3-Hydroxy-propanamidines, a New Class of Orally Active Antimalarials Targeting Plasmodium falciparum

Abstract

3-Hydroxypropanamidines are a new promising class of highly active antiplasmodial agents. The most active compound 22 exhibited excellent antiplasmodial in vitro activity with nanomolar inhibition of chloroquine-sensitive and multidrug-resistant parasite strains of Plasmodium falciparum (with IC₅₀ values of 5 and 12 nM against 3D7 and Dd2 strains, respectively) as well as low cytotoxicity in human cells. In addition, 22 showed strong in vivo activity in the Plasmodium berghei mouse model with a cure rate of 66% at 50 mg/kg and a cure rate of 33% at 30 mg/kg in the Peters test after once daily oral administration for 4 consecutive days. A quick onset of action was indicated by the fast drug absorption shown in mice. The new lead compound was also characterized by a high barrier to resistance and inhibited the heme detoxification machinery in P. falciparum.

Graphical Abstract

INTRODUCTION

Malaria, an infectious disease caused by protozoan parasites of the genus Plasmodium, affected an estimated 229 million people in 2019 and was responsible for more than 409,000 deaths globally.¹ Children aged under 5 years and pregnant women are the most vulnerable groups. Plasmodium falciparum the most prevalent species in Africa and Southeast Asia, is responsible for most fatalities worldwide.¹ In the absence of an effective vaccine, chemotherapy is of central importance for the treatment and control of malaria. Artemisinin-based combination therapies, which consist of a highly potent fast-acting artemisinin derivative coupled with a longer-acting partner drug, are the recommended first-line treatment for uncomplicated malaria in endemic regions.¹ However, P. falciparum resistance to artemisinin has spread across Southeast Asia, and resistance now affects all antimalarial drug classes in clinical use.² Although there has been significant success in malaria drug discovery in the past decade, there has not been a new class of antimalarials approved since the introduction of atovaquone in the 1990s. New drugs for malaria treatment and prophylaxis are urgently needed.

One commonly used strategy in antimalarial drug development is the structural optimization of established drugs to improve their antiplasmodial and physicochemical properties. Lumefantrine (LUM) belongs to the long-used drug class of arylamino alcohols and is commonly used as a partner drug with artemether, a derivative of artemisinin. LUM is currently undergoing Phase IIb clinical trials for the treatment of uncomplicated P. falciparum malaria in combination with the antimalarial drug candidate ganaplacide (KAF156).³ The shortcomings of LUM include its variable bioavailability due to fat-dependent absorption with high plasma protein binding and poor water solubility because of its high lipophilicity.^4,5 On the other hand, LUM is mostly devoid of the cardiac toxicity of halofantrine (HF), which is no longer in clinical use.⁶ Previously, we reported the development of 3-hydroxypropanehydrazonamide (1), a chemotype with potent antiplasmodial in vivo activity that has some structural resemblance to LUM and HF (Figure 1).⁷ Due to the potential formation of toxic hydrazine upon hydrazonamide hydrolysis, we have focused mainly on the replacement of the hydrazonamide motif by different alkylaryl- and aryl-substituted amidine (6–23), N-piperidinoamidine (24), and N-(4-fluoro)benzyloxyamidine (25) moieties (Figure 1). Herein, we describe the discovery of 3-hydroxypropanamidines (3-HPAs; 6–23) as a new class of compounds with potent antiplasmodial in vivo activity. The most active amidine 22 contained a para-methoxy-substituted benzamidine moiety instead of a hydrazonamide structure. In contrast, the substituted N-piperidinoamidine 24 and N-(4-fluoro)benzyloxyamidine 25 derivatives were less active.

Figure 1.

Structural comparison between the known antiplasmodial compounds halofantrine, lumefantrine, lead compound 1, 3-hydroxypropane-N-(arylalkyl)(aryl)amidines 6–23, 3-hydroxy-propane-N-piperidinoamidine 24, and 3-hydroxypropane-N-(4-fluoro)benzyloxyamidine 25.

RESULTS

Compound Design and Synthesis.

Based on the previously established structure–activity relationship (SAR) in the hydrazonamide series, 3-HPAs 6–23 were synthesized using two different linear synthetic routes (Scheme 1).^7–9 Different aromatic aldehydes were used as starting materials (2a–h). The aldehydes were either commercially available (2a, d, and e) or synthesized according to previously published methods (2b, c, and f–h),^9–11 followed by their conversion to 3-hydroxypropanenitriles according to Leven et al. (3a–h).⁷ 3-Hydroxypropanenitriles 3a–d were converted into imidoester hydrochlorides (4a–d) by Pinner reactions and the resulting imidoesters 4a–d were reacted directly with the respective arylalkyl amines and substituted anilines to yield amidine hydrochlorides (6–13, 22, and 23) as final compounds (route A).⁹

Scheme 1. Synthesis of 3-HPAs 6–23^a.

^a(i) ACN, n-BuLi, THF, −78 °C to rt, N₂, 1 h. (ii) MeOH, HCl in Et₂O, CH₂Cl₂/THF, −10 °C to rt. (iii) R¹-NH₂, DCM, 0 °C to rt, 12 h. (iv) 4-Fluorobenzylhydroxylamine, DCM, rt, 12 h. (v) 5 (4-Amino-2-((diethylamino)methyl)phenol) or R¹–NH₂, Al(CH₃)₃, toluene/THF, 0 to 50 °C, 5 h. (vi) 1-Aminopiperidine, Al(CH₃)₃, toluene/THF (1:1), 50 °C, 5 h.

Benzamidines 14–21 were synthesized based on the method of Korbad et al. (Scheme 1, route B) by reacting nitriles (3b, e–h) with the respective aryl amine (5, 4-fluoroaniline, 4-methoxyaniline, and 4-tert-butylaniline) in the presence of trimethylaluminum in toluene.¹² Although yields from route B were lower than those from route A, route B had the advantage that no storage-sensitive intermediates (imidoesters) were used and synthesis is less time-consuming. However, benzamidine 22 could only be synthesized in a low yield of 12% via synthesis route B (not shown), while route A afforded an improved yield of 39%. In alignment with the findings of Korbad et al., alkylarylamidines (6–13) could not be synthesized via route B. To yield the N-piperidinoamidine 24, nitrile 3b was directly reacted with 1-aminopiperidine in the presence of trimethylaluminum (see the Supporting Information). However, N-(4-fluoro)benzyloxyamidine 25 could not be prepared under these conditions. Therefore, imidate-HCl 4b was converted to 25 by its reaction with 4-fluorobenzylhydroxylamine (see the Supporting Information).

In this project, our SAR studies focused mostly on the replacement of the unfavorable hydrazonamide motif by differently substituted amidine moieties. Earlier studies demonstrated that a hydroxyl group at a defined distance to a basic nitrogen is essential for potent antimalarial activity.⁹ The novel 3-HPAs 6–23 were evaluated for their antiplasmodial in vitro activity against P. falciparum asexual blood stages (ABS) of the chloroquine (CQ)-sensitive strain Pf 3D7 and the multidrug-resistant strain PfDd2 [resistant to chloroquine, pyrimethamine, and sulfadoxine and with low-grade resistance to quinine (QN) as well as for their in vitro cytotoxicity against human cells (HepG2, HEK293, and HeLa). The hERG-potassium channel inhibition was tested as a potential off-target effect. Additionally, the most promising compound 22 was tested in vivo in the Plasmodium berghei mouse model with three different oral doses. Furthermore, the snapshot in vivo pharmacokinetic (PK) profile of 22 was determined in mice after oral administration. CQ, artesunate (AS), LUM, and HF were used as reference compounds.

Antiplasmodial In Vitro Evaluation of Amidines 6–23.

Except for the weakly active compound 10, N-arylalkyl-substituted 3-HPAs 6–9 and 11–13 exhibited significant antiplasmodial activity with IC₅₀ values mostly in the higher nanomolar range against the ABS of the CQ-sensitive strain Pf 3D7 (IC₅₀: 0.07–0.98 μM). Overall, the activity toward the multidrug-resistant line PfDd2 was substantially weaker and the parasite selectivity of series 1 (compounds 6–13) was less pronounced compared to lead compound 1 (Table 1).

Table 1.

In Vitro Evaluation of Amidines 6–13

Compd	X	Y	R2	Pf3D7a IC50±SD [μM]	PfDd2a IC50±SD [μM]	HepG2b IC50(n1; n2) [μM]	SIc (HepG/Pf3D7)
CQd				0.006 ± 0.002	0.23 ± 0.10	73.7 (74.6; 72.7)	12267
LUMe				0.006 ± 0.003	0.002 ± 0.0005	> 20 μM	> 20 μM
1f				0.002 ± 0.0005	0.0029 ± 0.0022	27.7	12043
6	H	H		0.98 ± 0.40	20.0 ± 0.4	9.2 (6.2; 12.2)	9
7	H	H		0.72 ± 0.10	10.0 ± 0.2	5.7 (5,0; 6,3)	8
8	H	H		0.48 ± 0.1	1.5 ± 0.6	13.0 (13.2; 12.8)	27
9	H	H		0.33b (0.14; 0.51)	1.1b (1.7; 0.5)	5.0 (4.2; 5.9)	16
10	H	H		15.0 ± 3.5	22.8 ± 4.5	50.5 (64.5; 36.7)	3
11	H	H		0.07 ± 0.01	0.7 ± 0.4	9.0 (8.9; 9.9)	129
12	F	CF3		0.24 ± 0.14	0.12 ± 0.001	n.d.	n.d.
13	Cl	CF3		0.35 ± 0.17	0.23 ± 0.11	7.3 (11.2; 3.4)	21

^aValues show the mean ± standard deviation (SD) of at least three independent experiments conducted in duplicate (11-points, serial dilution).

^bValues show the mean plus single values (n₁; n₂) of two independent experiments.

^cSelective Index (SI).

^dChloroquine.

^eLumefantrine.

^fPreviously published by Leven et al.;⁷ n.d., not determined.

According to the previously established SAR for the hydrazonamide chemotype, several benzamidine-based analogues with various aromatic and heteroaromatic substituents (R¹) were synthesized (14–23) and tested against P. falciparum ABS parasites (Table 2). The benzamidines depicted in Table 2 exerted equivalent potency against Pf 3D7 and Pf Dd2. The most active inhibitors of P. falciparum ABS growth contained a bulky tricyclic aromatic moiety (20–24), whereas derivatives with tert-butylphenyl, di-tert-butyl-phenyl, and quinoline moieties (14–18) exhibited activity only in the micromolar range. An exception was compound 16, with two tert-butylphenyl substituents (IC₅₀ = 0.05 μM; Pf Dd2 IC₅₀ = 0.05 μM). The most active 3-hydroxypropanamidine derivatives 22 and 23 showed potent antiplasmodial in vitro activity against P. falciparum ABS parasites up to single-digit nanomolar concentrations (22, 23: Pf 3D7 IC₅₀ = 0.005 μM; 22: Pf Dd2 IC₅₀ = 0.012 μM; 23: Pf Dd2 IC₅₀ = 0.009 μM). The 1,3-dichloro-6-(trifluoromethyl)phenanthrene-N-piperidinoamidine derivative 24 exhibited antiplasmodial activity in the single-digit nanomolar range against the P. falciparum 3D7 strain (Pf 3D7 IC₅₀ = 0.007 μM), whereas the respective N-(4-fluoro)benzyloxyamidine 25 was less active (Pf 3D7 IC₅₀ = 0.28 μM). To assess the parasite selectivity, the in vitro cytotoxicity of all compounds was initially tested against HepG2 cells, demonstrating low cytotoxicity and good to excellent parasite selectivity (Tables 1 and 2). Additionally, the cytotoxicity of the most active compounds 22 and 23 was evaluated on two additional human cell lines (HeLa and HEK293), confirming the lack of cytotoxic effects (Table 3).

Table 2.

In Vitro Evaluation of Amidines 14–23, N-Piperidinoamidine 24, and N-(4-Fluoro)benzyloxyamidine 25

Compd	R1	R2	Pf3D7a IC50±SD [μM]	PfDd2a IC50±SD [μM]	HepG2b IC50(n1; n2) [μM]	SIc (HepG/Pf3D7)
CQ			0.006 ± 0.002	0.23 ± 0.095	73.7 (74.6; 72.7)	12283
LUM			0.006 ± 0.003	0.003 ± 0.0005	>20	>20
1			0.002 ± 0.0005	0.003 ± 0.002	27.7	12043
14			1.5 ± 0.3	0.940 ± 0.509	57.5 (72.0; 43.0)	38
15			3.4 ± 0.8	2.4 ± 1.2	56.2 (72.8; 39.6)	17
16			0.05 ± 0.02	0.05 ± 0.01	11.8 (16.6; 7.0)	236
17-HCl			0.33 ± 0.08	0.35 ± 0.21	n.d.	n.d.
18-2HCl			2.1 ± 0.1	5.1 ± 0.9	179.8 (254.9; 104.6)	86
19			3.4 ± 1.1	5.6 ± 2.5	> 200	> 200
20			0.027 ± 0.009	0.04 ± 0.01	9.3 (8.6; 10.0)	344
21-2HCl			0.05 ± 0.03	0.64 ± 0.21	15.2 (13.1; 17.2)	304
22-HCl			0.005 ± 0.001	0.012 ± 0.006	4.3 (4.4; 4.2)	860
23-HCl			0.005 ± 0.002	0.009 ± 0.003	7.5 (10.7; 4.2)	1500
24-HCl			0.007 ± 0.002	0.014 ± 0.001	8.3 (13.1; 3.4)	1186
25			0.28 ± 0.1	0.29 ± 0.1	34.1 (57.7; 10.5)	122

^aValues show the mean ± standard deviation (SD) of at least three independent experiments conducted in duplicate (11-points, serial dilution).

^bValues show the mean ± standard deviation of two independent experiments.

^cSelective Index (SI).

Table 3.

Cytotoxicity of the Most Active Antiplasmodial Compounds on Different Cell Lines

compd	HepG2a IC50 (n1; n2) [μM]	SIb (HepG2/Pf3D7)	HeLaa IC50 (n1; n2)	SIb (HeLa/Pf3D7)	HEK239a IC50 (n1; n2) [μM]	SIb (HEK239/Pf3D7)
CQ	73.7 (74.6; 72.7)	12283	>1000	>1000	>1000	>1000
22	4.3 (4.4; 4.2)	860	46.1 (46.3; 45.8)	9220	36.8 (40.4; 33.2)	7360
23	7.5 (10.7; 4.2)	1500	127.2 (123.3; 131.0)	25,440	78.3 (79.3; 77.4)	15,660

^aValues are mean plus single values (n₁; n₂) of two independent experiments conducted in duplicate (11-points, serial dilution).

^bSelective Index.

Chemical Stability of Compound 22.

To study the stability of 22 at physiological pH, this compound was dissolved in a mixture of Tween20/ethanol/phosphate buffer pH 7.5 (7/3/90) and monitored over a period of 24 h. After this time, we detected nearly zero decomposition (1% drug decomposition, n = 2). The stability of 22 at acidic pH was determined by dissolving 22 in a mixture of Tween20/ethanol/phosphate buffer pH 2 (7/3/90) and monitoring over a period of 24 h. Again, after 24 h, nearly zero decomposition could be detected (0.1% drug decomposition, n = 2).

In Vitro Metabolic Stability of Compound 22.

Metabolic stability screening of 22 in human liver microsomes revealed 86% stability after 40 min of incubation. Propranolol, a reference drug with medium to high metabolic stability, showed that 76% of the parent compound remained and therefore demonstrated comparable stability to 22.¹³ Calculated results for the intrinsic clearance suggest that 22 (9 μL/min/mg) is a low-clearance compound with an estimated long half-life (n = 2).

In Vitro Plasma-Protein-Binding Study.

Human plasma protein binding was assessed for 22 as well as for the reference drug HF using a dialysis approach at pH 7.4. At a concentration of 1 μM, both compounds showed 100% plasma protein binding after 5 h of equilibration period. The obtained plasma protein binding of HF was consistent with reported values and subsequently confirmed the assay’s reliability.¹⁴

hERG Potassium Channel Inhibition.

A known liability of some arylamino alcohol antimalarials is the inhibition of the hERG potassium channel that can be associated with cardiotoxic effects. Thus, a fluorescence polarization assay was carried out to determine the potential hERG potassium channel inhibition of 22 in comparison with lead compound 1 and the references HF and LUM (Table 4). Compound 22 showed reduced hERG channel inhibition compared to lead compound 1. Interestingly, the hERG channel inhibition of LUM was also in the sub-micromolar range, although the drug does not show cardiotoxic effects in clinical use.

Table 4.

hERG Potassium Channel Inhibition of 1 and 22

compd	IC50 [μM]a
HFb	0.16
LUMc	0.38
1	0.07
22	0.19

^ahERG fluorescence polarization assay; IC₅₀ values were derived from an experiment conducted with 10-point titrations (each time point conducted in duplicate) with 3-fold serial dilutions.

^bHalofantrine (HF).

^cLumefantrine (LUM).

Inhibition of Hemozoin Formation.

The host hemoglobin breakdown by the parasite releases free ferrous heme that is rapidly oxidized to the insoluble ferric form and subsequently biomineralized into chemically inert hemozoin crystals, thus protecting the parasite from the oxidative cytotoxicity of free heme.¹⁵ The arylamino alcohol antimalarials LUM and MEF bind to heme, and the inhibition of heme detoxification is suspected to contribute to their mode of action. On this basis of structural similarity, we employed a pyridine-based heme fractionation protocol to assess whether inhibition of hemozoin formation contributes to the mode of action of lead compound 1 and amidine derivative 22. This experiment involved spectrophotometrically tracking the levels of free heme and hemozoin in synchronized early ring-stage parasites (~3 h post-invasion) that were incubated for 28 h at varying drug concentrations, whereupon the inhibitors of heme detoxification exhibit a dose-dependent signature of increasing “free” heme with a corresponding decrease in hemozoin levels.^16,17 Assays were performed using a range of drug concentrations, whereby 1 × IC₅₀ was defined as the inhibitory concentration of compounds against Pf NF54, as measured in a 72 h SYBR Green-based parasite susceptibility assay. Compound 1, 22, and CQ had IC₅₀ values of 5.8, 8.4, and 16.3 nM, respectively, against Pf NF54 in the 72 h assay. The measurements of heme and hemozoin quantify the parasite’s ability to detoxify reactive free heme through hemozoin formation in the presence of drug, where an increase in free heme corresponds to a decrease in parasite survival. The levels of hemoglobin, free heme, and hemozoin observed in untreated parasites indicate the amount of each species produced under physiological culture conditions. As expected, we observed an increase in the proportion (%) of “free” heme and a decrease in hemozoin levels upon treatment with increased concentrations of the positive control CQ (from 1.0 to 2.5 × IC₅₀; Figure 2). Treatment with compound 1 yielded no significant increase in “free” heme compared to untreated parasites at all concentrations. In contrast, parasites treated with 22 had a build-up in heme levels that was most pronounced at 2.5 × IC₅₀, with a significant decrease in hemozoin, a finding similar to that of CQ. However, 22 only showed very large amounts of free heme accumulation and a corresponding substantial decrease in hemozoin at 2.5 × IC₅₀, relative to untreated controls, whereas CQ showed a greater increase in free heme across multiple lower concentrations. Both 22 and CQ showed some inhibition of hemoglobin degradation at 1.5 to 2.5 × IC₅₀. These profiles provide evidence that 22 and CQ differ subtly in their degree of inhibition of heme detoxification. This distinction might reflect different interactions with heme molecules to form a complex or vacuolar accumulation dynamics resulting from pH trapping that could perhaps be more favorable for CQ or poor physicochemical features that limited the optimal permeation of 22 into the parasite’s digestive vacuole.

Figure 2.

Proportions of heme species observed as hemoglobin, free heme, or hemozoin in Pf NF54 trophozoites treated with increasing doses of compound 1 (black), 22 (red), or chloroquine (blue). Proportions are expressed as a percentage of total heme isolated from trophozoites after 28 h of incubation (that started with 3 to 4 h post-invasion ring-stage parasites). Data are presented as mean ± SEM (calculated from three independent biological repeats with technical duplicates). Statistical comparisons of the drug-treated parasites with their untreated controls were performed using two-tailed t-tests with Welch’s correction. *P < 0.05; **P < 0.01; ***P < 0.001.

In Vitro Resistance Development with 22 and Target Identification Studies.

To assess whether the parasite could develop resistance to lead compound 1 and amidine derivative 22, triplicate flasks of 2 × 10⁹ P. falciparum parasites of the multidrug-resistant Pf Dd2 clone B2 (Dd2-B2) were selected at 3× to 5× the compound IC₅₀ values. Upon recovering no viable parasites after 60 days, in vitro selections were repeated but with discrete increments of drug pressure and cycles of free-drug media. However, clonal lines displayed a substantial growth defect and quickly died.

In light of the difficulty in generating resistant parasites, we sought to use a hyper-mutating P. falciparum Dd2 line with a proofreading-deficient DNA polymerase δ (Pf_Polδ).^18,19 Both the Pf Dd2-B2 and Pf_Polδ lines exhibited similar IC₅₀ values in 72 h growth inhibition assays for both compound 1 (24.4 and 26.8 nM, respectively) and 22 (15.3 and 15 nM, respectively), but there was a significant reduction in the IC₅₀ values for the new amidine 22 (Figure 3A). Selections used triplicate flasks of 2 × 10⁹ Pf_Polδ parasites pressured with 3 to 5× the compound’s IC₅₀ values. Compound 1 yielded resistant parasites that showed 2- and 3-fold increase in IC₅₀ values and that were subsequently cloned by limiting dilution. Derivative 22 was unable to readily generate resistant parasites when Pf_Polδ parasites were exposed to 45 nM (3 × IC₅₀). Of note, the difficulty in selecting for resistance suggests that these mutations would not be easily favored by nature.

Figure 3.

In vitro resistance selection studies with Pf_Polδ parasites identify low-grade resistance mediated in part by pfmdr1. (A) Half-maximal inhibitory concentration of compounds 1 and 22 against multidrug-resistant parental lines. Error bars indicate the SEM based on three independently repeated experiments with technical duplicates. P values are based on parametric unpaired t tests. (B) Gain of resistance of compound 1-resistant clones and cross-resistance assays against compound 22. Error bars indicate the SEM based on two and three independently repeated experiments with technical duplicates for compounds 1 and 22, respectively. P values are based on parametric unpaired t tests. (C) qPCR analysis of resistant clones. Error bars indicate the SEM based on three independently repeated experiments with technical triplicates. (D) Potency of compounds against FCB lines with different pfmdr1 copy numbers. Error bars indicate the SEM based on two independent experiments with technical duplicates. P values are based on parametric unpaired t tests. *P < 0.05; **P < 0.01; ****P < 0.0001.

As a proxy for amidine 22 resistance, compound 1-resistant clones were tested against 22 in a standard 72 h drug-response assay. Clones Fl2-C4 and Fl3-C3 exhibited an average 1.6-fold shift in the IC₅₀ value (Figure 3B). These same clones demonstrated a 3-fold IC₅₀ increase when tested against compound 1. The whole-genome sequencing analysis of two resistant clones from two separate flasks, Fl1-D6 and Fl2-C4, with a high average depth of coverage (42–47×), identified a single-point mutation, S451F, in the multidrug-resistant protein 2 (also known as pfmdr2; PF3D7_1447900). This mutation was observed in both independently derived clones (Supplementary Tables S1 and S2). In addition, both clones revealed an amplification of a 81 kb segment on chromosome 5 that contained the multidrug-resistant protein 1 (pfmdr1; PF3D7_0523000; see Supplementary Table S3 Figure S1). These two clones gained three additional copies of pfmdr1 in addition to the two copies already present in the parental Pf_Polδ line. To interrogate the role of pfmdr1 amplification, additional clones (Fl1-G2 and Fl3-Ce) not analyzed by whole-genome sequencing were evaluated for copy number in a qPCR assay. Whole-genome sequenced clones and parental lines were included as controls. All clones showed between four to five copies of pfmdr1 (Figure 3B,C), confirming a gain of two to three–copies. We also tested compounds 1 and 22 against isogenic parasite lines known to have two copies of pfmdr1 (FCB WT) or one fully functional copy (FCB KD) with reduced pfmdr1 mRNA levels.²⁰ Although both lines were more sensitive than the parental line Dd2-B2, the decrease from two to one pfmdr1 copy caused a further sensitization (Figure 3D).

In Vivo Evaluation of Amidines 22 and 23.

On the basis of their subnanomolar antiplasmodial in vitro activity and excellent parasite selectivity, we selected compounds 22 and 23 for in vivo efficacy studies in P. berghei-infected mice. In a standard 4-day Peters test, mice (n = 3) were treated orally with 50, 30, 10, and 3 mg/kg bodyweight of 22 and 50 mg/kg bodyweight of 23 (Table 5; Figure 4). The 3-hydroxypropanamidine derivative 22 exhibited no obvious signs of toxicity during the in vivo studies. 22 exhibited excellent activity with partially curative in vivo efficacy at doses of 50 and 30 mg/kg (66 and 33% cure rates) along with mean survival time in days (MSDs) of 27 and 28 days, respectively. At 10 mg/kg, the in vivo activity was still high (99.9%, MSD 17 days), whereas no significant in vivo efficacy was observed at 3 mg/kg. At the highest dose of 50 mg/kg bodyweight 23 showed only a limited in vivo antimalarial activity of 44%.

Table 5.

In Vivo Evaluation of Lead Compounds 1 and 22, alongside Comparator Antimalarials and 23, Tested at Different Oral Doses

compd	dosea	activity [%]b	MSD [d]c	curedd
CQ	4 × 30	99.9	21	0/10
AS	4 × 30	99	9	0/10
LUM	4 × 30	99.8	>30	3/3
1	4 × 50	99.9	22	1/3
1	4 × 30	99.4	26	1/3
1	4 × 10	99.5	25	2/3
22	4 × 50	99.8	27	2/3
22	4 × 30	99.9	28	1/3
22	4 × 10	99.9	17	0/3
22	4 × 3	0	euthanized	0/3
23	4 × 50	~40	euthanized	0/3
control		0	4

^aStandard 4-day Peters test using daily oral doses of 3, 10, 30 and 50 mg/kg body weight. Experimental groups (n = 3 mice) were treated 4 × (4, 24, 48 and 72 h) post-infection.

^bBlood for parasitemia determination by FACS analysis was collected on day 4 (96 h after infection).

^cMean survival time in days (MSD). Control mice and mice where the antimalarial activity was <40% were euthanized on day 4 post-infection in order to prevent death otherwise occurring on day 6.²² The survival of the animals was monitored by microscopy up to 30 days. A compound was considered curative (>30), if the animal survived up to day 30 post-infection with no detectable parasites (detection limit: 1 parasite in 10,000 erythrocytes).

^dNumber of parasite-free mice on day 30.

Figure 4.

Example of dried blood spot PK concentration–time profiles after a single oral administration of 22 (10 mg/kg, n = 1).

Snapshot In Vivo PK of Amidine 22.

To investigate the drug concentration after a single oral administration in P. berghei-infected mice, a snapshot PK screening was performed using blood samples of dried blood spots, which were analyzed via liquid chromatography-tandem mass spectrometry (LC–MS/MS).

For compound 22, a tailored low-volume LC–MS/MS was developed and validated to measure PK parameters. After a single oral dose of 3, 10, and 30 mg/kg of 22, we collected venous blood 1, 4, and 24 h after drug administration. This study employed six mice, with two per administered dose. This snapshot in vivo PK analysis revealed fast absorption with a t_max of ~1 h (first sampling time point) in all six mice (Figure 4; Table 6). The mean elimination half-life (k_e) was estimated at ~6.2 h based on the three sampling time points, while the dose-dependent C_max concentrations (at 1 h post-dosing) averaged 3.5, 26, and 67.7 ng/mL for 3, 10, and 30 mg/kg dosing, respectively (Table 6).

Table 6.

PK Parameters after a Single Oral Dose of 3, 10, and 30 mg/kg of 22 Administered to P. berghei-Infected Mice

parameter		3 mg/kg		10 mg/kg		30 mg/kg
		mouse I	mouse II	mouse I	mouse II	mouse I	mouse II
ke	[1/h]	0.1328	0.1264	0.1677	0.1318	0.0697	0.0988
t1/2	[h]	5.21	5.49	4.13	5.26	9.95	7.02
Tmax	[h]	1	1	1	1	1	1
Cmax	[ng/mL]	3.7	3.3	29.1	23.0	78.1	57.3
AUC 0-ta	[ng/mL*h]	28.9	20.4	360.8	192.1	1083.2	783.1
AUC 0-inf_obsb	[ng/mL*h]	30.0	21.5	365.0	199.7	1313.6	848.1

^aAUC0-t: The area under the dried-blood-spot concentration–time curve from time zero to the last quantifiable observed concentration. AUC0-inf_obs: The area under the dried-blood-spot concentration–time curve from time zero to infinity. Results were obtained from non-compartmental analyses.

DISCUSSION

Our SAR study demonstrated that the replacement of the unfavorable hydrazonamide moiety of lead compound 1 by a benzamidine motif provided a novel chemotype with promising antimalarial in vivo activity after oral administration. These data confirm that a hydroxyl group at a defined distance to a basic functionality is essential for potent antimalarial activity. Our present SAR results revealed that the 1,3-dihalo-6-trifluoromethyl-phenanthrenyl moiety and the 4-methoxyphenyl substituent (22) are key structural elements for potent in vivo activity, excellent parasite selectivity, and high barrier to resistance in P. falciparum parasites. Furthermore, compound 22 demonstrated high chemical and metabolic stability. A liability of the novel lead compound 22 was its moderate inhibition of the hERG K⁺ channel in an in vitro fluorescence polarization assay. Although our preliminary data from initial mechanistic studies do not directly identify the molecular target, they indicate a possible contribution of inhibition of heme detoxification to the mode of action of compound 22. One future study to investigate the involvement of other determinants in the mode of action and/or mechanism of resistance of this class of compounds would involve in vitro parasite selection with LUM and other analogues of 22 using a hypermutator line with a genetic lineage likely to be permissive to LUM-induced genomic changes. In addition, the whole-genome sequencing data further suggest that pfmdr1 amplification provides a genetic background that favors low-grade resistance. Of note, the increased pfmdr1 copy number has previously been correlated with reduced in vitro parasite susceptibility to MEF, LUM, HF, QN, artesunate, and artemisinin.^21,23,24 This ABC transporter resides on the membrane of the digestive vacuole, wherein the endocytosed host hemoglobin is degraded into heme and hemozoin; here, globin is digested into peptides for parasite protein synthesis.^25,26

Our data also suggest a role of the S451F mutation in the related ABC transporter pfmdr2, which was observed in clones from two independent drug-pressured flasks. PfMDR2 is thought to localize onto the parasite plasma membrane and has previously been implicated in parasite susceptibility to several antimalarials including MEF and QN.^27–30 The strenuous efforts for selecting resistant parasites to either compound 1 or 22 and the poor growth observed with resistant clones indicate that the studied chemical scaffolds exhibit biological constraints for antimalarial drug resistance. Resistance was not achieved with Pf Dd2 parasites with parasite numbers up to 6 × 10⁹. By comparison, other selection studies conducted by our group over the same period with the same Dd2 line regularly obtained mutants resistant to KAE609, DSM265, or M5717 (also known as DDD107498) with parasite inocula of as low as 10⁷ parasites.^25,31–33 Indeed, we could only obtain resistance using the Pf_Polδ line that was recently shown to have an increased mutation rate and that at a starting inoculum of 10⁹ parasites enabled the selection of resistance to salinomycin A that had failed to select for resistance at inocula of up to 3 × 10¹⁰ Pf Dd2 parasites.^19,34 This feature positions these compounds well for future development in terms of showing a relatively low risk of resistance compared to several other advanced antimalarial drug candidates.

CONCLUSIONS

Herein, we describe the development of 3-HPAs as a novel class of orally active antimalarial agents, which are structurally related to arylamino alcohol antimalarials. Several benzamidine derivatives (16, 20–23) exhibited potent in vitro antiplasmodial activity against multidrug-resistant (Pf Dd2) and -sensitive (Pf 3D7) strains of P. falciparum along with good to excellent parasite selectivity. The novel lead compound 22 contains a 1,3-difluoro-6-trifluoromethyl-phenanthrenyl moiety and a 4-methoxyphenyl-substituted benzamidine functionality. 22 demonstrated potent activity with partially curative in vivo efficacy at doses of 50 and 30 mg/kg (66 and 33% cure rates) along with mean survival times of 27 and 28 days, respectively. A snapshot in vivo PK analysis in mice suggested good drug absorption and a fast onset of action after oral administration. Furthermore, both compounds 1 and 22 presented a high barrier to resistance in comparison with other experimental and clinical compounds. 22 also showed inhibition of the heme detoxification machinery, albeit not quite as efficient as CQ. The hERG K⁺ channel inhibition and the improvable drug-likeness (e.g., high fluorination status) of 22 will be addressed along with further SAR and additional mode of action studies in a follow-up lead optimization program. This new class of orally active antimalarials with high barrier to resistance is a promising starting point for further lead optimization of a preclinical candidate.

EXPERIMENTAL SECTION

General Procedures.

Proton (¹H) and carbon (¹³C) NMR spectra were recorded on a Bruker AVANCE 300 (300.13 MHz for ¹H and 75.47 MHz for ¹³C) and 600 MHz (600.22 MHz for ¹H und 151.93 MHz for 13C) NMR spectrometer using DMSO-d₆ or chloroform-d as a solvent. All spectra were recorded at room temperature in the specified solvents. The coupling constants between two nuclei over n bonds (ⁿJ) are given in Hertz (Hz) and chemical shifts are given in past per million (ppm). Elemental analysis was performed on a PerkinElmer PE 2400 CHN elemental analyzer. Mass spectroscopy was performed with a Bruker Daltonics UHR-QTOF maXis 4G or a Thermo Quest ion-trap-API-mass spectrometer Finnigan LCQ Deca. If necessary, the purity was determined by high-performance liquid chromatography (HPLC, method 1). Instrument: a Knauer HPLC system in combination with a Knauer UV Detector K-2600. Column: Vertex Plus (150 × 4 mm with a precolumn, Eurospher II 100–5 C18). Mobile phase 1: linear gradient (90–0%) of water with 0.1% trifluoroacetic acid. Mobile phase 2: linear gradient (10–100%) of acetonitrile with 0.1% of trifluoroacetic acid. Run time: 20 min, followed by an isocratic elution with 100% acetonitrile for 10 min. Flow rate: 1 mL/min. Detection: 254 nm. The purity of all final compounds determined by HPLC analysis was ≥95%. The bioanalytical separation and quantification were carried out on an Agilent 1200 Series HPLC system (method 2, Agilent, Waldbronn, Germany) coupled to a triple–quadrupole tandem mass spectrometer API 4000 (SCIEX, Concord, Canada) with an electrospray ionization (ESI) interface. The mass-to-charge ratio of 475.2 to 310.0 m/z for 22 was monitored utilizing multiple-reaction monitoring and a positive ionization mode. The validated method was characterized by a linear range from 0.025 to 100 ng/mL. The intra-run accuracy and inter-run accuracy varied from −4.5 to 13.2 and from −4.5 to 13.6% (relative error). The collected data were analyzed using Analyst 1.6.2 (Applied Biosystems/MDS SCIEX, Concord, Canada) with IntelliQuan as the integration algorithm without smoothing. Punching 6 mm samples out of the dried blood spots facilitated obtaining a uniform amount of sample during sample preparation. Punches were extracted using 600 μL of acetonitrile/water 80/20 (v/v) as the extraction solvent and an overhead vortexer with 99 rpm for 90 min. The supernatant was transferred into a deep well plate and was evaporated to dryness under a gentle stream of nitrogen while shaking at 550 rpm at 40 °C. The residue was reconstituted with 200 μL of a mixture of formic acid, dimethyl sulfoxide, acetonitrile, and water (0.1:1:98.9:10, v/v). Furthermore, a LC–MS/MS [Shimadzu LC coupled with API 4000 QTRAP (SCIEX, Concord, Canada)] was applied for plasma-protein-binding experiments as well as for metabolic stability tests. The collected data were analyzed using Analyst 1.6.3 (Applied Biosystems/MDS SCIEX, Concord, Canada). Microsomal incubation was done in five aliquots of 40 μL each. The liver microsomal incubation medium consisted of phosphate buffer (100 mM, pH 0.4), MgCl₂ (3.3 mM), NADPH (3 mM), glucose-6-phosphate (5.3 mM), and glucose-6-phosphate dehydrogenase (0.67 units/mL) with 0.42 mg of liver microsomal protein per mL. In the control reactions, the NADPH-cofactor system was substituted with phosphate buffer. Test compounds (2 μM; final solvent concentration, 1.6%) were incubated with microsomes at 37 °C, shaking at 100 rpm. The stability was observed over 40 min using diclofenac and propranolol as reference substances. The supernatant after protein precipitation by 90% acetonitrile–water was determined by LC–MS/MS. Using a Dialyzer HTD 96b (HTDialysis, Gales Ferry, USA), the plasma protein binding was determined. 120 μL of spiked plasma (1 μM of each compound) was added to one chamber and the same volume of pH 7.4 PBS buffer was added to the other chamber. After 5 h at 37 °C and 100 rpm, supernatants, following acetonitrile-triggered protein precipitation, were analyzed using a HPLC system coupled with a tandem mass spectrometer. Additionally, the recovery was monitored and HF and verapamil were used as reference compounds for plasma-protein-binding experiments.

Experimental Data.

Compounds 2a, 2d, and 2e were purchased from commercial suppliers in purities of 95–99%. Known compounds 2b, 2c, 2f–h, 3a–c, and 4a–c were synthesized according to the procedures of Leven, Totten, and Tagawa (Scheme 1).^7–11 The synthesis of starting materials 3d–h, 4d, and 5 and biological evaluation, cellular heme fractionation assay, in vitro selection of resistance, ¹H NMR and ¹³C NMR spectra, and HPLC profiles of the target compounds are reported in the Supporting Information.

General Procedure for the Synthesis of Final Compounds 6–13, 22, and 23.

To a suspension of imidomethyl ester hydrochloride (4a–d, 1.0 equiv) in anhyd. DCM (10 mL), the respective amine (1.0 equiv) was added under ice cooling. The reaction mixture was stirred for 12 h at rt, initially forming a clear solution from which a white solid precipitated over time. This precipitate was filtered off and recrystallized from methanol/diethyl ether.

N′-(4-Fluorobenzyl)-3-hydroxy-3-(phenanthrene-9-yl)-propanimidamide Hydrochloride (6).

White solid, yield: 86%, mp: 147 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 8.92 (m, 1H), 8.85 (m, 1H), 8.53 (d, J = 7.6 Hz, 1H), 8.05 (s, 1H), 7.99 (m, 1H), 7.79–7.62 (m, 4H), 7.40 (m, 2H), 7.16 (m, 2H), 5.86 (dd, J = 10.2, 3.9 Hz, 1H), 4.55 (m, 2H, NCH2Ar), 3.15 (dd, J = 13.9, 3.9 Hz, 1H), 2.88 (dd, J = 13.9, 9.9 Hz, 1H). ¹³C NMR (151 MHz, DMSO-d₆): δ [ppm] = 165.35, 162.08 (d, ¹J_CF = 243.5 Hz), 138.66, 132.28, 131.44, 130.54, 130.25 (d, ³J_CF = 7.7 Hz), 130.06, 129.48, 129.10, 127.47, 127.40, 127.35, 127.08, 125.29, 124.50, 123.91, 123.26, 115.68 (d, ²J_CF = 21.5 Hz), 67.51, 44.76, 41.61. Anal. calcd. for C₂₂H₂₄ClFN₂O, [%] C 70.50, H 5.42, N 6.85; found, [%] C 70.22, H 5.44, N 7.01.

3-Hydroxy-3-(phenanthrene-9-yl)-N′-phenethylimidamide Hydrochloride (7).

White solid, yield: 61%, mp: 167 °C, ¹H NMR (600 MHz, DMSO-d₆): δ [ppm] = 9.71 (br-s, 1H), 9.35 (br-s, 1H), 8.96–8.89 (m, 1H), 8.85 (d, J = 8.2 Hz, 1H), 8.49 (d, J = 7.9 Hz, 1H), 8.04 (d, J = 2.3 Hz, 1H), 8.01 (dd, J = 7.8, 1.5 Hz, 1H), 7.78–7.65 (m, 4H), 7.36–7.30 (m, 4H), 7.28–7.23 (m, 1H), 6.16 (s, 1H), 5.79 (d, J = 9.4 Hz, 1H), 3.51 (d, J = 8.2 Hz, 2H), 3.04 (dt, J = 13.8, 3.0 Hz, 1H), 2.91–2.78 (m, 3H). ¹³C NMR (151 MHz, DMSO-d₆): δ [ppm] = 165.07, 138.72, 131.42, 130.54, 130.04, 129.48, 129.43, 129.31, 129.08, 128.88, 127.50, 127.40, 127.34, 127.09, 126.96, 125.22, 124.41, 123.93, 123.27, 67.45, 43.77, 41.50, 33.79. Anal. calcd. for C₂₅H₂₅ClN₃O, [%] C 74.15, H 6.22, N 6.92; found, [%] C 73.94, H 6.27, N 6.92.

N′-Benzyl-3-hydroxy-3-(phenanthrene-9-yl)propanimidamide Hydrochloride (8).

White solid, yield: 72%, mp: 139 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 10.06 (br-s, 1H), 9.48 (br-s, 1H), 8.95–8.89 (m, 1H), 8.85 (dd, J = 8.2, 1.5 Hz, 1H), 8.57–8.51 (m, 1H), 8.06 (s, 1H), 8.04–7.97 (m, 1H), 7.79–7.62 (m, 4H), 7.34 (s, 5H), 6.23 (br-s, 1H), 5.86 (dd, J = 10.0, 3.8 Hz, 1H), 4.56 (d, J = 3.4 Hz, 2H), 3.16 (dd, J = 14.0, 3.9 Hz, 1H), 2.92 (d, J = 10.0 Hz, 1H). ¹³C NMR (75 MHz, DMSO-d₆): δ [ppm] = 164.95, 138.11, 135.40, 130.91, 130.04, 129.55, 128.95, 128.59, 128.47, 127.58, 127.51, 126.99, 126.91, 126.83, 126.59, 124.66, 123.99, 123.43, 122.76, 67.02, 45.04, 41.14. HPLC (Meth. 1): t_R = 11.0 min, AUC ≥99%.

N′-(4-Fluorophenethyl)-3-hydroxy-3-(phenanthrene-9-yl)-propanimidamide Hydrochloride (9).

White solid, yield: 86%, mp: 157 °C, ¹H NMR (600 MHz, DMSO-d₆): δ [ppm] = 9.60 (br-s, 1H), 9.28 (br-s, 1H), 8.93 (d, J = 8.0 Hz, 1H), 8.85 (d, J = 8.2 Hz, 1H), 8.50–8.41 (m, 1H), 8.05–7.98 (m, 2H), 7.78–7.64 (m, 4H), 7.36–7.31 (m, 2H), 7.15 (dd, J = 10.1, 7.5 Hz, 2H), 6.14 (s, 1H), 5.76 (s, 1H), 3.48 (s, 2H), 3.03 (dd, J = 14.1, 3.6 Hz, 1H), 2.89–2.77 (m, 2H). ¹³C NMR (151 MHz, DMSO-d₆): δ [ppm] = 165.15, 161.54 (d, ¹J_CF = 242.1 Hz), 138.59, 134.82, 131.40, 131.22(d, ³J_CF = 7.8 Hz), 130.55, 130.04, 129.43, 129.07, 127.52, 127.43, 127.34, 127.10, 125.10, 124.40, 123.96, 123.28, 115.54 (d, ²J_CF = 21.0 Hz), 67.44, 43.75, 41.55, 32.91. Anal. calcd. for C₂₅H₂₄ClFN₂O, [%] C 71.00, H 5.72, N 6.62; found, [%] C 70.74, H 5.73, N 6.61.

(Z)-3-Hydroxy-3-(phenanthrene-9-yl)-N′-(2-(pyridin-2-yl)ethyl)-propanimidamide Hydrochloride (10).

White solid, yield: 84%, mp: 166 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 8.96–8.88 (m, 1H), 8.88–8.80 (m, 1H), 8.53 (dd, J = 4.8, 1.6 Hz, 1H), 8.46 (d, J = 6.9 Hz, 1H), 8.03 (s, 1H), 8.02–7.98 (m, 1H), 7.80–7.62 (m, 6H), 7.36 (d, J = 7.9 Hz, 1H), 7.28 (dd, J = 7.6, 5.0 Hz, 1H), 5.78 (d, J = 8.7 Hz, 1H), 3.67 (t, J = 7.3 Hz, 2H), 3.10–2.97 (m, 3H), 2.81 (dd, J = 14.0, 10.1 Hz, 1H). ¹³C NMR (75 MHz, DMSO-d₆): δ [ppm] = 164.69, 157.79, 149.01, 138.11, 136.62, 130.90, 130.01, 129.51, 128.95, 128.56, 126.97, 126.88, 126.83, 126.56, 124.66, 123.88, 123.49, 123.41, 122.75, 121.86, 66.90, 41.37, 40.98, 35.22. Anal. calcd. for C₂₄H₂₄ClN₃O, [%] C 71.01, H 5.96, N 10.35; found, [%] C 70.65, H 5.98, N 10.21.

(Z)-3-Hydroxy-N′-(4-methoxyphenethyl)-3-(phenanthrene-9-yl)-propanimidamide Hydrochloride (11).

White solid, yield: 70%, mp: 101 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 8.96–8.89 (m, 1H), 8.84 (d, J = 7.8 Hz, 1H), 8.46 (d, J = 7.3 Hz, 1H), 8.06–7.97 (m, 2H), 7.81–7.59 (m, 4H), 7.21 (d, J = 8.4 Hz, 2H), 6.97–6.79 (m, 2H), 5.77 (d, J = 9.3 Hz, 1H), 3.73 (s, 3H), 3.49–3.41 (m, 2H), 3.10–2.98 (m, 1H), 2.89–2.72 (m, 3H). 13C NMR (75 MHz, DMSO-d₆): δ [ppm] = 164.50, 157.91, 138.12, 130.91, 130.03, 129.82, 129.60, 129.53, 128.96, 128.57, 126.98, 126.89, 126.83, 126.58, 124.65, 123.91, 123.43, 122.76, 113.76, 66.93, 54.99, 43.55, 40.99, 32.43. HPLC (meth. 1): t_R = 11.5 min, AUC = 95.1%.

(Z)-3-(1,3-Difluoro-6-(trifluoromethyl)phenanthrene-9-yl)-N′-(4-fluorophenethyl)-3-hydroxpropanimidamide Hydrochloride (12).

White solid, yield: 80%, mp: 122 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 9.74 (s, 1H), 9.40 (s, 1H, NH), 9.30 (s, 1H), 8.84 (t, J = 10.4 Hz, 2H), 8.32 (s, 1H), 8.05 (dd, J = 8.9, 1.8 Hz, 1H), 7.79–7.64 (m, 1H), 7.40–7.29 (m, 2H), 7.21–7.07 (m, 2H), 6.35 (s, 1H), 5.84 (d, J = 9.4 Hz, 1H), 3.48 (t, J = 7.5 Hz, 2H), 3.09–2.97 (m, 1H), 2.91–2.77 (m, 3H). ¹³C NMR (126 MHz, DMSO-d₆): δ [ppm] = 164.72, 161.44 (d, ¹J_CF = 242.4 Hz), 161.28 (d, ¹J_CF = 245.3 Hz), 159.23 (d, ¹J_CF = 251.8 Hz), 138.71, 134.67 (d, ⁴J_CF = 3.3 Hz), 132.10, 131.03 (d, ³J_CF = 7.7 Hz), 130.91 (d, 3 J = 7.8 Hz), 129.31, 127.84 (q, ²J_CF = 32.1 Hz), 127.11, 123.91, 123.66 (q, ¹J_CF = 271.6 Hz), 122.47, 118.16–117.75 (m), 115.63 (d, ²J_CF = 21.5 Hz), 115.40 (d, ²J_CF = 21.0 Hz), 105.78 (dd, ²J_CF = 23.1 Hz, ⁴J_CF = 3.6 Hz), 103.64 (dd, ²J_CF = 28.3, 24.9 Hz), 67.48, 43.63, 41.35, 32.80. HPLC (meth. 1): t_R = 13.4 min, AUC = 95.5%. m/z: calcd for [C₂₆H₂₁F₆N₂O]⁺,: 491.1553; found, 491.1551.

3-(1,3-Dichloro-6-(trifluoromethyl)phenanthren-9-yl)-N′-(4-fluorophenethyl)-3-hydroxy-propanimidamide Hydrochloride (13).

White solid, yield: 50%, mp: 198 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 9.41 (d, J = 1.8 Hz, 1H), 9.27 (d, J = 2.0 Hz, 1H), 8.85 (d, J = 8.8 Hz, 2H), 8.55 (s, 1H), 8.17–7.98 (m, 2H), 7.45–7.29 (m, 2H), 7.22–7.08 (m, 2H), 5.87 (d, J = 9.9 Hz, 1H), 3.49 (t, J = 7.5 Hz, 2H), 3.10–2.98 (m, 1H), 2.85 (dd, J = 8.8, 6.0 Hz, 2H), 2.76 (dd, J = 14.0, 10.0 Hz, 1H). ¹³C NMR (300 MHz, DMSO-d₆): δ [ppm] = 164.20, 161.03 (d, ¹J_CF = 242.1 Hz), 140.32, 134.27 (d, ⁴J_CF = 3.1 Hz), 132.68, 132.25, 131.80, 131.22, 130.68 (d, ³J_CF = 8.2 Hz), 128.89, 128.18, 127.71 (q, ²J_CF = 31.8 Hz), 126.95, 126.53, 124.34 (q, ¹J_CF = 272.8 Hz), 123.75–123.48 (m), 122.74, 122.36–122.06 (m), 121.09, 115.03 (d, ²J_CF = 21.0 Hz), 67.03, 43.25, 40.98, 32.39. Anal. calcd. for C₂₆H₂₁Cl₃F₄N₂O, [%] C 55.78, H 3.78, N 5.00; found, [%] C 55.51, H 4.00, N 5.01.

(Z)-3-(1,3-Difluoro-6-(trifluoromethyl)phenanthrene-9-yl)-3-hydroxy-N′-(4-methoxyphenyl)propanimidamide Hydrochloride (22).

White solid, yield: 39%, mp: 179 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 11.30 (br-s, 1H), 9.60 (br-s, 1H), 9.31 (br-s, 1H), 8.88 (t, J = 8.7 Hz, 2H), 8.53 (s, 1H), 8.39 (s, 1H), 8.06 (d, J = 8.9 Hz, 1H), 7.82–7.67 (m, 1H), 7.28–7.18 (m, 2H), 7.17–7.05 (m, 2H), 6.50 (br-s, 1H), 5.97 (d, J = 9.0 Hz, 1H), 3.82 (s, 3H), 3.26–3.16 (m, 1H), 3.02–2.84 (m, 1H). ¹³C NMR (75 MHz, DMSO-d₆): δ [ppm] = 164.71, 160.81 (dd, J_CF = 245.2, 14.3 Hz), 158.86 (dd, J_CF = 251.8, 13.8 Hz), 158.84, 138.45, 131.85 (d, ⁴J_CF = 4.7 Hz), 131.66, 129.06–128.70 (m), 127.41 (d, ²J_CF = 31.8 Hz), 126.83, 126.78, 123.80 (q, ¹J_CF = 272.6 Hz), 123.59–123.45 (m), 122.09, 117.70, 117.51, 117.43, 115.12, 105.42 (dd, J_CF = 20.2, 3.1 Hz), 103.28 (dd, ²J_CF = 29.0, 24.8 Hz), 67.17, 55.44, 41.04. HPLC: t_R = 12.8 min, AUC = 99.1%. Anal. calcd. for C₂₅H₂₀ClF₅N₂O₂, [%] C 58.77, H 3.95, N 5.48; Found, [%] C 58.36, H 3.79, N 5.42.

(Z)-3-(Anthracene-9-yl)-N′-(4-fluorophenyl)-3-hydroxypropanimidamide Hydrochloride (23).

White solid, yield: 52%, mp: 188 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 11.55 (br-s, 1H), 9.84 (br-s, 1H), 8.97 (s, 2H), 8.68 (s, 1H), 8.62 (s, 1H), 8.16–8.08 (m, 2H), 7.62–7.51 (m, 4H), 7.48–7.34 (m, 4H), 6.62 (d, J = 10.3 Hz, 1H), 6.48 (s, 1H), 3.62–3.46 (m, 1H), 3.19–3.06 (m, 1H). ¹³C NMR (75 MHz, DMSO-d₆): δ [ppm] = 165.19, 161.40 (d, ¹J_CF = 244.9 Hz), 134.06, 131.14, 130.71 (d, ⁴J_CF = 2.8 Hz), 128.89, 128.71, 128.17, 128.05, 127.92, 125.52, 124.94, 116.85 (d, ²J_CF = 22.9 Hz), 66.68. HPLC (meth. 1): t_R = 10.7 min, AUC ≥99%.

General Procedure for the Synthesis of Compounds 14–21.

Under a nitrogen atmosphere, the respective amine (2.0 equiv) was dissolved in anhyd. toluene (5 mL) at 0 °C. Trimethylaluminum (2.0 equiv, 2 M in toluene) was subsequently added and the solution was stirred at 0 °C for 5 min. Then, a solution of the corresponding nitrile (3d–f, 1.0 equiv) in dry THF (5 mL) was added dropwise to the reaction mixture and stirred at 50 °C for 5 h. After cooling to room temperature, the reaction mixture was poured onto an ice–water mixture (10 mL). Subsequently, the phases were separated. The aqueous phase was adjusted to pH 8 with a sat. NaHCO₃ solution and then extracted with AcOEt (4 × 50 mL); the combined organic phases were washed with sat. sodium chloride solution and dried over anhyd. Na₂SO₄. After filtration, the solvent was removed under reduced pressure and the oily residue was purified by flash chromatography on silica gel (dichloromethane/methanol (0 to 30%). If necessary, the product was recrystallized from methanol/diethyl ether.

(Z)-3-(4-(tert-Butyl)phenyl)-N′-(4-fluorophenyl)-3-hydroxypropanimidamide (14).

White solid, yield: 68%, mp: 122 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 7.39–7.28 (m, 4H), 7.10–7.02 (m, 1.7 H_major), 6.91 (t, J = 8.7 Hz, 0.3 H_minor), 6.75 (s, 1.7H_major), 6.43 (s, 0.3 H_minor), 6.08 (d, J = 43.0 Hz, 2H), 4.99 (dd, J = 7.6, 5.4 Hz, 0.8 H_major), 4.84 (s, 0.2 H_minor), 2.48–2.44 (m, 2 H), 1.29 (s, 9H). ¹³C NMR (75 MHz, DMSO-d₆): δ [ppm] = 157.85, 157.68 (d, ¹J_CF = 236.9 Hz), 149.08, 145.82, 142.05, 125.56, 124.62, 123.09 (d, ³J_CF = 8.1 Hz), 115.50 (d, ²J_CF = 21.8 Hz), 70.31, 44.11, 34.13, 31.19. HPLC (meth. 1): t_R = 10.0 min, AUC ≥99%.

(Z)-3-(4-(tert-Butyl)phenyl)-3-hydroxy-N′-(4-methoxyphenyl)-propanimidamide (15).

White solid, yield: 36%, mp: 123 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 7.35 (t, J = 5.9 Hz, 4H), 6.85 (d, J = 8.2 Hz, 2H), 6.74 (s, 2H), 6.13 (br-s, 3H), 4.97 (t, J = 6.4 Hz, 1H), 3.71 (s, 3H), 2.46 (d, J = 7.2 Hz, 2H), 1.28 (s, 9H). ¹³C NMR (151 MHz, DMSO-d₆): δ [ppm] = 158.35, 154.98, 149.58, 142.54, 126.01, 125.12, 123.17, 114.90, 70.78, 55.56, 44.29, 34.63, 31.70. HPLC (meth. 1): t_R = 10.3 min, AUC = 96.3%.

(Z)-N′−3-Bis-(4-(tert-butyl)phenyl)-3-hydroxypropanimidamide (16).

White solid, yield: 68%, mp: 110 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 7.34 (d, J = 2.3 Hz, 4H), 7.28 (d, J = 8.3 Hz, 2H), 6.74 (s, 2H), 6.05 (br-s, 3H), 4.98 (t, J = 6.4 Hz, 1H), 2.47 (d, J = 6.7 Hz, 2H), 1.29 (s, 9H), 1.27 (s, 9H). ¹³C NMR (151 MHz DMSO-d₆): δ [ppm] = 157.91, 149.58, 146.90, 144.36, 142.54, 126.21, 126.06, 125.12, 121.87, 70.84, 44.39, 34.63, 34.34, 31.85, 31.70. HPLC (meth. 1): t_R = 12.73 min, AUC = 94.8%.

(Z)-3-(3,5-Di-tert-butylphenyl)-N′-(4-fluorophenyl)-3-hydroxypropanimidamide Hydrochloride (17).

The formation of the hydrochloride of the free base of 17 was carried out by dissolving the solid in methanol (5 mL) and adding HCl (1.5 equiv, 6N in isopropanol) to the solution. After stirring for 5 min, 17 was precipitated by the slow addition of diethyl ether. After filtration, the product was recrystallized from methanol/diethyl ether to yield 17. White solid, yield: 50%, mp: 122 °C, ¹H NMR (600 MHz, DMSO-d₆): δ [ppm] = 11.23 (br-s, 1H), 9.50 (br-s, 1H), 8.60 (br-s, 1H), 7.42–7.34 (m, 3H), 7.32–7.25 (m, 4H), 5.96 (s, 1H), 5.11 (s, 1H), 2.90 (d, J = 12.8 Hz, 1H), 2.80 (d, J = 12.0 Hz, 1H), 1.31 (s, 18H). ¹³C NMR (151 MHz, DMSO-d₆): δ [ppm] = 165.67, 161.67 (d, ¹J_CF = 245.8 Hz), 150.60, 143.67 131.66, 128.05 (d, ³J_CF = 8.4 Hz), 121.48, 120.28, 117.29 (d, ²J_CF = 22.9 Hz), 71.29, 43.21, 35.05, 31.80. HPLC (meth. 1): t_R = 12.7 min, AUC = 96.4%. m/z calcd for [C₂₃H₃₂FN₂O]⁺,: 371.2493; found, 371.2499.

(Z)-N′-(4-Fluorophenyl)-3-hydroxy-3-(6-methoxyquinoline-4-yl)-propanimidamide Dihydrochloride (18).

Conversion of the free base of 18 was carried out by dissolving the solid in methanol (5 mL) and adding HCl (2.5 equiv, 6N in isopropanol) to the solution. After stirring for 5 min, 18 was precipitated by the slow addition of diethyl ether. After filtration, the product was recrystallized from methanol/diethyl ether to yield 18. Yellow solid, yield: 36%, mp: 189 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 12.04 (s, 1H), 9.96 (s, 1H), 9.13 (d, J = 5.5 Hz, 1H), 8.72 (s, 1H), 8.38 (dd, J = 9.3, 2.0 Hz, 1H), 8.17 (d, J = 5.4 Hz, 1H), 8.09 (d, J = 2.6 Hz, 1H), 7.81 (dd, J = 9.3, 2.5 Hz, 1H), 7.49–7.27 (m, 4H), 6.33 (dd, J = 9.7, 4.2 Hz, 1H), 4.11 (s, 3H), 3.35 (dd, J = 13.9, 4.2 Hz, 1H), 2.96 (dd, J = 13.8, 9.9 Hz, 1H). ¹³C NMR (75 MHz, DMSO-d₆): δ [ppm] = 164.12, 161.31 (d, ¹J_CF = 244.9 Hz), 159.37, 157.95, 141.71, 134.41, 130.54 (d, ⁴J_CF = 3.0 Hz), 127.80 (d, ³J_CF = 8.9 Hz), 127.18, 126.37, 123.63, 119.39, 116.81 (d, ²J_CF = 22.9 Hz), 103.79, 65.95, 57.06. HPLC (meth. 1): t_R = 5.2 min, AUC = 95.6%. m/z calcd for [C₁₉H₁₈FN₃O₂]⁺,: 340.1456; found, 340.1457.

(Z)-N′-(4-Fluorophenyl)-3-hydroxy-3-(quinoline-4-yl)-propanimidamide (19).

White solid, yield: 57%, mp: 135 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 8.92 (d, J = 4.4 Hz, 1H), 8.29 (dd, J = 8.4, 1.4 Hz, 1H), 8.06 (dd, J = 8.4, 1.3 Hz, 1H, Harom), 7.77 (dd, J = 8.4, 1.4 Hz, 1H), 7.71–7.60 (m, 2H), 7.11–7.01 (m, 2H), 6.81–6.62 (m, 2H), 6.31 (d, J = 22.3 Hz, 1H), 6.09 (br-s, 2H), 5.90–5.79 (m, 1H), 2.72 (dd, J = 14.7, 4.3 Hz, 1H), 2.58 (dd, J = 14.7, 8.2 Hz, 1H). ¹³C NMR (151 MHz, DMSO-d₆): δ [ppm] = 158.23 (d, ¹J_CF = 237.3 Hz), 157.67, 150.81, 150.73, 148.22, 130.19, 129.43, 126.86, 125.71, 124.34, 123.66, 123.64, 118.54, 116.03 (d, ²J_CF = 21.9 Hz), 67.28, 43.64. Anal. calcd. for C₁₈H₁₆FN₃O, [%] C 69.89, H 5.21, N 13.58; found, C 69.60, H 5.50, 13.49.

(Z)-N′-(4-Fluorophenyl)-3-hydroxy-3-(phenanthren-9-yl)-propanimidamide (20).

White solid, yield: 28%, mp: 113 °C, ¹H NMR (300 MHz, Chloroform-d): δ [ppm] = 8.77 (d, J = 8.7 Hz, 1H), 8.67 (d, J = 8.1 Hz, 1H), 8.09 (s, 1H), 7.92 (d, J = 7.3 Hz, 1H), 7.76–7.55 (m, 4H), 7.08–6.85 (m, 4H), 5.99 (d, J = 6.7 Hz, 1H), 3.03–2.70 (m, 2H). ¹³C NMR (151 MHz, Chloroform-d): δ [ppm] = 159.29 (d, ¹J_CF = 241.8 Hz), 158.19, 143.56, 136.74, 131.51, 130.70, 129.95, 129.27, 128.85, 126.80, 126.65 (d, ³J_CF = 6.6 Hz), 126.21, 124.19, 123.46, 123.39, 123.32, 122.41, 116.21 (d, ²J_CF = 22.1 Hz), 67.95, 41.62. HPLC (meth. 1): t_R = 10.4 min, AUC = 97.5%.

(Z)-N-(3-((Diethylamino)methyl)-4-hydroxyphenyl)-3-hydroxy-3-(phenanthrene-9-yl)-propanimidamide Dihydrochloride (21).

Compound 21 was isolated as a dihydrochloride. To a solution of the free base 21 (441.6 mg, 1.0 mmol) in methanol, HCl [0.42 mL, 2.5 mmol (6N in isopropanol)] was added. After stirring for 5 min, the solvent was evaporated in vacuum. The remaining sticky, solid residue was dissolved in distilled water (5 mL), followed by lyophilization of the aqueous solution to yield 21. Brown solid, yield: 22%, mp: 31 °C, ¹H NMR (600 MHz, DMSO-d₆): δ [ppm] = 11.45 (s, 1H), 10.91 (s, 1H), 10.37 (s, 1H), 9.79 (s, 1H), 8.92 (dd, J = 8.0, 1.6 Hz, 1H), 8.89–8.81 (m, 2H), 8.62 (dd, J = 8.0, 1.6 Hz, 1H), 8.12 (s, 1H), 8.03 (dd, J = 7.7, 1.5 Hz, 1H), 7.78–7.66 (m, 4H), 7.53 (d, J = 2.5 Hz, 1H), 7.22–7.16 (m, 2H), 6.29 (d, J = 4.6 Hz, 1H), 5.99–5.91 (m, 1H), 4.23 (s, 2H), 3.24 (dd, J = 13.8, 3.9 Hz, 1H), 3.16–3.07 (m, 4H), 3.00 (dd, J = 13.8, 10.2 Hz, 1H), 1.30 (t, J = 7.2 Hz, 6H). ¹³C NMR (151 MHz, DMSO-d₆): δ [ppm] = 164.80, 156.89, 138.65, 131.42, 130.54, 130.13, 130.05, 129.44, 129.11, 128.32, 127.55, 127.48, 127.38, 127.15, 125.91, 125.18, 124.51, 123.96, 123.28, 117.97, 117.46, 67.66, 49.85, 46.43, 45.86, 41.74, 8.91, 8.83. HPLC (meth. 1): t_R = 8.0 min, AUC = 97.4%.

(Z)-3-(1,3-Dichloro-6-(trifluoromethyl)phenanthrene-9-yl)-3-hydroxy-N′-(piperidin-1-yl)-propanimidamide Hydrochloride (24).

Under a nitrogen atmosphere, 1-aminopiperidine (0.216 mL, 2 mmol) was dissolved in dry toluene (5 mL) at 0 °C. Then, trimethylaluminum (2.00 mmol, 2 M in toluene) was added. The solution was stirred for 5 min at 0 °C. Subsequently, 3b (384.2 mg, 1.00 mmol) dissolved in dry THF (5 mL) was added dropwise. The reaction mixture was heated to 50 °C and stirred for 5 h. After cooling to rt, the mixture was poured onto an ice–water mixture (10 mL). After phase separation, the aqueous phase was adjusted to pH 8 with a sat. NaHCO₃ solution and then extracted with AcOEt (4 × 50 mL). The combined organic phases were washed with sat. sodium chloride solution and dried over anhydrous Na₂SO₄. After filtration, the solvent was removed under reduced pressure and the oily residue was purified by flash chromatography [dichloromethane/methanol (10%)]. The conversion of the free base of 24 into the hydrochloride salt was carried out by dissolving the oily residue in methanol (5 mL) and adding HCl (1.5 equiv, 6N in isopropanol) to the solution. After stirring for 5 min, 24 was precipitated by the slow addition of diethyl ether. After filtration, the crude product was recrystallized from methanol/diethyl ether to yield 24. White solid, yield: 38%, mp: 210 °C, ¹H NMR (300 MHz, DMSO-d₆): δ [ppm] = 9.39 (s, 1H), 9.25 (d, J = 1.9 Hz, 1H), 8.80 (d, J = 8.8 Hz, 1H), 8.52 (s, 1H), 8.06 (dd, J = 10.1, 1.7 Hz, 2H), 5.85 (dd, J = 9.5, 3.8 Hz, 1H), 3.02 (dd, J = 14.0, 3.8 Hz, 1H), 2.79–2.57 (m, 4H), 1.69–1.57 (m, 4H), 1.40 (s, 2H). ¹³C NMR (75 MHz, DMSO-d₆): δ [ppm] = 162.18, 140.19, 132.67, 132.23, 131.79, 131.23, 128.88, 128.16, 127.70 (q, ²J_CF = 32.4 Hz), 126.91, 126.41, 124.34 (q, ¹J_CF = 272.4 Hz), 123.82–123.52 (m), 122.71, 122.36–122.05 (m), 121.07, 66.85, 54.84, 38.50, 24.69, 22.65. HPLC (meth. 1): t_R = 14.2 min, AUC = 97.9%.

(Z)-3-(1,3-Dichloro-6-(trifluoromethyl)phenanthrene-9-yl)-N′-((4-fluorobenzyl)oxy)-3-hydroxypropanimidamide (25).

The imidomethyl ester hydrochloride 4b (452.7 mg, 1.00 mmol) was suspended in DCM (10 mL), followed by dropwise addition of 4-fluorobenzylhydroxylamine (141.1 mg, 1.00 mmol) in DCM (5 mL). The reaction mixture was stirred for 12 h at rt. Afterward, the solvent was removed under reduced pressure. The solid raw product was recrystallized from diethyl ether to yield 25 as a white solid. White solid, yield: 50%, mp: 189 °C, ¹H NMR (600 MHz, DMSO-d₆): δ [ppm] = 9.40 (s, 1H), 9.27 (s, 1H), 8.80 (br-s, 1H), 8.53 (s, 1H), 8.08 (d, J = 1.8 Hz, 1H), 8.05 (d, J = 8.5 Hz, 1H), 7.58–7.46 (m, 2H), 7.31 (br-s, 1H), 7.21 (t, J = 8.7 Hz, 2H), 7.14 (br-s, 1H) 6.41 (br-s, 1H), 5.83 (d, ³J = 8.2 Hz, 1H), 4.93 (s, 2H), 3.02–2.94 (m, 1H), 2.64–2.56 (m, 1H). ¹³C NMR (75 MHz, DMSO-d₆): δ [ppm] = 162.36 (d, ¹J_CF = 245.0 Hz), 159.84, 140.24, 132.69, 132.24, 131.77 (d, ⁴J_CF = 4.1 Hz), 131.17, 130.80, 128.86, 128.16, 127.70 (d, ²J_CF = 32.1 Hz), 126.93, 126.54, 124.33 (q, ¹J_CF = 272.7 Hz), 123.69, 122.71, 122.19, 121.00, 115.63, 115.17 (d, ²J_CF = 21.5 Hz), 76.29, 66.70, 37.47. HPLC (meth. 1): t_R = 20.0 min, AUC = 96.3% (method run time: 25 min).

ABBREVIATIONS

abs

absolute

ABS

asexual blood stages

ACN

acetonitrile

AcOEt

ethyl acetate

AS

artesunate

br

broad

BW

bodyweight

CQ

chloroquine

d

doublet

DCM

dichloro-methane

dd

doublet of doublet

dt

doublet of triplet

hERG

human ether-à-go-go-related gene

HF

halofantrine

3-HPAs

3-hydroxypropanamidine

HPLC

high-pressure liquid chromatography

HRMS

high-resolution mass spectroscopy

Hz

Hertz

LUM

lumefantrine

m

multiplet

MEF

mefloquine

MeOH

methanol

MSD

mean survival time

n-Buli

n-Butyl lithium

PK

pharmacokinetic

QN

quinine

rt

room temperature

s

singlet

SAR

structure–activity relationship

q

quartet

t

triplet

THF

tetrahydrofuran