Investigating the antiplasmodial activity of primary sulfonamide compounds identified in open source malaria data.

In the past decade there has been a significant reduction in deaths due to malaria, in part due to the success of the gold standard antimalarial treatment - artemisinin combination therapies (ACTs). However the potential threat of ACT failure and the lack of a broadly effective malaria vaccine are driving efforts to discover new chemical entities (NCEs) to target this disease. The primary sulfonamide (PS) moiety is a component of several clinical drugs, including those for treatment of kidney disease, glaucoma and epilepsy, however this chemotype has not yet been exploited for malaria. In this study 31 PS compounds sourced from the GlaxoSmithKline (GSK) Tres Cantos antimalarial set (TCAMS) were investigated for their ability to selectively inhibit the in vitro growth of Plasmodium falciparum asexual stage malaria parasites. Of these, 14 compounds were found to have submicromolar activity (IC50 0.16-0.89 μM) and a modest selectivity index (SI) for the parasite versus human cells (SI > 12 to >43). As the PS moiety is known to inhibit carbonic anhydrase (CA) enzymes from many organisms, the PS compounds were assessed for recombinant P. falciparum CA (PfCA) mediated inhibition of CO2 hydration. The PfCA inhibition activity did not correlate with antiplasmodial potency. Furthermore, no significant difference in IC50 was observed for P. falciparum versus P. knowlesi (P > 0.05), a Plasmodium species that is not known to contain an annotated PfCA gene. Together these data suggest that the asexual intraerythrocytic stage antiplasmodial activity of the PS compounds examined in this study is likely unrelated to PfCA inhibition.

Introduction

Malaria remains one of the world's most important infectious diseases, causing approximately 438,000 deaths in 2015, mainly African children under the age of five (WHO, 2015). While it is possible that the first generation RTS,S malaria vaccine will be employed in some regions in the future, the World Health Organization (WHO) remains cautious and recommends that other malaria prevention and treatment strategies continue, including the development of new drugs (WHO, 2016). This recommendation is driven by the threat of malaria parasite resistance or reduced clinical efficacy emerging to all current antimalarial drugs including the gold standard artemisinin combination therapies (ACTs) (WHO, 2015, Fairhurst and Dondorp, 2016). Added to this, the majority of agents in the current antimalarial drug development portfolio are based on known antimalarial pharmacophores (Wells et al., 2015), which may compromise their widespread use due to potential issues of cross resistance. With few antimalarial chemical classes (e.g. spiroindoline, imidazolepiperazine and triazolopyrimidine chemotypes) presently under advanced development (MMV, 2016) there is an urgent need to ensure that the antimalarial drug discovery pipeline is primed with new chemical entities, ideally those with novel modes of action to avoid cross-resistance to existing drugs. The primary sulfonamide (PS) chemotype is not currently used for malaria prevention or treatment, but has a proven track record for treatment of other diseases, including glaucoma, renal disorders and epilepsy (Poulsen, 2010, Supuran, 2010, Pastorek and Pastorekova, 2015, Supuran and Winum, 2015).

In previous work we identified several PS glycosides with moderate in vitro antiplasmodial activity (50% growth inhibitory concentration (IC50) ∼1 μM) and selectivity for the parasite versus human cells (Selectivity Index (SI) > 40) (Andrews et al., 2013). Published work from another group identified a thioureido benzenesulfonamide with similar in vitro activity against P. falciparum (IC50 ∼1 μM) and in vivo activity against P. berghei in a mouse malaria model (ID50 10 mg/kg) (Krungkrai et al., 2008). Additional evidence that PS compounds have antimalarial potential comes from high throughput screening of a GlaxoSmithKline (GSK) library of ∼2,000,000 compounds. The results of the GSK screen led to compilation of the Tres Cantos antimalarial set (TCAMS), with data made publicly available as a resource for antimalarial lead identification and basic research into the “druggable” genome of P. falciparum through deposition in the open access European Bioinformatics Institute ChEMBL Neglected Tropical Disease archive. The TCAMS dataset contains ∼13,500 compounds that inhibit the in vitro growth of drug-sensitive (3D7) and multi-drug resistant (Dd2) P. falciparum parasites (≥80% and ≥50% at 2 μM, respectively) (Gamo et al., 2010). Following a substructure search of this open source malaria data, we identified 31 PS compounds (Fig. 1, Fig. 2) that were subsequently provided by GSK and investigated in this study.

Fig. 1

Structures of TCAMS PS compounds with an antimalarial pharmacophore. Diaminopyrimidine based compounds highlighted in blue; 4-aminoquinoline based compounds highlighted in red. Compound 4 and 6 were tested as the formate salt, compounds 3, 5, 7 were tested as the trifluoracetate salt. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

Structures of TCAMS PS compounds with no antimalarial pharmacophore. Group 1 - Imidazo[1,2-b]pyridazines; Group 2 - 2-Pyrimidinyl pyrazolopyridines; Group 3 – Quinolines; Group 4 – Benzimidazoles; Group 5 – Quinazolinamines and Group 6. Trifluoracetate salt - 18, 24 Hydrochloride salt -23.

PS compounds are known to inhibit carbonic anhydrase (CA) enzyme activity in many organisms (Supuran, 2008). CA enzymes maintain an important physiological equilibrium: the hydration of carbon dioxide to bicarbonate anion and a proton: H2O + CO2 ⇆ HCO3− + H+ and are responsible for HCO3− and pH homeostasis, including within erythrocytes. Malaria parasite CA inhibitors were first suggested as a potential new class of antimalarials in 1998 (Sein and Aikawa, 1998) and later the esterase activity of P. falciparum CA (PfCA; PlasmoDB (Aurrecoechea et al., 2009) gene ID number PF3D7_1140000) was shown to be inhibited by a series of PS compounds (Krungkrai et al., 2008). We further characterized PfCA from P. falciparum 3D7, identifying this protein as the first member of a new family of CAs (the η-CA family) (Del Prete et al., 2014). Using purified recombinant PfCA expressed in E. coli, hydration of CO2 was shown to be inhibited by clinically used PS compounds (Vullo et al., 2015). However, a limitation of previous studies was that in vitro antiplasmodial activity had not been correlated with inhibition of PfCA-mediated CO2 hydration. In this study the antiplasmodial potency (in vitro IC50) of PS compounds from the TCAMS set were determined and compared with enzymatic activity against human CA I (hCA I), human CA II (hCA II) and PfCA. Based on these data, structure activity relationship (SAR) analysis was carried out in order to assess potential correlation of in vitro IC50 values with recombinant PfCA inhibition data. Finally, to assess the ability of PS compounds to target different Plasmodium species, three of the most potent antiplasmodial PS compounds were also assessed for activity against laboratory adapted P. knowlesi A1H.1 (Moon et al., 2013), a species that normally causes malaria in macaque monkeys but that can also cause severe disease and death in humans (Jongwutiwes et al., 2004, Jongwutiwes et al., 2011, Luchavez et al., 2008, Kantele and Jokiranta, 2011, Khim et al., 2011, Barber et al., 2013, Lee et al., 2013).

Materials and methods

Compounds

GSK PS compounds

31 PS compounds identified in the European Bioinformatics Institute ChEMBL Neglected Tropical Disease TCAMS database (www.ebi.ac.uk/chemblntd/download/#tcams) were supplied by GSK (Gamo et al., 2010) under a Material Transfer Agreement (MTA). All compounds were provided as 10 mM stocks diluted in 100% DMSO and diluted as required. Chloroquine and pyrimethamine were purchased from Sigma-Aldrich®, USA and stock solutions of 10 mM were prepared in 100% DMSO (Sigma-Aldrich®, USA) and diluted as required for assays.

In vitro P. falciparum growth inhibition assays

The antiplasmodial activity of compounds was tested against P. falciparum using a 72 h isotopic microtest, as previously described (Fisher et al., 2014). Briefly, synchronous ring-stage P. falciparum infected RBCs (0.25% parasitemia and 2.5% final hematocrit) were seeded into 96-well tissue culture plates (3596 Costar®, Corning, USA) containing serial dilutions of control or test compounds. After incubating for 48 h under standard P. falciparum culture conditions in hypoxanthine-free parasite culture media, 0.5 μCi [3H]-hypoxanthine (Perkin Elmer®, USA) was added to each well followed by culturing for a further 24 h. Cells were harvested onto 1450 MicroBeta filter mats (Wallac, USA) and [3H]-incorporation was determined using a 1450 MicroBeta liquid scintillation counter (Perkin Elmer®, USA). The percentage inhibition of growth compared to that of matched DMSO controls (0.5%; Sigma-Aldrich, USA) was determined for at least three independent experiments, each carried out in triplicate wells. IC50 values were calculated using linear interpolation of inhibition curves (Huber and Koella, 1993). Chloroquine was used as a positive control in all assays.

In vitro P. knowlesi growth inhibition assays

The in vitro antimalarial growth inhibition activity of compounds was tested against P. knowlesi infected erythrocytes using a modified [3H]-hypoxanthine incorporation assay (Arnold et al., 2016). Briefly, asynchronous P. knowlesi A1H.1 parasites (0.25% parasitemia and 2% final hematocrit) were seeded into 96-well tissue culture plates (3596 Corning®, USA) containing serial dilutions of control or test compounds in parasite culture media (Moon et al., 2013). Assay plates were incubated for 48 h with 0.5 μCi [3H]-hypoxanthine (PerkinElmer®, USA) added at the 24 h time point. All other assay procedures, including data analysis, controls and independent replicates, were as described for P. falciparum above.

CA enzyme assays

An SX.18MV-R Applied Photophysics (Oxford, UK) stopped-flow instrument was used to assay inhibition of CA isozymes for each test compound (Khalifah, 1971). A saturated CO2 solution in water at 25 °C was used as substrate. Phenol Red (0.2 mM) was used as indicator, working at the absorbance maximum of 557 nm, 10 mM Hepes (pH 7.4) as buffer, Na2SO4 or NaClO4 (0.1 M, anions are not inhibitory at this concentration) to maintain constant ionic strength was used to follow the CA-catalyzed CO2 hydration reaction for 5–10 s. A stock solution of each test compound was prepared in DMSO-water 1:1, v/v (10 mM). At least 7 different inhibitor concentrations (prepared by serial dilution of stock solutions with the assay buffer) in triplicate were used to measure the CA inhibition constant. Inhibitor (I) and enzyme (E) solutions were preincubated for 10 min at room temperature prior to assay to allow for the formation of the E-I complex. The IC50 values were obtained by non-linear least-squares methods as reported earlier (Del Prete et al., 2014, Vullo et al., 2015), and represent the mean from at least three different determinations. All CA isozymes were recombinant proteins obtained as reported earlier by our group (Del Prete et al., 2014, Vullo et al., 2015). The concentration of enzyme in the assay system was: 12.3 nM for hCA I, 8.7 nM for hCA II and 14.5 nM for PfCA.

Results and discussion

Identification of PS compounds in the Tres Cantos antimalarial set (TCAMS)

Data mining of the GSK TCAMS open source antimalarial data using the PS chemotype (-SO2NH2) as the search fragment resulted in identification of 31 PS-containing compounds (Fig. 1, Fig. 2). Of these, seven compounds contain a known antimalarial pharmacophore, either the diaminopyrimidine scaffold (diaminopyrimidine group; compounds 1-6) or a 4-aminoquinoline scaffold (4-aminoquinoline group; compounds 7 and 8) (Fig. 1A and B, respectively). Compounds lacking a known antimalarial pharmacophore were grouped where possible according to structural similarity (Groups 1-5; Fig. 2). Seven TCAMS PS compounds did not share structural similarity with other compounds and were separately grouped (Group 6, Fig. 2). As published data for GSK TCAMS compounds was limited to percentage inhibition at 2 μM, the in vitro IC50 of compounds was determined against P. falciparum drug sensitive (3D7) and multi-drug resistant (Dd2) parasite lines (Table 1).

Table 1

In vitro antiplasmodial activity and selectivity of PS compounds 1–31.

CPD #	CHEMBL ID	MW	cLogP	HepG2a %Inhib @ 10μM	Pf3D7b IC50 μM	PfDd2c IC50 μM	SIe
DAP/PS
1	546456	467	3.2	21	0.62 ± 0.14	0.42 ± 0.01	>16
2	587256	459	3.7	74	1.31 ± 0.18	<2d	<8
3	524967	626	4.8	71	>5	<2d	<2
4	547936	558	4.5	14	0.28 ± 0.04	0.37 ± 0.09	>36
5f	548709	401	−0.5	7	0.26 ± 0.06	0.22 ± 0.09	>38
6	588030	441	1.9	2	0.84 ± 0.02	0.48 ± 0.08	>12
AQ/PS
7	548576	637	4.9	56	0.57 ± 0.12	>2d	<18
8	589061	362	3.2	5	1.57 ± 0.02	>2d	>6
Group 1
9	530148	414	2.9	26	0.84 ± 0.01	<2d	>12
10	537064	463	4.3	10	0.47 ± 0.01	0.41 ± 0.04	>21
11	533004	410	2.2	19	1.89 ± 0.13	<2d	>5
12	530711	404	2.1	7	0.46 ± 0.01	0.43 ± 0.02	>22
13	536144	410	1.6	16	0.46 ± 0.01	0.86 ± 0.08	>22
Group 2
14f	587021	598	4.5	83	1.09 ± 0.13	<2d	<9
15f	582203	598	4.5	81	3.04 ± 0.01	<2d	<3
Group 3
16	530087	454	2.5	39	3.21 ± 0.03	>2d	>3
17	533876	440	2.5	40	0.79 ± 0.03	<2d	>13
18	529796	484	1.5	35	>5	<2d	<2
Group 4
19	586746	437	4.6	0	1.59 ± 0.15	<2d	>6
20	529534	390	2.0	16	0.85 ± 0.05	<2d	>12
21	528642	442	5.2	24	2.00 ± 0.22	>2d	>5
22	529358	420	4.9	0	3.57 ± 0.11	>2d	>3
Group 5
23	587392	540	4.4	18	>5	<2d	<2
24	587392	518	4.2	54	1.92 ± 0.39	ND	<5
Group 6
25	533539	317	0.6	17	0.69 ± 0.15	<2d	>14
26	586070	301	1.9	6	0.81 ± 0.02	<2d	>12
27	580443	370	1.3	17	1.55 ± 0.12	<2d	>6
28	533079	367	0.9	39	0.23 ± 0.01	0.16 ± 0.04	>43
29	548077	538	3.2	1	3.02 ± 0.23	<2d	>3.3
30	531249	236	1.3	0	>5	>2d	<2
31	530277	442	5.2	24	>5	>2d	<2
CQ	76	320	4.3	0	0.01 ± 0.003	0.05 ± 0.01	>1000
PYR	36	249	2.8	0	0.03 ± 0.01	17.28 ± 6.90	>333

HepG2% inhibition at 10 μM data derived from ChEMBL GSK TCAMS data (Gamo et al., 2010); Mean 50% inhibitory concentration (IC50) ± standard deviation against P. falciparum line 3D7b and Dd2c for three independent experiments, each performed in triplicate (this study); P. falciparum Dd2 IC50 data derived from ChEMBL GSK TCAMS data (Gamo et al., 2010) SI selectivity index (HepG2 IC50/Pf IC50); Compounds have potential serine/threonine protein kinase activity. All other data derived from ChEMBL GSK TCAMS data (Gamo et al., 2010). DAP/PS PS compounds with a diaminopyrimidine group; AQ/PS PS compounds with a 4-aminoquinoline group; CQ chloroquine; PYR pyrimethamine.

Antiplasmodial structure activity relationship (SAR) analysis of compounds containing an antimalarial pharmacophore

Diaminopyrimidine group

Diaminopyrimidines are a class of compound that comprise two amino substituents directly attached to a pyrimidine ring (Fig. 1A; highlighted in blue), and include the antimalarial drug pyrimethamine which is a dihydrofolate reductase inhibitor (Falco et al., 1951, Gregson and Plowe, 2005). Six of the PS compounds (1-6) contained a diaminopyrimidine scaffold as part of their structure. Of these, compounds 4 and 5 displayed the most potent IC50s against P. falciparum 3D7 parasites (IC50 0.28 and 0.26 μM, respectively; Table 1) and the highest selectivity for P. falciparum parasites versus the mammalian cell line HepG2 (SI > 35; Table 1). It should be noted that compounds 4 and 5 have opposing cLog P values (+4.5 and −0.5, respectively) indicating that there may be potential to improve the potency and selectivity of compound 5 via rational medicinal chemistry approaches to improve membrane permeability. Compound 1 (Pf3D7 IC50 0.62 μM) and 6 (Pf3D7 IC50 0.84 μM) showed lower antiplasmodial activity and selectivity (SI < 20; HepG2/Pf3D7; Table 1) compared to 4 and 5. Compound 2 had the poorest antiplasmodial activity and selectivity of the diaminopyrimidines (Pf3D7 IC50 1.3 μM; HepG2, 74% inhibition at 10 μM; Table 1). Compounds 1, 4, 5, and 6 also displayed submicromolar antiplasmodial activity against the P. falciparum Dd2 parasite line (IC50s 0.22–0.48 μM), which is resistant to drugs including chloroquine, pyrimethamine and sulfadoxine (Noedl et al., 2003). The resistance index (Ri), which is the ratio of the IC50s of the resistant line Dd2 to the sensitive line 3D7, for these compounds ranged from 0.6 to 1.3, indicating a lack of cross resistance with Dd2.

4-Aminoquinoline group

The 4-aminoquinoline scaffold is another known antimalarial pharmacophore (Fig. 1B; highlighted in red), with the antimalarial drug chloroquine being a 4-aminoquinoline derivative (Greenwood, 1992). Compounds 7 and 8 contain a 4-aminoquinoline and PS group (IC50 0.57 and 1.57 μM, respectively; Table 1). Although compound 7 has a submicromolar IC50 against P. falciparum 3D7, it has only moderate selectivity for the parasite versus the HepG2 mammalian cell line (56% inhibition at 10 μM; SI ∼18; Table 1) and the TCAMS database data indicates that 7 is less active against chloroquine resistant P. falciparum Dd2 parasites (IC50 > 2 μM; Ri > 3.5). Given that P. falciparum Dd2 is resistant to chloroquine, (Noedl et al., 2003), it is possible that the 4-aminoquinoline contributes to its activity. As the P. falciparum Dd2 IC50 was not calculated for this group, the activity data for compound 8 are not sufficient to be able to make any conclusions about cross resistance.

SAR analysis of GSK PS compounds lacking an antimalarial drug pharmacophore

GSK PS compounds lacking a known antimalarial drug pharmacophore (Fig. 2) were categorised into groups according to structure similarity.

Group 1 - Imidazo[1,2-b]pyridazines

Group 1 compounds 9-13 are all differently substituted phenyl/pyridinyl imidazo[1,2-b]pyridazines with the PS attached to a benzyl amine substituent. The common imidazopyridazine core has previously been associated with inhibition of P. falciparum calcium dependent protein kinase 1 (PfCDPK1) inhibition, however compounds 9-13 were found not to inhibit several Plasmodium kinases including PfCDPK1 and PfCDPK4 (Crowther et al., 2016). Of the five compounds in this group, compounds 10 (trifluoromethoxyphenyl substituent), 12 (nitrilephenyl substituent), and 13 (methoxypyridinyl substituent) displayed the greatest level of antiplasmodial activity (Pf3D7 IC50 0.46–0.47 μM; PfDd2 IC50 0.41–0.86 μM; Table 1) and moderate parasite-specific selectivity (SI > 20 HepG2/Pf3D7; Table 1) while, the chlorophenyl and methoxyphenyl variants (9 and 11) had reduced antiplasmodial activity (Pf3D7 IC50 0.84 μM and 1.89 μM respectively; Table 1) indicating that the different substituents of these compounds had a small impact on antiplasmodial activity.

Group 2 - 2-pyrimidinyl pyrazolopyridines

Group 2 comprises compounds 14 and 15 (Pf3D7 IC50 1.09 and 3.04 μM, respectively; PfDd2 IC50 < 2 μM; Table 1), which are 2-pyrimidinyl pyrazolopyridines. The compounds differ in the position of the PS group substituent of the benzene ring. Position 4 (14) on the benzene ring may be responsible for the somewhat better Pf3D7 activity with respect to position 5 (15) for the placement of the PS group, however additional analogues would be required for SAR to be confirmed. However, these compounds have poor selectivity for Pf3D7 versus human cells (SI < 10 HepG2/Pf3D7; Table 1) and therefore are of limited interest as lead compounds. No SAR conclusions can be made for compounds 14 and 15 with respect to PfDd2 activity as accurate IC50s were not determined.

Group 3 - quinolines

Group 3 consists of substituted quinoline compounds 16, 17 and 18. Of this group compound 17, a 3-[4-(3-sulfamoylphenyl)quinolin-6-yl]benzenesulfonamide, displayed the best antiplasmodial activity (Pf3D7 IC50 0.79 μM; Table 1) and moderate selectivity (SI > 10 HepG2/Pf3D7; Table 1) and minimal cross-resistance (IC50 < 2 μM; Ri < 2.5). In contrast, compound 16 which is a 3-[4-[3-(sulfamoylmethyl)phenyl]quinolin-6-yl]benzenesulfonamide displayed approximately 4-fold decrease in antiplasmodial activity (Pf3D7 IC50 3.2 μM; PfDd2 IC50 > 2 μM; Table 1) and poor selectivity (SI ∼3: HepG2/Pf3D7; Table 1), indicating that the additional methylene group of 16 caused reduced antiplasmodial activity and lower parasite selectivity. Compound 18 comprises a dimethylaminosulfonamide pyridine moiety and has the lowest antiplasmodial activity of the group (Pf3D7 IC50 > 5 μM; PfDd2 IC50 < 2 μM; Table 1).

Group 4 - benzimidazoles

Group 4 consists of compounds 19-22 all of which contain a benzimidazole-5-sulfonamide group with different substitutions. Of this group compound 20, 2-[2-amino-5-(3-cyanophenyl)pyridin-3-yl]-3H-benzimidazole-5-sulfonamide, displayed the best antiplasmodial activity (Pf3D7 IC50 0.85 μM; Table 1) and moderate parasite selectivity (SI > 12 HepG2/Pf3D7; Table 1) with no cross resistance to P. falciparum Dd2 (Table 1; IC50 < 2 μM; Table 1). Compound 20 is a novel benzimidazolyl-pyridine and it should be noted that this class of compound has also been implicated in the inhibition of mammalian serum and glucocorticoid-regulated kinase 1 (SGK-1) activity (Gaulton et al., 2012). However, this target has yet to be validated in Plasmodium. The remaining compounds in Group 4 (19, 21 and 22) showed poor antiplasmodial activity and parasite selectivity (Pf3D7 IC50 1.59–3.57 μM; SI > 3 - >6 HepG2/Pf3D7; Table 1).

Group 5 - Quinazolinamines

Group 5 comprised two compounds, 23 and 24, that contain a 4-quinazolinamine pharmacophore. The compounds differ by fluorination on the 2-phenyl substituent, with two hydrogens of 24 replaced with fluorine in 23. Additionally the pyrrolidine fragment of compound 24 has an additional methylene spacer to the 4-quinazolinamine pharmacophore.

Compound 24 displayed better antiplasmodial activity (Pf3D7 IC50 1.92 μM; Table 1) than compound 23 (Pf3D7 IC50 > 5 μM; Table 1), suggesting that the fluorine groups of compound 23 may contribute to reduced antiplasmodial activity.

Group 6

The remaining seven compounds from the GSK PS compounds (Group 6; 25-31; Fig. 2) do not share structural similarity with a known antimalarial pharmacophore or with other PS compounds to allow informative SAR analysis. Of this group, compound 28, a 3-[(4-pyrazolo[1,5-b]pyridazin-3-ylpyrimidin-2-yl)amino]benzenesulfonamide had the best antiplasmodial activity (Pf3D7 IC50 0.23 μM), highest selectivity (SI > 44: HepG2/Pf3D7) and no evidence of cross resistance to P. falciparum Dd2 parasites (IC50 0.16 μM) (Table 1). Compound 25, a N-methyl-5-[(E)-2-(2-sulfamoylphenyl)ethenyl]pyridine-3-carboxamide and compound 26 a 4-hydroxyquinazoline displayed 3- to 4-fold lower antiplasmodial activity (Pf3D7 IC50 0.69–0.81 μM) and selectivity (SI > 12: HepG2/Pf3D7) than 28 (Table 1). The remaining compounds 27, 29, 30 and 31 showed poor antiplasmodial activity (Pf3D7 IC50 1.55 - >5 μM) and low selectivity (SI > 3 to >7 HepG2/Pf3D7). Compounds 25-29 also demonstrated some antiplasmodial activity against the drug resistant P. falciparum parasites (PfDd2 IC50 < 2 μM; Table 1).

Relationship between PS compound inhibition of P. falciparum growth in vitro and biochemical potency against PfCA

The PS moiety is in the structure of several drugs where the mechanism of action is attributed to CA inhibition (Supuran, 2008). CAs from several pathogens including Plasmodium are now recognised as potential anti-infective drug targets (Capasso and Supuran, 2013). This, together with the current dogma that PfCA is likely to be essential for parasite survival as it is believed to be involved in the first step of the crucial pyrimidine synthesis pathway in P. falciparum (Krungkrai et al., 2003), led us to investigate PfCA as a possible target of 26 (compounds 3, 6, 9, 12 and 19 were not assayed due to insufficient stocks) of the 31 GSK PS compounds in this study. In addition, a potential limitation of PS-based drugs for malaria use is that they may also inhibit hCA I and hCA II found in erythrocytes (and/or other human CAs), hence CA inhibition data was also investigated against hCA I and hCA II (Table 2). Acetazolamide (AZA), the par excellence therapeutically established CA inhibitor, was included as a reference compound (Supuran, 2008).

Table 2

Carbonic anhydrase enzyme inhibition of PS compounds.

Compd	IC50 (μM)			Selectivity
	hCA Ia	hCA IIb	PfCAc	hCA I/PfCA	hCA II/PfCA
DAP/PS
1	4.40	2.47	8.00	0.55	0.31
2	2.29	6.39	8.63	0.27	0.74
4	8.59	2.54	7.48	1.15	0.34
5	9.52	7.30	6.81	1.40	1.07
AQ/PS
7	1.69	3.98	8.52	0.20	0.47
8	0.21	0.35	0.53	0.39	0.67
Group 1
10	0.89	3.59	6.16	0.14	0.58
11	3.11	0.69	6.37	0.49	0.11
13	0.76	0.64	1.90	0.40	0.34
Group 2
14	3.96	0.70	8.69	0.46	0.08
15	3.14	1.17	8.13	0.39	0.14
Group 3
16	9.40	1.27	0.91	10.29	1.39
17	0.95	0.17	5.19	0.18	0.03
18	0.75	0.31	6.10	0.12	0.05
Group 4
20	8.401	0.44	0.70	11.99	0.62
21	6.23	6.01	5.30	1.18	1.14
22	3.39	2.60	2.63	1.29	0.99
Group 5
23	0.69	7.29	0.90	0.76	8.06
24	6.13	7.48	6.91	0.89	1.08
Group 6
25	4.31	4.83	4.73	0.91	1.02
26	0.38	0.65	4.21	0.09	0.16
27	0.71	2.58	5.63	0.13	0.46
28	0.91	2.21	6.78	0.13	0.33
29	8.03	5.08	7.75	1.04	0.66
30	0.40	0.29	0.78	0.52	0.38
31	9.19	9.66	0.90	10.27	10.80
AZA	0.28	0.02	0.37	0.76	0.05

Mean % inhibition against human ahuman CA I, bhuman CA II and cP. falciparum CA (±standard deviation) for three independent experiments, each performed in triplicate. DAP/PS PS compounds with a diaminopyrimidine group; AQ/PS PS compounds with a 4-aminoquinoline group; AZA acetazolamide.

PfCA inhibition for PS compounds ranged from IC50 0.53–8.69 μM, with six compounds (8, 16, 20, 23, 30 and 31) having submicromolar activity against the enzyme (Table 2). Only one of these contained an antimalarial pharmacophore (4-aminoquinoline, compound 8) and this compound displayed 16-fold greater inhibition for PfCA (IC50 0.53 μM) than the other 4-aminoquinoline, compound 7 (IC50 8.52 μM). However compound 8 showed no selectivity at the enzyme level for PfCA compared to hCA I or hCA II (IC50 0.21 μM and 0.35 μM, respectively; Table 2). Group 3 quinoline compound 16 has submicromolar inhibitory activity against PfCA, ∼5- to 6-fold better than for quinolines 17 and 18. Of note, compound 16 is 10-fold more selective for PfCA over hCA I and non-selective for hCA II, while the two analogues 17 and 18 have 5- to 31-fold and 8- to 20-fold, respectively, better inhibition for the human CAs. At least for compounds 16 and 17, this may suggest that the different orientation of one of the benzenesulfonamide groups leads to improved PfCA inhibitory activity and selectivity over human CA I for 16 (PfCA IC50 0.91 μM SI; hCA I IC50/PfCA IC50 ∼10) compared to 17 (PfCA IC50 5.19 μM SI; hCA I IC50/PfCA IC50 < 1). Group 4 benzimidazole 20 has submicromolar inhibitory activity against PfCA (IC50 0.71 μM; Table 2) and like 16 from Group 3, is more selective for PfCA over hCA I but not hCA II (SI; hCA I IC50/PfCA IC50 ∼12; SI; hCA II IC50/PfCA IC50 < 1). In Group 5, the submicromolar inhibitory activity of quinazolinamine 23 compared with ∼8-fold less activity for analogue 24 (IC50 0.90 and 6.90 μM, respectively) indicates that replacement of hydrogen with fluorine may improve PfCA inhibition. In contrast to the PfCA selectivity of 16 and 20 for hCA I over hCA II, 23 shows the opposite profile having 8-fold better selectivity for PfCA over hCA II (IC50 0.90 and 7.29 μM, respectively) and no selectivity for PfCA over hCA I (IC50 0.90 and 0.69 μM, respectively). Two Group 6 compounds, structurally unrelated compounds 30 and 31, exhibited submicromolar PfCA inhibition activity with 31 being ∼10-fold more selective for PfCA over both human CAs. Additionally, it should be noted that two of the most potent antiplasmodial compounds that did not contain an antimalarial pharmacophore (10 and 28; Pf3D7 IC50 0.47 and 0.23 μM, respectively) had poor PfCA inhibition (IC50 6.16–6.78 μM) but displayed submicromolar inhibition for hCA I (IC50 0.89–0.91 μM). This potential off target effect needs to be taken in to consideration if these compounds are to be further investigated as candidates for malaria therapy.

Despite individual inhibitors having inhibitory activity in PfCA enzymatic assays, no correlation between PfCA enzymatic IC50s and those determined against P. falciparum infected erythrocytes was found for compounds for which IC50s were available for both assays (i.e. for 22 of the PS compounds (1, 2, 4, 5, 7, 8, 10, 11, 13-17, 20-22, 24-29); Fig. 3). At the Group level, either no correlation was found or numbers were too small to analyse. Likewise, there was no correlation between either of the human CA IC50s and the P. falciparum 3D7 IC50s (not shown).

Fig. 3

Relationship between inhibitor potency against Spearman correlation (two-tailed) analysis of the mean IC50 against P. falciparum infected erythrocytes and PfCA enzyme for 22 PS compounds (r = −0.1316; P = 0.556).

Antiplasmodial activity of selected GSK PS compounds against P. knowlesi

One of the desired attributes of new antimalarial leads is the ability to target multiple human infecting Plasmodium species. P. knowlesi is a zoonotic species that can cause death in humans (Barber et al., 2011, William et al., 2011) and is also a possible model for P. vivax which causes significant malaria morbidity (WHO, 2015). P. knowlesi is the only Plasmodium species other than P. falciparum that is amenable to long term continuous culture in vitro (Moon et al., 2013). In addition, of relevance to this study, whole genome sequencing has failed to identify a PfCA homologue in P. knowlesi (Aurrecoechea et al., 2009). Three of the GSK PS compounds that lack an antimalarial drug pharmacophore and with the highest potency against P. falciparum (10, 12 and 28) were selected for in vitro testing for activity against P. knowlesi infected erythrocytes. No significant difference (P > 0.05) in IC50s was seen for the three compounds for P. knowlesi (IC50 0.46 ± 0.07, 0.62 ± 0.15 and 0.19 ± 0.05 μM, respectively) compared to those obtained for P. falciparum 3D7 (IC50 0.47 ± 0.01, 0.46 ± 0.01 and 0.23 ± 0.01 μM, respectively) (Table 3). Together with the lack of correlation of biological activity with enzyme activity discussed above, and the absence of a PfCA homologue in P. knowlesi, these data are consistent with the PS compounds in this study having an alternative target to PfCA. It should be noted, however, that we cannot discount the presence of an orthologue of PfCA in P. knowlesi given that a large proportion of the genome cannot yet be annotated based on amino acid sequence homology (Aurrecoechea et al., 2009). In addition, this study focuses on asexual intraerythrocytic stage Plasmodium parasites and it may be that the enzymatic activity observed here is of relevance to other Plasmodium lifecycle stages as PfCA is also reported to be expressed in the ookinete life cycle stage in the mosquito (Aurrecoechea et al., 2009).

Table 3

Comparison of the activity of selected PS compounds against P. falciparum and P. knowlesi in vitro.

Compound	P. falciparum IC50 (μM)a	P. knowlesi IC50 (μM)b
10	0.47 ± 0.01	0.46 ± 0.07
12	0.46 ± 0.01	0.62 ± 0.15
28	0.23 ± 0.01	0.19 ± 0.05
CQ	0.02 ± 0.001	0.01 ± 0.003

Mean % inhibition against aP. falciparum line 3D7 and bP. knowlesi (±standard deviation) for three independent experiments, each performed in triplicate.

Conclusion

Information sharing of antimalarial high throughput screening data by pharmaceutical company GSK allowed us the opportunity to source and investigate the activity and potential mode of action of a panel of compounds containing a novel pharmacophore, the primary sulfonamide (PS) moiety. Whilst the PS moiety is known to target CA in other organisms (Supuran, 2008) most compounds did not selectively inhibit PfCA over human red blood cell CAs, with compound 31 being an exception. However the poor antiplasmodial activity of 31 (Pf3D7 IC50 > 5μM) and the good cell permeability properties (LogP 5.2) suggests no correlation with PfCA inhibition, as least in asexual blood stage parasites. Nonetheless, several PS compounds with submicromolar antiplasmodial activity (IC50 0.16–0.47 μM) and some selectivity (SI > 21 - >43) were identified and these compounds could be investigated further for their potential to be developed as antimalarial lead candidates. This includes PS diaminopyrimidines 4 and 5 and compounds 10, 12 and 28. While compounds 4 and 5 are structurally similar to the antimalarial drug pyrimethamine they did not exhibit cross resistance with a pyrimethamine resistant parasite line. It should be also noted that compounds 10, 12 and 28 displayed promising in vitro activity and selectivity across two Plasmodium species (P. falciparum and P. knowlesi IC50 < 0.5 μM; SI > 21). This, along with structural novelty, warrants investigating these three compounds further to determine if they possess the appropriate pharmacokinetic properties and in vivo efficacy to progress as antimalarial drug leads.