavity and differentiated in a conditioned medium (medium RMPI) as described in previous methodologies.³⁰⁻³³ The macrophages were provided by Instituto de Biomedicina “Dr. Jacinto Convit” (Hospital Vargas, Caracas, Venezuela). The P. berghei ANKA strain and mice for in vivo assay were provided by Instituto de Higiene (Universidad Central de Venezuela, Caracas). Mice (average body weight, 16.7 g) were infected with P. berghei parasites by intraperitoneal (IP) inoculation of donor mouse blood diluted in normal saline to contain 3 × 10⁶ infected red blood cells (RBCs). The BALB/c female mice used in this assay were 4 weeks old and pathogen-free (Instituto de Higiene, Universidad Central de Venezuela, Caracas).

Inhibition of β-Hematin Formation

The β-hematin formation assay was performed according to the literature.⁵⁵⁻⁵⁷ Briefly, a solution of hemin chloride (50 μL, 4 mM), dissolved in DMSO (5.2 mg mL^–1), was distributed in 96-well microplates. Different concentrations (5–100 μM) of quinoline compounds dissolved in DMSO were added in triplicate in test wells (50 μL). Controls contained either water (50 μL) or DMSO (50 μL). β-Hematin formation was initiated by the addition of acetate buffer (100 μL 0.2 M, pH 4.4). Plates were incubated at 37 °C for 48 h to allow completion of the reaction and centrifuged (4000 rpm × 15 min, IEC-CENTRA, MP4R). After discarding the supernatant, the pellet was washed twice with DMSO (200 μL) and finally dissolved in NaOH (200 μL, 0.2 N). The solubilized aggregates were further diluted 1:2 with NaOH (0.1 N) and absorbance was recorded at 405 nm (Microplate Reader, BIORAD-550). The results were expressed as a percentage of inhibition of β-hematin crystallization (% IHC).

Parasite, Experimental Host, and Strain Maintenance

Male BALB/c mice weighing 18–22 g were maintained on a commercial pellet diet and housed under conditions approved by Ethics Committee. P. berghei (ANKA strain), a rodent malarial parasite, was used for infection. Mice were infected by ip injection with 1 × 10⁶ infected erythrocytes diluted in phosphate-buffered saline solution (PBS, 10 mM, pH 7.4, 0.1 mL). Parasitemia was monitored by microscopic examination of Giemsa-stained smears.

Four Day Suppressive Test

BALB/c mice (18–22 g) were infected (using caudal vein) with 10⁶ infected RBCs with P. berghei (n = 6). Two hours after infection, treatment began with the best compounds tested resulting from in vitro assays. These compounds were dissolved in DMSO (0.1 M), diluted with saline-Tween 20 solution (2%). Each compound (20 mg kg^–1) was administered once by ip for 4 days. At day 4, parasitemia was counted by examination of Giemsa-stained smears. Chloroquine (20 mg kg^–1) was used as a positive control. The survival time beyond the control group (saline-treated) was recorded. The results were expressed as a percentage of parasitemia (% of parasitemia) and survival days of each compound-treated group over the control (saline-treated group).⁶⁰

Toxicity Assay

Peritoneal murine macrophages were grown in RPMI 1640 medium (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum, 1% l-glutamine, 1% streptomycin, and 100 units/mL penicillin (all obtained from Sigma-Aldrich, USA). Cell viability was assessed using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) for assay, which is based on the ability of viable cells to metabolically reduce a yellow tetrazolium salt (MTT; Sigma) to a purple formazan product. This reaction takes place when mitochondrial reductase enzymes are active. Cells were grown in 96-well plates (5 × 10⁴ cells/well) for 24 h. Cultures were carried out at 37 °C in a humidified atmosphere with 5% CO₂ and incubated with the most active compounds in 100 μL of complete culture medium containing 25, 50, 75, 100, and 200 μM concentrations of each compound for 48 h. After incubation, the medium was removed and the cells were treated with 100 μL of 0.4 mg/mL MTT for 4 h at 37 °C. Subsequently, 100 μL of DMSO was added to the mixture. The solubilized formazan product was quantified with the help of a microtiter plate reader TECAN-Sunrise at 570 nm. In all cases, the compounds were dissolved in DMSO at the final concentration in the culture medium lower than 1%, a concentration that neither had a cytotoxic effect nor caused any interference with the colorimetric detection method. All experiments were performed at least three times. Untreated control parasites were used to calculate the relative proliferation. Dose–response curves were recorded and the CC₅₀ values were determined using GraphPad Prism version 6.00 for Windows (GraphPad Software, La Jolla California, USA).⁶⁷ The results are presented as average SDs with SD below 10%.³⁰⁻³³

In Vitro Toxicity on Human Red Blood Cells (RBCs)

To evaluate the in vitro toxicological effect of the active 1,4-disubstitutedftalazines 2a–2h, we used a model based on the lysis of red blood cells (RBCs), measuring the hemoglobin released in the supernatant fraction.⁶⁸ The hemoglobin released was measured using a spectrophotometer at 550 nm. Lysis of RBCs in 50% Alsever’s solution were centrifuged at 800g for 10 min and then washed three times with saline solution to obtain RBCs. The synthesized compounds at three concentrations (0.1, 1, and 10 mM) were incubated with a 2% final suspension of RBCs at 37 °C for 45 min. The release of hemoglobin by an equal number of RBCs by hypotonic lysis in 0.05 vol of water was used as a 100% positive control, while RBCs treated with saline solution served as negative controls. Results were expressed as a hemolysis percentage (%) for each concentration of each tested compound.

Statistical Analysis

All experiments were performed at least three times. The results are expressed as mean ± SD. The ANOVA test was performed. Only post hoc Dunnett test p < 0.01 was considered to be statistically significant. The dose–response curves as well as the 50% growth inhibitory concentrations (IC₅₀ or CC₅₀) of the compounds were determined by a nonlinear regression of individual experiments calculated through computation with the GraphPad Prism v.5.02 software⁶⁷ program (Intuitive Software for Science, San Diego, CA, USA). The OriginPro2021b software package was employed for construction of graphic plots.⁶⁹

Animal Ethics

All experiments adhered to the French guidelines for animal research and were approved by the ethical committee of the Instituto de Higiene (Universidad Central de Venezuela, Caracas).

Theoretical Calculations

All DFT calculations were carried out at the Copernico Cluster of the computational center at Science Faculty (Universidad Central Venezuela, Caracas, Venezuela). All theoretical calculations, in the gas phase, were performed by using density functional theory (DFT) with the Gaussian 09 quantum chemistry software.⁶⁵ To achieve a balance between the computational cost and the accuracy of the calculations, the B3LYP functional developed by Becke⁶³ in conjunction with the 6-31G(d,p) basis set⁶⁴ was used in the present study. By comparison purpose, the calculations were focused in the most active molecules including the 1,4-disubstituted phthalazines (2a, 2b, 2d, 2e, 2f, and 2h), 1-substituted phthalazine-hydrazone (1d), 6-(N′-methylpiperazine)-3-(4-fluorophenyl)-1,2,4-triazolophthalazine (3c), N¹-aryl-2-trifluoromethyl-benzo[b]naphthyridin-4(1H)-ones (5a), and other non-active molecules such as 1-substituted phthalazine-hydrazone (1p), the 1,4-disubstituted phthalazine (2c), 6-chloro-3-phenyl-1,2,4-triazolophthalazine (3a), the 6-(N′-methylpiperazine)-3-(4-nitrofluorophenyl)-1,2,4-triazolophthalazine (3c), and the 2-(4-fluorophenyl)quinazolin-4(3H)-one (4b). The geometry of the ground state of all mentioned ligands was optimized. The complexation ability of their optimized geometries with one molecule of iron(II) or iron(III) was determined through an interactive model. To obtain the corrected interaction energy, the basis set superposition error (BSSE) approach was incorporated into the calculations via the counterpoise (CP) method.⁶⁶ For the ligand–iron(II) and ligand–iron(II) complexes, the complexation energy was included in the calculation process. Theoretical complexation energies can be found in Table 5. The calculation output can be found in the Supporting Information.

Experimental Complexation Determination in Solution

The complexation ability of a selected group of active ligands (1d, 2a, 2b, 2c, 2d, 2e, 2f, 2h, 3c, 3d, 4b, and 5k) with iron(III) was determined through a reported protocol with a few modifications by using UV–Vis spectroscopic measurements.⁷⁰⁻⁷² Titration with the iron(III) metal from 0 to 640 μM (10 concentrations) was performed at a constant ligand concentration: 1d (200 μM), 2a (140 μM), 2b (120 μM), 2c (200 μM), 2d (100 μM), 2e (200 μM), 2f (150 μM), 2g (100 μM), 2h (82 μM), 3b (100 μM), and 5k (50 μM). Maximum absorption for each ligand can be found in Table S6. The association constant (K_a) was obtained from eq 1 that is derived from the equilibrium between iodine and dye in eq 2 as follows:

1

2

Complexation stoichiometry was determined for each ligand according to a reported procedure.⁷¹ Results can be found in Figures S4–S7. A stoichiometry value of n = 1 was found for the ligands 1d, 3b, 3k, and CQ. The disubstituted phthalazin-hydrazones 2a, 2b, 2c, 2d, 2e, 2f, 2g, and 2h displayed a stoichiometry value of n = 2. Taking advantage of the discrete absorption of iron solution, association constants were determined from the (A_obs – A_ligand)/(A_ligand) variation as a function of iron concentration according to eq 1 and the stoichiometry of the complexation. CQ was used as a positive control, whereas the 2-arylquinazolin-4(3H)-one 4b was employed as a negative control. Some details can be found in the Supporting Information.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.1c05393.

• Full experimental details, further theoretical information, structure of 2-arylquinazolin-4(3H)-ones 4a–4t, in silico data details for active compounds, full hemolysis data, absorption spectra in the absence and presence of iron(III), and output for the optimized structures of ligands and those derived from corresponding complexes with iron(II) and iron(III) (PDF)

Author Present Address

^∥ Present address: Grupo de Química Orgaínica Medicinal, Instituto de Química Biológica, Facultad de Ciencias, Universidad de la República, Iguaí 4225, 11400 Montevideo, Uruguay (A.H.R.)

Author Contributions

M.E.A. performed the in vivo assay for the P. berghei model, G.C.H. performed the inhibitory activity toward β-hematin crystallization, N.G. supervised in vivo experiments, J.R.R. provided funding, supervised the β-hematin crystallization assay, and revised the final manuscript, L.G. performed all theoretical calculations, G.J.C. provided the financial resource for theoretical calculations, and A.H.R. organized the investigation, analyzed all experimental and theoretical data, and prepared and revised the manuscript.

The authors declare no competing financial interest.