Novel 2-Phenoxyanilide Congeners Derived from a Hit Structure of the TCAMS: Synthesis and Evaluation of Their in Vitro Activity against Plasmodium falciparum

The Tres Cantos Antimalarial Compound Set (TCAMS) is a publicly available compound library which contains 13533 hit structures with confirmed activity against Plasmodium falciparum, the infective agent responsible for malaria tropica. The TCAMS provides a variety of starting points for the investigation of new antiplasmodial drug leads. One of the promising compounds is TCMDC-137332, which seemed to be a good starting point due to its antiplasmodial potency and its predicted physicochemical properties. Several new analogues based on a 2-phenoxyanilide scaffold were synthesized by standard amide coupling reactions and were fully characterized regarding their identity and purity by spectroscopic and chromatographic methods. Furthermore, the results of the biological evaluation of all congeners against Plasmodium falciparum NF54 strains are presented. The findings of our in vitro screening could not confirm the presumed nanomolar antiplasmodial activity of TCMDC-137332 and its derivatives.


Introduction
Malaria is still one of the most severe infectious diseases. Approximately 3.2 billion people are at risk of being infected. Despite a decreasing number of mortal cases in the last decade due to better vector control and artemisinine-based combination therapy (ACT), still 584,000 deaths caused by malaria infection were reported in 2013 [1]. Although there are effective chemotherapeutics for the treatment of malaria, the discovery of new drugs is important due to the increasing resistances against available drugs [2,3]. A good strategy to circumvent the problem of resistance is to develop compounds acting with new mechanisms of action. Hence there is an urgent need for compounds with novel chemotypes which differ from scaffolds of existing drugs.
Major sources for drug discovery are phenotypic screening libraries, providing a huge number of new starting points. In the last few years extensive libraries with compounds showing antimalarial activity have been published by Novartis [4], St. Jude's Children Research Hospital [5], the Medical Research Council Technology [6] and GlaxoSmithKline (GSK) [7]. The latter was derived from a high throughput screening (HTS) of nearly 2 million substances of the GSK corporate collection. 13,533 of these, known as the Tres Cantos Antimalarial Compound Set (TCAMS), were active against the malaria parasite Plasmodium falciparum and inhibited the parasite growth by at least 80% at 2 µM concentration. In silico clustering and filtering of the TCAMS set performed by Calderon et al. resulted in 552 compounds which were declared as "quality starting points from the TCAMS" [8]. Since then, a few of the compound classes have been further examined, e.g., cyclopropylcarboxamides [9], 2-amino-1-phenylethanols [10], aminohydantoines [11] or carbamoyltriazoles [12].
One of the 552 promising compounds is TCMDC-137332 (1), which has an estimated IC50 value of 7 nM and therefore seemed to be one of the most potent structures of this collection [7,8]. We decided to further investigate this compound which matches nearly all criteria of the Lipinski rules for orally available drugs [13] (Figure 1) and is structurally dissimilar to currently known antimalarials [14].

Chemistry
To proof its activity, we resynthesized TCMDC-137332. We also prepared a series of congeners, all having a 2-phenoxyanilide scaffold ( Figure 2). The first group of ten compounds was substituted with chlorine in para position of the phenoxy residue, like in the scaffold of TCMDC-137332 (1-10). Another equivalent series was synthesized, in which the chlorine was replaced by a more polar methoxy group (11)(12)(13)(14)(15)(16)(17)(18)(19)(20). Major sources for drug discovery are phenotypic screening libraries, providing a huge number of new starting points. In the last few years extensive libraries with compounds showing antimalarial activity have been published by Novartis [4], St. Jude's Children Research Hospital [5], the Medical Research Council Technology [6] and GlaxoSmithKline (GSK) [7]. The latter was derived from a high throughput screening (HTS) of nearly 2 million substances of the GSK corporate collection. 13,533 of these, known as the Tres Cantos Antimalarial Compound Set (TCAMS), were active against the malaria parasite Plasmodium falciparum and inhibited the parasite growth by at least 80% at 2 µM concentration. In silico clustering and filtering of the TCAMS set performed by Calderon et al. resulted in 552 compounds which were declared as "quality starting points from the TCAMS" [8]. Since then, a few of the compound classes have been further examined, e.g., cyclopropylcarboxamides [9], 2-amino-1-phenylethanols [10], aminohydantoines [11] or carbamoyltriazoles [12].
One of the 552 promising compounds is TCMDC-137332 (1), which has an estimated IC50 value of 7 nM and therefore seemed to be one of the most potent structures of this collection [7,8]. We decided to further investigate this compound which matches nearly all criteria of the Lipinski rules for orally available drugs [13] (Figure 1) and is structurally dissimilar to currently known antimalarials [14].

Chemistry
To proof its activity, we resynthesized TCMDC-137332. We also prepared a series of congeners, all having a 2-phenoxyanilide scaffold ( Figure 2). The first group of ten compounds was substituted with chlorine in para position of the phenoxy residue, like in the scaffold of TCMDC-137332 (1-10). Another equivalent series was synthesized, in which the chlorine was replaced by a more polar methoxy group (11)(12)(13)(14)(15)(16)(17)(18)(19)(20).   The synthesis of phenoxyanilides 1-8 and 11-18 was achieved by standard amide coupling reactions of commercially available 2-(4-chlorophenoxy)aniline I or 2-(4-methoxyphenoxy)aniline hydrochloride II with acid chlorides in presence of triethylamine, as outlined in Scheme 1A. Compounds 9 and 19 were synthesized by the reaction of 2-(4-substituted)phenoxyanilines and Boc-protected glycine in the presence of PyBOP (benzotriazol-1-yl-oxytripyrrolidinophosphonium hexafluorophosphate) together with DIPEA (Diisopropylethylamine). Cleavage of the protecting group was accomplished by treatment of 9 and 19, respectively, with trifluoroacetic acid and subsequent precipitation of the hydrochloride salts 10 and 20 with a hydrochloric acid solution in propan-2-ol (Scheme 1B). The compounds were synthesized in satisfactory to excellent yields (53%-93%) and the purity (determined with elemental analyses and HPLC) of most of the products was sufficient for biological evaluation even before the final purification step (recrystallization, flash chromatography). The synthesis of phenoxyanilides 1-8 and 11-18 was achieved by standard amide coupling reactions of commercially available 2-(4-chlorophenoxy)aniline I or 2-(4-methoxyphenoxy)aniline hydrochloride II with acid chlorides in presence of triethylamine, as outlined in Scheme 1A. Compounds 9 and 19 were synthesized by the reaction of 2-(4-substituted)phenoxyanilines and Boc-protected glycine in the presence of PyBOP (benzotriazol-1-yl-oxytripyrrolidinophosphonium hexafluorophosphate) together with DIPEA (Diisopropylethylamine). Cleavage of the protecting group was accomplished by treatment of 9 and 19, respectively, with trifluoroacetic acid and subsequent precipitation of the hydrochloride salts 10 and 20 with a hydrochloric acid solution in propan-2-ol (Scheme 1B). The compounds were synthesized in satisfactory to excellent yields (53%-93%) and the purity (determined with elemental analyses and HPLC) of most of the products was sufficient for biological evaluation even before the final purification step (recrystallization, flash chromatography).

Calculations of Rule-of-Five-properties
We determined the properties concerning Lipinski's Rule of Five (RO5) for orally active drugs for all derivatives. The RO5 can be a useful reference to predict the quality of a structure to be an orally available drug-like compound [13,16], which was outlined by Calderon et al. for TCMDC-137332 [8]. Considering the molecular weight (<380 g/mol), the amount of H-bond acceptors (ď6) and H-bond donors (ď3), the values of every compound in this series comply with the proposed values of Lipinski et al. [13]. Estimation of lipophilicity has been carried out using chemicalize.org for calculation of logP [15]. According to the RO5 the logP should have a maximum value of 5. The calculated logP compared to TCMDC-137332 was reduced by replacement of the tert-butyl residue by smaller residues, cyclic aliphatic residues like cyclopropyl-or cyclobutyl-groups, and more polar groups. Introducing a methoxy group instead of the chlorine residue also reduced the calculated logP value. Some examples of the calculated octanol/water partition coefficients are outlined in Table 1. 1 Cl all derivatives. The RO5 can be a useful reference to predict the quality of a structure to be an orally available drug-like compound [13,16], which was outlined by Calderon et al. for TCMDC-137332 [8].
Considering the molecular weight (<380 g/mol), the amount of H-bond acceptors (≤6) and H-bond donors (≤3), the values of every compound in this series comply with the proposed values of Lipinski et al. [13]. Estimation of lipophilicity has been carried out using chemicalize.org for calculation of logP [15]. According to the RO5 the logP should have a maximum value of 5. The calculated logP compared to TCMDC-137332 was reduced by replacement of the tert-butyl residue by smaller residues, cyclic aliphatic residues like cyclopropyl-or cyclobutyl-groups, and more polar groups. Introducing a methoxy group instead of the chlorine residue also reduced the calculated logP value. Some examples of the calculated octanol/water partition coefficients are outlined in Table 1.

In Vitro Assay
After the full spectroscopic characterization of all congeners and validation of their purity by elemental analyses and chromatographic methods in vitro assays were carried out using P. falciparum NF54-luc erythrocyte stage parasites expressing luciferase. The derivatives 1-20 were examined in 3µM concentration performing viability assays measuring the bioluminescence in a luciferase assay system as described earlier [17][18][19][20]. A first screening, in which the cultures were incubated for 48 h, gave surprisingly low inhibition rates of the parasite growth (see Figure 3). With 22% reduction of the parasite viability, 5 was the most effective compound in this screening, followed by 13 and 15 with about 18% inhibition. The resynthesized TCMDC-137332 (1) failed to show any antiplasmodial activity, just as compounds 10 and 20 which contain polar aminomethyl residues. Due to the unexpected outcome of the first screening session, we performed another assay in which the cultures were incubated for a longer period (96 h), to identify slow acting compounds. However, the prolongation of the incubation time did not improve the inhibitory effect of the derivatives on the parasite growth significantly, the only exception being compound 16, which diminished the proliferation of P. falciparum NF54 by 31%. Compared to the first screening, 16 showed a 3-fold better inhibition rate, indicating that it is probably a rather slow acting inhibitor. However, the results of most of the derivatives were slightly inferior compared to the first screening. The assumed high antiplasmodial activity of TCMDC-137332 (1) could not be confirmed by these assays. 5 available drug-like compound [13,16], which was outlined by Calderon et al. for TCMDC-137332 [8].
Considering the molecular weight (<380 g/mol), the amount of H-bond acceptors (≤6) and H-bond donors (≤3), the values of every compound in this series comply with the proposed values of Lipinski et al. [13]. Estimation of lipophilicity has been carried out using chemicalize.org for calculation of logP [15]. According to the RO5 the logP should have a maximum value of 5. The calculated logP compared to TCMDC-137332 was reduced by replacement of the tert-butyl residue by smaller residues, cyclic aliphatic residues like cyclopropyl-or cyclobutyl-groups, and more polar groups. Introducing a methoxy group instead of the chlorine residue also reduced the calculated logP value. Some examples of the calculated octanol/water partition coefficients are outlined in Table 1.

In Vitro Assay
After the full spectroscopic characterization of all congeners and validation of their purity by elemental analyses and chromatographic methods in vitro assays were carried out using P. falciparum NF54-luc erythrocyte stage parasites expressing luciferase. The derivatives 1-20 were examined in 3µM concentration performing viability assays measuring the bioluminescence in a luciferase assay system as described earlier [17][18][19][20]. A first screening, in which the cultures were incubated for 48 h, gave surprisingly low inhibition rates of the parasite growth (see Figure 3). With 22% reduction of the parasite viability, 5 was the most effective compound in this screening, followed by 13 and 15 with about 18% inhibition. The resynthesized TCMDC-137332 (1) failed to show any antiplasmodial activity, just as compounds 10 and 20 which contain polar aminomethyl residues. Due to the unexpected outcome of the first screening session, we performed another assay in which the cultures were incubated for a longer period (96 h), to identify slow acting compounds. However, the prolongation of the incubation time did not improve the inhibitory effect of the derivatives on the parasite growth significantly, the only exception being compound 16, which diminished the proliferation of P. falciparum NF54 by 31%. Compared to the first screening, 16 showed a 3-fold better inhibition rate, indicating that it is probably a rather slow acting inhibitor. However, the results of most of the derivatives were slightly inferior compared to the first screening. The assumed high antiplasmodial activity of TCMDC-137332 (1) could not be confirmed by these assays. 4 available drug-like compound [13,16], which was outlined by Calderon et al. for TCMDC-137332 [8].
Considering the molecular weight (<380 g/mol), the amount of H-bond acceptors (≤6) and H-bond donors (≤3), the values of every compound in this series comply with the proposed values of Lipinski et al. [13]. Estimation of lipophilicity has been carried out using chemicalize.org for calculation of logP [15]. According to the RO5 the logP should have a maximum value of 5. The calculated logP compared to TCMDC-137332 was reduced by replacement of the tert-butyl residue by smaller residues, cyclic aliphatic residues like cyclopropyl-or cyclobutyl-groups, and more polar groups. Introducing a methoxy group instead of the chlorine residue also reduced the calculated logP value. Some examples of the calculated octanol/water partition coefficients are outlined in Table 1.

In Vitro Assay
After the full spectroscopic characterization of all congeners and validation of their purity by elemental analyses and chromatographic methods in vitro assays were carried out using P. falciparum NF54-luc erythrocyte stage parasites expressing luciferase. The derivatives 1-20 were examined in 3µM concentration performing viability assays measuring the bioluminescence in a luciferase assay system as described earlier [17][18][19][20]. A first screening, in which the cultures were incubated for 48 h, gave surprisingly low inhibition rates of the parasite growth (see Figure 3). With 22% reduction of the parasite viability, 5 was the most effective compound in this screening, followed by 13 and 15 with about 18% inhibition. The resynthesized TCMDC-137332 (1) failed to show any antiplasmodial activity, just as compounds 10 and 20 which contain polar aminomethyl residues. Due to the unexpected outcome of the first screening session, we performed another assay in which the cultures were incubated for a longer period (96 h), to identify slow acting compounds. However, the prolongation of the incubation time did not improve the inhibitory effect of the derivatives on the parasite growth significantly, the only exception being compound 16, which diminished the proliferation of P. falciparum NF54 by 31%. Compared to the first screening, 16 showed a 3-fold better inhibition rate, indicating that it is probably a rather slow acting inhibitor. However, the results of most of the derivatives were slightly inferior compared to the first screening. The assumed high antiplasmodial activity of TCMDC-137332 (1) could not be confirmed by these assays. 4 Considering the molecular weight (<380 g/mol), the amount of H-bond acceptors (≤6) and H-bond donors (≤3), the values of every compound in this series comply with the proposed values of Lipinski et al. [13]. Estimation of lipophilicity has been carried out using chemicalize.org for calculation of logP [15]. According to the RO5 the logP should have a maximum value of 5. The calculated logP compared to TCMDC-137332 was reduced by replacement of the tert-butyl residue by smaller residues, cyclic aliphatic residues like cyclopropyl-or cyclobutyl-groups, and more polar groups. Introducing a methoxy group instead of the chlorine residue also reduced the calculated logP value. Some examples of the calculated octanol/water partition coefficients are outlined in Table 1.

In Vitro Assay
After the full spectroscopic characterization of all congeners and validation of their purity by elemental analyses and chromatographic methods in vitro assays were carried out using P. falciparum NF54-luc erythrocyte stage parasites expressing luciferase. The derivatives 1-20 were examined in 3µM concentration performing viability assays measuring the bioluminescence in a luciferase assay system as described earlier [17][18][19][20]. A first screening, in which the cultures were incubated for 48 h, gave surprisingly low inhibition rates of the parasite growth (see Figure 3). With 22% reduction of the parasite viability, 5 was the most effective compound in this screening, followed by 13 and 15 with about 18% inhibition. The resynthesized TCMDC-137332 (1) failed to show any antiplasmodial activity, just as compounds 10 and 20 which contain polar aminomethyl residues. Due to the unexpected outcome of the first screening session, we performed another assay in which the cultures were incubated for a longer period (96 h), to identify slow acting compounds. However, the prolongation of the incubation time did not improve the inhibitory effect of the derivatives on the parasite growth significantly, the only exception being compound 16, which diminished the proliferation of P. falciparum NF54 by 31%. Compared to the first screening, 16 showed a 3-fold better inhibition rate, indicating that it is probably a rather slow acting inhibitor. However, the results of most of the derivatives were slightly inferior compared to the first screening. The assumed high antiplasmodial activity of TCMDC-137332 (1) could not be confirmed by these assays. 3 Considering the molecular weight (<380 g/mol), the amount of H-bond acceptors (≤6) and H-bond donors (≤3), the values of every compound in this series comply with the proposed values of Lipinski et al. [13]. Estimation of lipophilicity has been carried out using chemicalize.org for calculation of logP [15]. According to the RO5 the logP should have a maximum value of 5. The calculated logP compared to TCMDC-137332 was reduced by replacement of the tert-butyl residue by smaller residues, cyclic aliphatic residues like cyclopropyl-or cyclobutyl-groups, and more polar groups. Introducing a methoxy group instead of the chlorine residue also reduced the calculated logP value. Some examples of the calculated octanol/water partition coefficients are outlined in Table 1.

In Vitro Assay
After the full spectroscopic characterization of all congeners and validation of their purity by elemental analyses and chromatographic methods in vitro assays were carried out using P. falciparum NF54-luc erythrocyte stage parasites expressing luciferase. The derivatives 1-20 were examined in 3µM concentration performing viability assays measuring the bioluminescence in a luciferase assay system as described earlier [17][18][19][20]. A first screening, in which the cultures were incubated for 48 h, gave surprisingly low inhibition rates of the parasite growth (see Figure 3). With 22% reduction of the parasite viability, 5 was the most effective compound in this screening, followed by 13 and 15 with about 18% inhibition. The resynthesized TCMDC-137332 (1) failed to show any antiplasmodial activity, just as compounds 10 and 20 which contain polar aminomethyl residues. Due to the unexpected outcome of the first screening session, we performed another assay in which the cultures were incubated for a longer period (96 h), to identify slow acting compounds. However, the prolongation of the incubation time did not improve the inhibitory effect of the derivatives on the parasite growth significantly, the only exception being compound 16, which diminished the proliferation of P. falciparum NF54 by 31%. Compared to the first screening, 16 showed a 3-fold better inhibition rate, indicating that it is probably a rather slow acting inhibitor. However, the results of most of the derivatives were slightly inferior compared to the first screening. The assumed high antiplasmodial activity of TCMDC-137332 (1) could not be confirmed by these assays.

In Vitro Assay
After the full spectroscopic characterization of all congeners and validation of their purity by elemental analyses and chromatographic methods in vitro assays were carried out using P. falciparum NF54-luc erythrocyte stage parasites expressing luciferase. The derivatives 1-20 were examined in 3 µM concentration performing viability assays measuring the bioluminescence in a luciferase assay system as described earlier [17][18][19][20]. A first screening, in which the cultures were incubated for 48 h, gave surprisingly low inhibition rates of the parasite growth (see Figure 3). With 22% reduction of the parasite viability, 5 was the most effective compound in this screening, followed by 13 and 15 with about 18% inhibition. The resynthesized TCMDC-137332 (1) failed to show any antiplasmodial activity, just as compounds 10 and 20 which contain polar aminomethyl residues. Due to the unexpected outcome of the first screening session, we performed another assay in which the cultures were incubated for a longer period (96 h), to identify slow acting compounds. However, the prolongation of the incubation time did not improve the inhibitory effect of the derivatives on the parasite growth significantly, the only exception being compound 16, which diminished the proliferation of P. falciparum NF54 by 31%. Compared to the first screening, 16 showed a 3-fold better inhibition rate, indicating that it is probably a rather slow acting inhibitor. However, the results of most of the derivatives were slightly inferior compared to the first screening. The assumed high antiplasmodial activity of TCMDC-137332 (1) could not be confirmed by these assays.

Apparatus and Materials
Starting materials were purchased from the suppliers indicated below and were used without further purification. Pivaloyl chloride was from Merck-Schuchardt (Hohenbrunn, Germany), 2,2-dimethylbutyryl chloride, 3-methylbutyryl chloride and 4-methoxybenzoyl chloride were from

Calculations of Lipinski Properties
Calculations of the octanol/water partition coefficients (logP) were performed using chemicalize.org [15]. The calculator plugins are based on Viswanadhan's fragmentation methods [22], the PHYSPROP © database [23] and an additional data set [24]. All three approaches were weighted equally for the calculations.

In Vitro Antimalarial Activity Assay
Erythrocytic stages of transgenic NF54-luc P. falciparum were used for the luciferase-based viability assay. These parasites constitutively express high levels of luciferase. The parasites were cultured as described earlier [19,25]. Firstly, the culture was dispensed in triplicate into white 96-well flat bottom plates (each well contains 250 µL) (NUNC, Roskilde, Denmark) with parasitemia of 0.5%-1%. Then the cultures were incubated in the presence of a 3 µM concentration of the test compounds for 48 h (37˝C, 90% N 2 , 5% CO 2 , and 5% O 2 ). Subsequently, 100 µL RPMI1640 media was removed from each well and a 100 µL volume of the Bright-Glo ® substrate solution was added to each well. One of the cleavage products of the reaction is light, which was measured by a FLUOROSKAN FL luminometer (Thermo) machine, thereby detecting the amount of living parasites. The experiments were repeated with an incubation of 96 h. Untreated cultures were used as negative controls and to calculate the inhibition rate (0% inhibition of parasite growth). Blasticidin (Sigma-Aldrich, St. Louis, MO, USA), a drug used for selection of transfected parasites, was included as a positive control on each plate and gave >90% inhibition of parasite growth at 2 µg/mL.

Conclusions
In conclusion, the previously published N-[2-(4-chlorophenoxy)phenyl]-2,2-dimethylpropanamide (TCMDC-137332, 1) [7] and 19 derivatives were synthesized in satisfactory to excellent yields and were fully characterized regarding identity and purity. The predicted antimalarial activity of the herein examined hit structure of the TCAMS (1) was not confirmed in biological evaluation experiments. The growth of Plasmodium falciparum NF54 strains was not diminished in the expected nanomolar concentration by the 2-phenoxyanilide congeners. However, in a prolonged in vitro assay, in which cultures were incubated for 96 h, N-[2-(4-methoxyphenoxy)phenyl]cyclobutanecarboxyamide (16) was identified to be a moderate inhibitor of the proliferation of Plasmodium falciparum NF54 strains.