Evaluating 5-Nitrothiazoles as Trypanocidal Agents

Abstract

The growth-inhibitory properties of a 5-nitrothiazole series were evaluated against Trypanosoma brucei. A subset of related compounds displayed the greatest potency toward the parasite while exhibiting little cytotoxic effect on mammalian cells, with this antiparasitic activity dependent on expression of a type I nitroreductase by the trypanosome. We conclude that the 5-nitrothiazole class of nitroheterocyclic drugs may represent a new lead in the treatment of human African trypanosomiasis.

TEXT

Spread via the blood-feeding habits of tsetse flies, parasites belonging to the Trypanosoma brucei complex are responsible for human African trypanosomiasis (HAT) (1). Drugs are the only option for combatting this infection, but their use is often problematic (2). One treatment that targets the cerebral stage of this disease is nifurtimox-eflornithine combination therapy (3, 4). In this medication, eflornithine acts as an inhibitor of ornithine decarboxylase, blocking polyamine biosynthesis (5, 6), while nifurtimox is converted to a toxic metabolite following activation by a type I nitroreductase (NTR) (7, 8). As type I NTRs are expressed by some unicellular eukaryotes but not by metazoan organisms, the bioreductive activity of this enzyme has been exploited to develop a series of novel antiparasitic agents that often exhibit little or no toxicity toward cultured mammalian cells (9, 10).

The 5-nitrothiazoles are a class of heterocyclic compounds of which niridazole and nitazoxanide display potent antimicrobial and anthelmintic activities (11, 12). The mode of action of these agents is unclear; both structures have been shown to inhibit key enzymes involved in energy metabolism (13, 14) and are able to function as prodrugs, undergoing reduction to form adduct-forming metabolites (15–17). To date, only niridazole and its derivatives have been screened for trypanocidal activity against T. brucei and, when combined with suramin, have cured mice of trypanosomiasis (18). However, concerns over its carcinogenic properties resulted in the suspension of trials using niridazole (19). Here, we assessed a 2-amide 5-nitrothiazole series for growth-inhibitory activity against bloodstream-form (BSF) T. brucei (Table 1). Out of the 15 compounds tested, 7 had no effect on trypanosomal growth at a concentration of 30 μM. For the remaining chemicals, detailed inhibition assays were conducted that generated dose-response curves from which 50% inhibitory concentrations (IC₅₀s) were determined (Table 1). For NT2, NT4, NT6, NT7, and NT11, appreciable trypanocidal activity (IC₅₀s, >10.0 μM) equivalent to the potency exhibited by nifurtimox was noted, with the other agents being less effective (IC₅₀s, ∼17 μM). Screening against two mammalian lines revealed that NT2, NT10, NT12, and NT15 displayed toxicity toward THP-1 or SK-N-SH cells (Table 1), with NT10 and NT12 showing growth-inhibitory effects against both lines. For the remaining agents, no growth-inhibitory effects at concentrations of up to 100 μM were observed.

TABLE 1.

Structure and growth-inhibitory properties of nitrothiazole compounds

^aAll compounds tested satisfy Lipinski's rule of 5 (see PubChem database [http://pubchem.ncbi.nlm.nih.gov/]). Susceptibility of parasites and mammalian cells to nitrothiazole compounds was assessed as previously described (7).

^bR₁ and R₂ represent substituent groups on the 2-amide 5-nitrothiazole core.

^cAverage IC₅₀s ± SD were calculated from dose-response curves performed in triplicate. TbNTR^ox, T. brucei cell line overexpressing the type I nitroreductase. The numbers in parentheses correspond to the fold difference in IC₅₀s of the TbNTR^ox, SK-N-SH, and THP-1 cell lines versus wild type.

Before mediating its trypanocidal effects, nifurtimox must undergo activation in a reaction catalyzed by a type I NTR (7). Using purified His-tagged T. brucei type I nitroreductase (TbNTR) (Fig. 1A), we evaluated whether the 2-amide 5-nitrothiazoles can serve as the substrates for this enzyme (Fig. 1B). Five compounds were shown to be good NTR substrates, generating a specific activity that was ∼3-fold greater than that noted for nifurtimox (Fig. 1B). Of these structures, NT2, NT4, NT6, and NT7 were related in that they contained a saturated unbranched hydrocarbon chain. However, the number of carbon atoms in this sequence and the associated increase in lipophilicity did not affect the specific activity displayed by TbNTR toward a given substrate. Of the remaining compounds, three yielded activities similar to that observed for nifurtimox, while the others were not metabolized by TbNTR at an appreciable rate under the conditions used here (Fig. 1B).

FIG 1.

Activity of TbNTR toward different nitrothiazoles. (A) Samples obtained during purification of recombinant TbNTR were analyzed by SDS-PAGE (10%) stained with Coomassie blue. Escherichia coli crude extract (lane 1) was loaded onto an Ni-nitrilotriacetic acid column and the flowthrough (lane 2) collected. The column was washed with 50 mM imidazole-containing (lane 3) and 100 mM imidazole-containing (lane 4) buffers. Recombinant protein was eluted in a buffer containing 500 mM imidazole with 0.5% Triton X-100 (lane 5). Markers (M) are in kilodaltons. The ∼30-kDa band corresponding to recombinant TbNTR is indicated. (B) Activity of purified recombinant TbNTR was assessed by using nitrothiazoles (NT1 to NT15) as the substrates (100 μM) at a fixed concentration of NADH (100 μM). Enzyme activity, expressed in nanomoles of NADH oxidized per minute per milligram TbNTR, was then calculated using an ε value of 6,220 M⁻¹ · cm⁻¹. Nifurtimox (Nfx) was used as the control, and enzyme activity was determined as previously described (7). The enzyme activity values are the means ± SD from 3 assays.

To investigate whether NTR plays a role in prodrug activation within the parasite itself, the susceptibility of BSF T. brucei engineered to overexpress this enzyme was evaluated (Table 1; Fig. 2) (8). Cells with elevated levels of TbNTR were up to 10-fold more sensitive than controls to NT2, NT4, NT6, or NT7. This effect was NTR specific, as recombinant and wild-type parasite lines displayed similar sensitivities to the nonnitroaromatic compound G418 (IC₅₀, ∼0.6 μM). When these studies were extended to test other trypanocidal nitrothiazoles, a lower (∼2-fold) or no difference in IC₅₀ was observed (Table 1; Fig. 2). This implies that for these less-effective trypanocidal compounds, NTR plays little or no role in the metabolism of these structures within the parasite itself.

FIG 2.

Susceptibility of bloodstream form T. brucei overexpressing TbNTR to nitrothiazoles. Dose-response curves of T. brucei (solid line) and parasites expressing an ectopic copy of TbNTR (dashed line) toward representative nitrothiazoles. The growth-inhibitory effect expressed as the IC₅₀ was determined. All data points are means ± SD from experiments performed in quadruplicate. Nifurtimox was used as the drug control.

By comparing the specific-activity values and growth-inhibitory effects of each compound, a number of structure-activity relationships (SARs) were identified. In contrast to their nonsubstituent counterparts, addition of a methyl or tert-butyl group at the 4 position on the thiazole ring generated compounds that were not TbNTR substrates and did not exhibit trypanocidal activities (e.g., NT2 versus NT3 and NT4 versus NT5). This lack of activity may be due to steric hindrance with the 4-alkyl side chain blocking the trypanosomal enzyme from gaining access to the adjacent 5-nitro grouping or may reflect an inductive effect with the alkyl substituent on the thiazole backbone, rendering nitroreduction energetically unfavorable. Extending the SAR studies to investigate grouping attached to the thiazole ring via a 2-amide linker revealed that compounds containing an unbranched, saturated hydrocarbon chain (NT2, NT4, NT6, or NT7) were efficiently metabolized by TbNTR, which translated to a trypanocidal effect equivalent to that of the reference nitrofuran. Encouragingly, these structures displayed little or no in vitro toxicity to mammalian cells, suggesting that they warrant in vivo analysis. Modification of this saturated linear hydrocarbon chain (incorporation of an unsaturated bond [NT8]) or an ether linkage [NT13], inclusion of halogen substituents [NT10 to NT12] or its replacement with a hydrogen atom [NT1], or a benzyl-containing grouping [NT9, NT15]) generated structures that displayed lower TbNTR activity and/or had reduced potency toward BSF trypanosomes. Presumably, such alterations to the saturated alkyl chain alter the affinity of these variants for the parasite oxidoreductase. As the broad-spectrum anti-infective agent nitazoxanide is structurally related to NT9 and NT15 (all contain a phenyl group attached to the amide linker), we predict that this particular antimicrobial agent is unlikely to function as an effective TbNTR substrate and/or display activity against BSF T. brucei. Intriguingly, despite being screened against a wide range of microbial infectious agents, including Trypanosoma cruzi and Leishmania, the potency of this particular nitrothiazole against T. brucei has not been reported.

There has been renewed interest in the use of nitroheterocyclic prodrugs for the treatment of trypanosomatid infections, with nifurtimox combined with eflornithine now being used to treat the form of HAT that is prevalent in West and Central Africa, while the nitroimidazole fexinidazole is under clinical evaluation against HAT, Chagas disease, and visceral leishmaniasis. In both cases, these nitroheterocyclics are converted to toxic metabolites by a type I NTR activity (7, 20). Here, we identified several trypanocidal nitrothiazoles, including some that are activated by the type I NTR, as being as potent as nifurtimox against BSF T. brucei. Promisingly, the most effective structures exhibited little or no toxicity to cultured mammalian cells, with trypanosomal expression of the type I NTR underlying their selectivity. As such, these compounds warrant further attention in terms of developing novel therapies targeting HAT and may potentially represent one component of a new combinatorial treatment against this disease.

Funding Statement

The FONDECYT grant was awarded to Benjamín Aguilera-Venegas to facilitate his visit to the Wilkinson lab in the United Kingdom.