EVALUATION OF AROMATIC 6-SUBSTITUTED THIENOPYRIMIDINES AS SCAFFOLDS AGAINST PARASITES THAT CAUSE TRYPANOSOMIASIS, LEISHMANIASIS, AND MALARIA

Abstract

Target repurposing is a proven method for finding new lead compounds that target Trypanosoma brucei, the causative agent of human African trypanosomiasis. Due to the recent discovery of a lapatinib-derived analog 2 with excellent potency against T. brucei (EC₅₀ = 42 nM) and selectivity over human host cells, we have explored other classes of human tyrosine kinase inhibitor scaffolds in order to expand the range of chemotypes for pursuit. Following library expansion, we found compound 11e to have an EC₅₀ of 84 nM against T. brucei cells while maintaining selectivity over human hepatocytes. In addition, the library was tested against causative agents of Chagas’ disease, leishmaniasis, and malaria. Two analogs with sub-micromolar potencies for T. cruzi (4j) and Plasmodium falciparum (11j) were discovered, along with an analog with considerable potency against Leishmania major amastigotes (4e). Besides identifying new and potent protozoan growth inhibitors, these data highlight the value of concurrent screening of a chemical library against different protozoan parasites.

Introduction

Human African trypanosomiasis (HAT) is a life-threatening parasitic disease that places 70 million people at risk in parts of Sub-Saharan Africa. Caused by the protozoan parasite Trypanosoma brucei, it is one of 17 neglected tropical diseases (NTDs) highlighted by the World Health Organization as needing new therapies. Current medicines for HAT are not orally bioavailable, have unfavorable toxicity profiles and drug resistant parasites are emerging.¹

Phosphodiesterase and kinase target repurposing² are rapid and pragmatic methods of lead drug discovery explored in our laboratory.^3-8 Launching new antiparasitic drug discovery efforts by redirecting inhibitors of human drug targets that are homologous to essential parasite targets or pathways can facilitate rapid development of structure-activity relationships (SAR), and accelerate “hit” optimization.

Inhibition of human receptor tyrosine kinases (RTKs) is a widely employed approach for cancer therapeutics.^9-12 Specifically, an epidermal growth factor receptor (EGFR) inhibitor lapatinib (GW572016, Tykerb, 1, Figure 1) was approved by the FDA in 2007 for solid tumors and breast cancer.^13-16 Compound 1 proved to be a promising lead for HAT due to its modest potency against T. brucei with an EC₅₀ of 1.54 μM, and its ability to cure 25% of infected mice in a murine model of HAT.¹⁷ Beginning with lapatinib, extensive SAR studies led to compound 2, which has an EC₅₀ of 42 nM and excellent selectivity over human cells. Importantly, compound 2 was orally bioavailable in mice and had a modest effect on parasitemia and life extension in a murine T. brucei bloodstream infection despite exhibiting high plasma protein binding (>99%) and a lack of central nervous system exposure.⁷ These observed pharmacokinetic properties are not surprising due to its high LogP and molecular weight.

Figure 1.

Project strategy.

The previous SAR explorations of the lapatinib chemotype focused on the “head” and “tail” regions of the molecule, described in Figure 1. Historical tyrosine kinase inhibitor lead discovery efforts have led to many quinazoline scaffold replacements in cancer chemotherapy discovery.^18-21 GW837016X (3) possesses a related core scaffold and was also found to be potent against T. brucei. It therefore seemed appropriate to utilize the thieno[3,2-d]pyrimidine (and the regioisomeric thieno[2,3-d]pyrimidine) scaffolds to launch a crossover SAR study, preparing analogs matched to those quinazoline-based inhibitors previously described⁷ (denoted in this article as 4b-m, Table 1).

Table 1.

Antiparasitic activities of compounds against T. brucei, T. cruzi, L. major, and P. falciparum.

Cmpd	Scaffolda	R	T. brucei EC50 (μM) or % inh (at 5 μM)b	T. cruzi EC50 (μM) or % inh (at 10 μM)b	L. major		P. falciparum D6 EC50 (μM)d
					Amast EC50 (μM)c	Promast EC50 (μM)c
2e			0.042 ± 0.010	1.8 ± 0.9	7.98 r2 = 0.75	2.97 r2 = 0.76	0.23 r2 = 0.99
11a			42%	>50	5.4 r2 = 0.81	>20	0.86 r2 = 0.98
18a			0.40 ± 0.05	2.2 ± 0.4	2.60 r2 = 0.94	>20	0.52 r2 = 0.99
4be			3.9 ± 0.3	1.8 ± 0.1	>15	4.05 r2 = 0.85	0.79 r2 = 0.90
11b			2.8 ± 0.1	0%	>15	8.14 r2 = 0.86	2.97 r2 = 0.98
18bf			1.7 ± 0.2	42%	>15	>20	>20
4ce			4.7 ± 0.2	6.4 ± 4.0	ndg	ndg	ng
11c			1.2 ± 0.1	0.67 ± 0.06	2.37 r2 = 0.92	>20	1.07 r2 = 0.96
18c			0.5 ± 0.2	>50.0	>15	>20	>15
4de			3.2 ± 0.1	5.0 ± 0.0	4.67 r2 = 0.91	3.49 r2 = 0.82	0.52 r2 = 0.98
11d			1.4 ± 0.0	2.7 ± 2.3	>3	>3	0.65 r2 = 0.96
18d			1.3 ± 0.1	2.4 ± 1.0	>15	11.75 r2 = 0.90	0.52 r2 = 0.97
4e			1.0 ± 0.1	49%	1.14 r2 = 0.85	>20	0.26 r2 = 0.96
11e			0.084 ± 0.0	3.3 ± 1.2	>3	>3	0.23 r2 = 0.98
18e			0.42 ± 0.1	29%	>3	>3	0.28 r2 = 0.99
4f			0.22 ± 0.00	2.2 ± 0.4	4.09 r2 = 0.92	>20	0.60 r2 = 0.96
11f			0.89 ± 0.10	30 ± 13	>15	>20	1.65 r2 = 0.98
18f			1.6 ± 0.0	17%	2.03 r2 = 0.84	>20	7.79 r2 = 0.93
4ge			5.3 ± 0.2	Nd	>15	6.86 r2 = 0.78	5.79 r2 = 0.98
11g			10%	>50.0	3.50 r2 = 0.84	>20	6.21 r2 = 0.46
18g			1.2 ± 0. 3	45%	>15	>20	10.78 r2 = 0.94
4he			0.14 ± 0.1	>50.0	>15	>3	0.51 r2 = 0.98
11h			46%	3.2 ± 0.5	3.10 r2 = 0.93	>20	1.17 r2 =0.94
18h			0.33 ± 0.03	>50.0	>15	>20	2.11 r2 = 0.98
4ie			6.0 ± 0.2	>50.0	>15	>20	0.91 r2 = 0.99
11i			8%	>50.0	>15	>20	>15
18i			2.9 ± 0.3	49%	>3	>3	>4
4je,f			1.0 ± 0.05	0.60 ± 0.15	5.91 r2 = 0.94	2.74 r2 = 0.88	0.063 r2 = 0.99
11j			1.1 ± 0.0	>50.0	>15.0	0.22 r2 = 0.84	0.027 r2 = 0.97
18j			60%	1.7 ± 0.1	4.20 r2 = 0.91	>20	0.089 r2 = 0.81
4ke,f			1.9 ± 0.9	10.1%	>15	4.40 r2 = 0.82	0.27 r2 = 0.98
11k			0.47 ± 0.25	0%	>15	>20	3.10 r2 = 0.98
4le			6.5 ± 1.0	>50.0	>15	>20	>20
11l			2.2 ± 0.1	0.75 ± 0.02	1.58 r2 = 0.83	>20	0.44 r2 = 0.98
4m		H	1.3 ± 0.1	2.4 ± 0.2	12.36 r2 = 0.78	1.75 r2 = 0.77	5.33 r2 = 0.69
11m			1%	27 ± 3.5	>15	5.68 r2 = 0.94	16.4 r2 = 0.93

^aAr is defined as 3-chloro-4-((3-fluorobenzyl)oxy)phenyl.

^bCompounds showing ±75% growth inhibition at 5 or 10 μM for T. brucei or T. cruzi, respectively, were tested for EC₅₀ values. T. brucei EC₅₀ values are the result of duplicate experiments, within ± 25%, with the exception of 11e (± 33%), and 11k (± 52%). T. cruzi EC₅₀ values are the result of duplicate experiments, within ± 50%, with the exception of 4c (± 63%), and 11d (± 85%).

^cCompounds screened against L. major amastigotes and promastigotes were tested in duplicate and had r² values >0.75.

^dCompounds were tested in duplicate against P. falciparum (D6 strain) and had r² values >0.90 except for 11g (r² = 0.46) and 18j (r² = 0.81) against P. falciparum.

^ePreviously reported data utilizing the same assay as described in the Electronic Supplementary Information.⁷

^fAll compounds were inactive against HepG2 cell lines except for 4b (TC₅₀ = 4.9 μM), 4j (12.9 μM), 4k (9.6 μM), and 18b (10.0 μM).

^gNot determined due to low solubility in the assay conditions.

We recently demonstrated the value of simultaneously testing compounds against related protozoan parasites, wherein it became evident that producing parasite-dependent SAR could direct future work against each pathogen.⁸ In addition to the screening of new compounds against T. brucei, compounds were screened against the kinetoplastid parasites Leishmania major (causative agent of cutaneous leishmaniasis) and T. cruzi (which causes Chagas disease), as well as the malaria-causing parasite Plasmodium falciparum. In this report we disclose the results of our efforts to prepare and test thienopyrimidine-based inhibitors, matching the “tail” replacements explored in the previous study that led to compound 2.⁷

Synthetic procedure

The requisite dihalo template precursors (9 and 16, Schemes 1 and 2) of the thieno[3,2-d]pyrimidine and thieno[2,3-d]pyrimidine scaffolds were made following previously disclosed protocols.^22-25 The thieno[3,2-d]pyrimidine intermediate 9 was synthesized starting from methyl 3-aminothiophene-2-carboxylate (5) as shown in Scheme 1. This commercially available material underwent formylation, followed by cyclization with formamide and ammonium formate to afford intermediate 7. The pyrimidinone 7 was chlorinated using phosphorous oxychloride to provide 8 and brominated using 1,2-dibromo-1,1,2,2-tetrafluoroethane to yield the key thienopyrimidine 9. From this point, halide displacement with the requisite aniline in refluxing isopropanol produced the bromide intermediate 10, amenable for subsequent Suzuki chemistry.

Scheme 1.

Synthesis of the thieno[3,2-d]pyrimidine template.

Scheme 2.

Synthesis of the thieno[2,3-d]pyrimidine Template

Reagents and conditions:(a) HCO₂H, Ac₂O, 0 °C to RT, 12 h, 79%; (b) HCONH₂, NH₄HCO₂, 150 °C to RT, 17 h, 56%; (c) POCl₃, 110 °C, 12 h, 90%; (d) LDA, C₂Br₂F₄, THF, -78 °C to RT, 13 h, 94%; (e) 3-chloro-4-((3-fluorobenzyl)oxy)aniline, IPA, 85 °C, 12 h, 88%; (f) Boronic ester, Na₂CO_3, Pd(PPh₃)₄, DME/EtOH (3:2), H₂O, 80-85 °C, 12 h, 8-36%

The thieno[2,3-d]pyrimidine template was synthesized from commercially available methyl 2-cyanoacetate (12, Scheme 2), which underwent Gewald reaction conditions with 1,4-dithiane-2,5-diol to produce the desired aminothiophene 13. Cyclization of the thiophene with formamide afforded intermediate 14, followed by sequential bromination (15), and chlorination to provide 16. As before, aniline displacement of the chloride was effected smoothly, providing 17 for further diversification via Suzuki reaction. For both scaffolds, intermediate 10 or intermediate 17 was coupled with boronic esters or acids using Suzuki reaction conditions to generate the respective final compounds 11a-m and 18a-j, which are matched with corresponding quinazoline analogs previously reported (Table 1).⁷

Reagents and conditions: (a) 1,4-dithiane-2,5-diol, TEA, DMF, 45 °C, 2 h, 58%; (b) HCONH₂, 170 °C, 6 h, 38%; (c) Br₂, AcOH, 80 °C, 2 h, 79%; (d) POCl₃, 100 °C, 12 h, 75%; (e) 3-chloro-4-((3-fluorobenzyl)oxy)aniline, IPA, 85 °C, 12 h, 88%; (f) Boronic ester, Na₂CO₃, Pd(PPh₃)₄, DME/EtOH (3:2), H₂O, 85 °C, 12 h, 20-46%

Results and discussion

Tests of the compounds against T. brucei revealed favorable growth inhibitory activity amongst three distinct scaffolds shown in Table 1. To make the direct comparison with 2, replacement of the quinazoline with thieno[2,3-d]pyrimidine (18a) yielded a 10x change in potency, whereas the thieno[3,2-d]pyrimidine replacement (11a) resulted in a dramatic loss of potency. Removal of the morpholine altogether (4b, 11b, 18b) or the tail altogether (4m, 11m) reduced the potency to the micromolar range.

An additional atom linker between the phenyl and morpholine was more favored on the quinazoline scaffold (4j) than the thienopyrimidines (11j and 18j) but still showed reduced potency. Evaluation of the morpholine regiochemistry revealed that para (4k, 11k) was less potent than meta (2, 11a, 18a), though alteration of this regiochemistry was better tolerated when attached to the thieno[3,2-d]pyrimidine 11k than the quinazoline scaffold 4k. The most potent compound from this series is 11e, an 84 nM inhibitor of T. brucei cell growth. This “tail” was not as potent in the quinazoline scaffold (4e); the morpholinophenyl tail present in 2 remains the most potent analog overall with an EC₅₀ of 42 nM. For the most potent analogs in the thieno[2,3-d]pyrimidine scaffold, both the 5-methyl-6-morpholinopyridin-3-yl tail (18e) and the morpholinophenyl tail (18a) were equipotent with EC₅₀ values of 420 and 400 nM respectively, but the 3-phenyl(piperidin-1-yl)methanone analog (18h) was slightly more potent with an EC₅₀ of 330 nM.

Figure 2 displays the effect of changing the scaffold, while maintaining the R-group substitution, upon anti-T. brucei activity. We can make a few observations from this qualitative analysis. First, in the majority of cases, switching from quinazoline (black circles) to thieno[2,3-d]pyrimidine (red circles) tends to improve compound potency (i.e., reduces EC₅₀ 3-fold or more). Second, when comparing matched analogs between the quinazoline and thieno[3,2-d]pyrimidine scaffolds, except for 18c versus 4c, which is about a 9-fold difference; 0.5 versus 4.7 μM), the difference in activity is seldom greater than two-fold. This indicates that thieno[3,2-d]pyrimidines, unlike thieno[2,3-d]pyrimidines, are equipotent to quinazolines when tested against T. brucei. Third, the range of activities for compounds with thienopyrimidine scaffolds is narrower (0.084 - 2.9 μM) than the quinazolines (0.042 – 6.5 μM). Fourth, for matched pairs of the [3,2-d] and [2,3-d] scaffold, there is seldom more than a 2-fold difference in EC₅₀ (the only exception being 11j (EC₅₀= 1.1 μM) vs 18j, which only inhibits parasite growth 60% at 5 μM ).

Figure 2.

Plot of anti-trypanosome activity as a function of R-group (X-axis) and scaffold (black-quinazoline, red = thieno[2,3-d]pyrimidine; green = thieno[3,2-d]pyrimidine. R group letter corresponds to compound substitutions listed in Table 1.

Table 1 also shows the results of testing these analogs against the other protozoan parasites. Compound 2 and its related thienopyrimidine-derived analogs 11a and 18a had shown sub-micromolar potency towards the malaria parasite P. falciparum. Removal of the “tail” (4m, 11m) or of the morpholine (4b, 11b, 18b) produced a significant loss in potency across all 3 scaffolds. Any ortho substitution on the phenyl portion of the tail was not well tolerated by any set. The three tails that produced the most potent analogs were the meta substituted methylsulfonyl (4d, 11d, 18d), the para substituted methyl-pyridinylmorpholine (4e, 11e, 18e), and the para substituted benzylmorpholine (4j, 11j, 18j). These tails consistently showed sub-micromolar EC₅₀ values across all three scaffolds, suggesting that appropriately placed H-bond acceptors may be crucial to antimalarial activity in this region. These compounds were also tested against two drug resistant strains of malaria (W2, C235) and were generally found to show potencies within 3-fold of the D6 strain (Table S2, Figure S1 Electronic Supplemental Information (ESI)).

Although 2 and the corresponding thieno[2,3-d]pyrimidine analog 18a displayed promising micromolar potency against T. cruzi, overall the thieno[3,2-d]pyrimidine analogs were more potent in this study, with the best compounds being 4j (EC₅₀ = 0.60 μM), 11c (0.67 μM), and 11l (0.75 μM). Quinazolines tolerated the removal of the morpholino group better than either of the thienopyrimidines. Though the para-morpholinophenyl tail (4k, 11k) was inactive in this assay, adding an intervening methylene group (4j and 11l) resulted in active analogs.

Screening of more than 35 compounds against amastigotes of Leishmania major failed to produce any sub-micromolar compounds. The most potent analog across all 3 scaffolds against Leishmania major amastigotes was 4e with an EC₅₀ of 1.14 μM. Against promastigotes of L. major only 11j had sub-micromolar potency (EC₅₀ = 220 nM). As seen with other drug sensitivity studies, there is often little-to-no correlation between the effect of compounds on intracellular amastigotes and the promastigote form of Leishmania.²⁶

Most compounds showed no toxicity to the HepG2 cell line except 4b, 4j, 4k, and 18b with TC₅₀ values (concentration leading to 50% growth inhibition of host cells) of 4.9, 12.9, 9.6, and 10.0 μM respectively.

We tested the two most potent T. brucei growth inhibitors 2 and 11e in in vitro ADME assays, and observed that, as predicted based on their high cLogP (7.6 and 7.4 respectively), both have poor solubility (< 1 μM), high protein binding (>99%), and modest metabolic stability (Table S3, ESI). Nonetheless, as the goal of this report was to describe SAR between the quinazoline and thienopyrimidine scaffolds, we are encouraged by the reasonable SAR tracking that is observed between the different scaffolds. Work continues on all three scaffolds to improve their physicochemical properties, primarily focused upon reducing size and lipophilicity: Results will be reported in due course.

Conclusions

We have previously shown that some lapatinib-derived analogs have high potency towards the HAT parasite T. brucei and with little host cell toxicity, typified by 2 (EC₅₀ = 42 nM). In assessing matched [3,2-d] and [2,3-d] thienopyrimidine scaffold replacements, we have discovered analogs, such as 11e, that have excellent potency against T. brucei cells (EC₅₀ = 84 nM). Also, a multi-pathogen testing campaign using these T. brucei inhibitors against select protozoan parasites produced compounds with a variety of favorable potencies: 4j, T. cruzi EC₅₀ = 0.60 μM; 4e, L. major amastigote EC₅₀ = 1.14 μM and 11j, P. falciparum EC₅₀ = 0.027 μM. While these data are indicative of good starting points for antiparasite drug discovery, attention must be paid to improvement of the physicochemical properties of these agents in order to produce effective drugs.