Scaffold and Parasite Hopping: Discovery of New Protozoal Proliferation Inhibitors

Abstract

Utilizing a target repurposing and parasite-hopping approach, we tested a previously reported library of compounds that were active against Trypanosoma brucei, plus 31 new compounds, against a variety of protozoan parasites including Trypanosoma cruzi, Leishmania major, Leishmania donovani, and Plasmodium falciparum. This led to the discovery of several compounds with submicromolar activities and improved physicochemical properties that are early leads toward the development of chemotherapeutic agents against kinetoplastid diseases and malaria.

Neglected tropical diseases (NTDs) are a group of life-threatening diseases affecting a significant population of the world. These diseases are especially devastating in developing and economically disadvantaged countries. Most of the treatments available to treat NTDs have serious adverse effects. Further, the issues of emerging resistance and a lack of affordable drugs are driving the search for new treatments or vaccines; this search is a priority for the World Health Organization (WHO). Chagas disease and leishmaniasis, caused by Trypanosoma cruzi and Leishmania spp., respectively, are among the 20 NTDs highlighted by the WHO as urgently needing new therapies.¹⁻³

Chagas disease, also known as American trypanosomiasis, is a life-threatening illness endemic mostly in Latin American countries that is transmitted primarily by the kissing bug (Reduviidae family). According to the WHO, 8 million people worldwide are estimated to be infected by T. cruzi. It is responsible for 41% of heart failure cases in Latin America^4,5 killing more than 10,000 people per year. More than 25 million people are at risk of acquiring Chagas disease.⁶ Leishmaniasis is a group of diseases caused by more than 20 different species of the Leishmania parasite transmitted by the sandfly and affecting humans in 98 countries. Approximately 1.3 million new cases and 20,000 to 30,000 deaths occur annually due to leishmaniasis.⁷ Though not considered an NTD, malaria is a life-threatening tropical disease caused by five species of protozoan Plasmodium parasites (mainly Plasmodium falciparum and P. vivax) that are transmitted by the bite of infected female Anopheles mosquitoes. In 2016, an estimated 216 million cases of malaria occurred worldwide leading to approximately 445,000 malaria-related deaths worldwide.⁸

As part of our drug discovery program for tropical diseases, we have employed a target class repurposing approach⁹ wherein we identified lapatinib as an antitrypanosomal agent.⁹ The synthetic reoptimization of lapatinib and its related analogs for activity against T. brucei resulted in the discovery of NEU-1953(10,11) (1, Figure 1), which was further optimized to improve aqueous solubility as an anti-HAT (human African trypanosomiasis) agent.¹² Here we report the assessment of this family of compounds against P. falciparum, Leishmania spp., and T. cruzi, along with some newly prepared analogs.

Figure 1.

Optimization of lapatinib leading to the identification of NEU-1953 (1)^10,11 as a multiparasite inhibitor

Synthesis

Compounds 17, 18, 22–27, 33–35, 39, 42–45, S21–S23, S25–S27, S31–S35, and S38–S41 were synthesized using previously reported protocols^10,12,13 with slight modifications as described in the Supporting Information (Schemes S1–S7). All other compounds (1–16, 28–32, 36–38, 40, 41, S1–S20, S28–S30, and S36–S37) were procured from our internally available T. brucei compound library.^11,12

Trypanosoma cruzi

Screening the compounds against intracellular T. cruzi amastigotes revealed 12 compounds with submicromolar activity and an additional 25 compounds with low micromolar activity (≤10 μM) (Tables 1–3 and Supporting Information Tables S1–S3). The hit selection criteria against T. cruzi is EC₅₀ < 10 μM with a 10-fold selectivity over host cells.¹⁴ Considering 1 as a starting point (T. cruzi EC₅₀ = 6.0 μM),¹¹ several modifications were made to understand the complete structure–activity relationships (SAR) profile. Removal of the N-methyl from the piperazine (8, EC₅₀ > 50 μM) resulted in a complete loss of activity; the N-methyl is slightly favored over N-ethyl (11, EC₅₀= 7.7 μM) or N-propyl(12, EC₅₀= 7.6 μM). Further, 11 and 12 were found to be toxic against the host cell line (C2C12; Supporting Information Table S7). The insertion of a bridging carbon in the piperazine ring (10) or increasing sp³ content by replacing the piperazine ring with the homopiperazine (9, EC₅₀ = 1.2 μM) and spirocyclic amines (13, EC₅₀ = 0.52 μM and 14, EC₅₀ = 0.22 μM) led to an increase in solubility, better ADME properties, and equipotent or improved potency against T. cruzi compared to 1. The N-atoms of the pyrimidine were not necessary for potency (cf. 2 and 18) but play an important role in the aqueous solubility (2, aq. sol. = 1.7 μM and 18, aq. sol. = 4.1 μM). The insertion of a hydrogen bond donor (−NH) in the tail region (6) abrogated the potency. The potency and selectivity data of tail group modifications and ortho-methyl substituted compounds are listed in Table 1.

Table 1. Inhibition Profile of Quinolines 1–18 with Modified Tail Regions against T. cruzi, L. major, and P. falciparum D6.

^aAll SEM values within 25%.

^bSI = C2C12 or NIH3T3 TC₅₀/T. cruzi EC₅₀.

^cAll r² values are >0.8 unless noted otherwise.

^dSI = HepG2 TC₅₀/L. major EC₅₀.

^eAll r² values are >0.9 unless noted otherwise.

^fSI = HepG2 TC₅₀/P. falciparum D6 EC₅₀.

Table 3. Inhibition Profile of C-6 Substituted Quinolines (33–35), Quinolinimines (36–39), Pseudoring/Open Ring Analogs (40–43), and Isocryptolepines (44–45) against T. cruzi, L. major, L. donovani, and P. falciparum D6ⁱ.

^aAll SEM values within 25%.

^bSI = C2C12 TC₅₀/T. cruzi EC₅₀.

^cAll r² values are >0.8 unless noted otherwise.

^dSI = HepG2 TC₅₀/L. major EC₅₀.

^eAll SEM values within 25%.

^fSI = B10R TC₅₀/L. donovani EC₅₀.

^gAll r² values are >0.9 unless noted otherwise.

^hSI = HepG2 TC₅₀/P. falciparum D6 EC₅₀.

ⁱnd = not determined.

Further modifications were performed on the headgroup, and some crossover compounds (19–32 and S24–S30) with different head and tail combinations were synthesized (Table 2 and Supporting Information Table S2). Replacement of the pyrazine with saturated rings led to the identification of 28 and 29 with EC₅₀ values of 2.1 and 1.2 μM, respectively, against T. cruzi. In general, substituted anilines such as 22 (EC₅₀ = 0.060 μM) and 23 (EC₅₀ = 0.080 μM) were preferred over heterocyclic head groups, but a decrease in aqueous solubility was observed (Supporting Information, Table S8).

Table 2. Inhibition Profile of Quinolines 19–32 with Crossover Head and Tail Combinations against T. cruzi, L. major, L. donovani, and P. falciparum D6ⁱ.

^aAll SEM values within 25%.

^bSI = C2C12 TC₅₀/T. cruzi EC₅₀.

^cAll r² values are >0.8 unless noted otherwise.

^dSI = HepG2 TC₅₀/L. major EC₅₀.

^eAll SEM values within 25%.

^fSI = B10R TC₅₀/L. donovani EC₅₀.

^gAll r² values are >0.9 unless noted otherwise.

^hSI = HepG2 TC₅₀/P. falciparum D6 EC₅₀.

ⁱnd = not determined.

In an attempt to explore the SAR around the scaffold, a range of quinolinimines, open chain compounds, isocryptolepines, and C-6 substituted quinolines were synthesized and the potency and selectivity data are listed in Table 3 and Supporting Information Table S3. Compound 33 (EC₅₀ = 0.33 μM) was the most potent compound with improved and acceptable physicochemical and ADME properties.

The quinolinimine analog of 1 (39) possessed better aqueous solubility and in vitro ADME properties but led to a loss in potency. These quinolinimine modifications led to the identification of two low micromolar compounds (37, EC₅₀ = 3.0 μM and 38, EC₅₀ = 2.1 μM).

In an attempt to increase the conformational flexibility of the quinoline series, we synthesized a series of open chain analogs (75–83) by breaking one of the central core rings. Only two compounds (42 and 43) from the open chain series were found to have micromolar activity against T. cruzi. Alternatively, isocryptolepine analogs (84–86) were designed and synthesized to lock the free rotation of the headgroup. This strategy identified compounds 44 (EC₅₀ = 0.24 μM) and 45 (EC₅₀ = 0.21 μM) as potent T. cruzi inhibitors, but the aqueous solubility was decreased.

The general SAR trends observed for this series against T. cruzi and a representative best compound with favorable ADME properties are summarized in Figure 2. The selectivity index (SI) of all compounds tested against T. cruzi was determined with reference to their toxicity against host cells (C2C12 or NIH3T3). While there was no particular trend observed for toxicity against host cells, several active compounds (13, 19, 22, 23, 24, 25, and 33) have SI values above our targeted threshold (SI > 10).

Figure 2.

SAR trends for T. cruzi summarized around 1 and representative best compound from the series (33).

Leishmania major

Screening against L. major intracellular amastigotes resulted in several compounds with potency ≤10 μM; the potency and selectivity data of all compounds tested are listed in Tables 1–3 and Supporting Information Tables S1–S3. The SAR trends observed for L. major were different from that of T. cruzi. Starting with modifications around 1 (L. major EC₅₀ > 15 μM), the insertion of a hydrogen bond donor (−NH) in the tail region (3, 4, and 6) was favored and resulted in improved potency against L. major with 6 showing improved aqueous solubility (129 μM) as well as other in vitro ADME properties.¹¹ The removal of nitrogen atoms from the pyrimidine of 6 (EC₅₀ = 1.7 μM) resulted in a reduction in potency (5, EC₅₀ = 7.9 μM).¹¹Ortho-methylation of the aromatic core resulted in entirely different responses, methylation at C-6 resulted in improved potency (15, EC₅₀ = 1.8 μM), whereas methylation at C-8 resulted in a loss of potency (16, EC₅₀ > 24 μM). The replacement of N-methyl in the tail region with the carbamate (7, EC₅₀ = 0.35 μM), N-propyl (12, EC₅₀ = 4.3 μM), and N-methylsulfonyl (17, EC₅₀= 0.73 μM) resulted in an enhancement of potency against L. major.

Three quinolinimines were active in the micromolar range; 36 (EC₅₀ = 2.3 μM), 38 (EC₅₀ = 1.5 μM), and 39 (EC₅₀ = 2.3 μM) were most promising. Analogs bearing substituted anilines (23, EC₅₀ = 1.6 μM and 24, EC₅₀ = 0.22 μM) resulted in improved potency but at the cost of significantly reduced solubility. Replacement of the pyrazine with saturated rings and the variation of 6-substituted quinolines did not result in any active compounds. One compound from the isocryptolepines (45, EC₅₀ = 0.44 μM) and open chain analogs (40, EC₅₀ = 0.36 μM) were found to be submicromolar inhibitors of L. major. The general SAR trends observed against L. major and representative best compound with improved ADME properties are summarized in Figure 3. The SI of all compounds tested against L. major was determined with reference to their toxicity against HepG2 cells.

Figure 3.

SAR trends for L. major around 1 and representative best compound from the series (17).

Leishmania donovani

Based on the positive results against L. major, we tested selected analogs against L. donovani, one of the species that causes visceral leishmaniasis (Tables 2–3 and Supporting Information, Tables S2, S3, and S7). The observed activity trends were not similar across the two species (albeit we note that the host cell lines utilized were different for each species). Screening against L. donovani resulted in three compounds with low micromolar inhibition (EC₅₀ ≤ 10 μM) and four compounds with submicromolar potency (EC₅₀ < 1 μM). Replacing the pyrazine of 1 with tetrahydropyran resulted in an active compound against L. donovani (29, EC₅₀= 4.0 μM). Replacing the piperazine with the homopiperazine and the pyrazine with 4-(trifluoromethoxy)aniline resulted in 22 with improved potency (EC₅₀ = 0.085 μM) though this was associated with toxicity against host cells (TC₅₀ = 0.66 μM; SI = 7.8). Further, modifying the core to the quinolinimine led to 36 (EC₅₀ = 0.02 μM); and different head and tail combinations identified 19 (EC₅₀ = 0.023 μM) as the most potent compound among all those tested. The general SAR trends observed for this series against L. donovani and the representative best compound with favorable ADME properties are summarized in Figure 4. The SI of all compounds tested against L. donovani was determined with reference to their toxicity against host cells (B10R).

Figure 4.

SAR trends for L. donovani around 1 and representative best compound from the series (36).

Based upon the high potency against L. donovani, two compounds (19 and 36) were selected for in vivo efficacy studies in an acute infection model in both female and male mice at 50 mg/kg b.i.d. for 14 days. The results are summarized in Supporting Information, Figure S1 and Table S4. Compound 36 has a better ADME and pharmacokinetic profile¹² over 19 but neither compound showed any significant effect on reducing the parasitemia in mice.

Plasmodium falciparum

We have also recently reported the medicinal chemistry optimization of lapatinib where 1 was found to be a potent inhibitor of P. falciparum.¹⁰ Continuing our program on antimalarial drug discovery and based upon our previous observations that 4-anilinoquinolines have activity against P. falciparum,^10,15 this library of compounds was also tested against three different strains of P. falciparum D6 (chloroquine sensitive, mefloquine resistant), W2 (chloroquine resistant, mefloquine sensitive), and C235 (chloroquine, mefloquine, and pyrimethamine resistant). The screening against P. falciparum resulted in 42 compounds with low nanomolar activity (<0.1 μM), 28 compounds with submicromolar activity (0.1–1 μM), and eight compounds with low micromolar activity (1–10 μM). The inhibition profile of all tested compounds against P. falciparum D6 and the associated SI values is presented in Tables 1–3 and Supporting Information Tables S1–S3; and the tabulated values against W2 and C235 are provided in the Supporting Information, Table S5. As previously reported, 1(10) was active across all three strains of P. falciparum (D6EC₅₀ = 0.026 μM, W2EC₅₀ = 0.062 μM, and C235EC₅₀ = 0.037 μM). The pyrimidine nitrogen atoms were not necessary for potency (2, D6EC₅₀ = 0.074 μM). Variations to the N-alkyl on the piperazine (N-ethyl, 11, D6EC₅₀ = 0.029 μM; N-propyl, 12, D6EC₅₀ = 0.023 μM) were favored over the free -NH (8, D6EC₅₀ = 0.13 μM) in terms of potency with little or no effect on the solubility. Increasing the sp3 content resulted in a significant increase in solubility and maintained the potency (9, D6EC₅₀ = 0.024 μM). Further, the replacement of the pyrazine with other amines such as methylated pyrazines (20 and 21), pyridines (26 and 27), and substituted anilines (22 and 23) was tolerated and led to the best compound in the series, i.e. 21 with excellent potency (D6EC₅₀ = 0.019 μM) and favorable ADME properties. Ortho-methylation of the aromatic core at C-8 (16) was favored over the C-6 position (15). Replacing the pyrazine headgroup with saturated rings (28, 29, 30, 31, and 32), C-6 substituted quinolines (34 and 35), scaffold modification with quinolinimine (38), isocryptolepines (44 and 45), and open chain analogs (41 and 43) were also successful and resulted in several compounds with low nanomolar to submicromolar potency across all three strains of P. falciparum. The general SAR trends observed for this series against P. falciparum D6 and a representative best compound with favorable ADME properties are summarized in Figure 5.

Figure 5.

SAR trends for P. falciparum D6 around 1 and representative best compound from the series (21).

A comparison of activity between P. falciparum D6, W2, and C235 (Figure 6) revealed most compounds tested had activities within 3-fold of the D6 potency value. Of those tested, 20% of the quinolines (14 of 69 compounds) were less potent against the W2 strain than D6, as was one of the quinolinimines (representing 25% of those tested) though this represents a small sample size and more data is required to understand the trend for this scaffold. Further, 4% of the quinolines (3 of 69 compounds) were less potent against the C235 strain than D6. This suggests a lack of sensitivity to chloroquine resistance for the quinolines presented herein. The open chain and isocryptolepine scaffolds all had potency values within 3-fold against each strain tested.

Figure 6.

Comparison of potency (compounds with EC₅₀ ≤ 0.5 μM shown) between (A) P. falciparum W2 vs D6 and (B) P. falciparum C235 vs D6. The blue line signifies a 3-fold change in potency between P. falciparum D6 and the strain in question; the black line signifies compounds that are equipotent with P. falciparum D6 and the strain in question.

In conclusion, utilizing a parasite-hopping approach, we have screened a previously reported library of T. brucei inhibitors along with newly synthesized compounds against T. cruzi, L. major, L. donovani, and three strains of P. falciparum. Assessment of these compounds produced 13 submicromolar active compounds against T. cruzi (EC₅₀ < 1 μM), five submicromolar active compounds against L. major (EC₅₀ < 1 μM), four submicromolar active compounds against L. donovani (EC₅₀ < 1 μM), and 42 low nanomolar compounds against P. falciparum D6 (EC₅₀ < 0.1 μM). Further, the cross-screening resulted in identification of several potent compounds with good selectivity index and improved ADME profile. Compounds 19 and 36 were tested in an efficacy model of L. donovani infection; both compounds failed to translate in vitro potency to in vivo activity. Work is ongoing to improve compound properties toward discovery of antiparasitic agents that are effective in animal efficacy studies.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.9b00453.

• Details of synthetic chemistry, biological assay protocols, biological data of all tested compounds, in vitro cell toxicity data, in vitro ADME properties, and other biological data (annotated with NEU registry numbers) (PDF)

Funding from the National Institutes of Health (R01AI082577, R56AI099476, R01AI124046, R21AI127594, R01AI126311 to M.P.P) is gratefully acknowledged.

The authors declare no competing financial interest.