Research ArticleTuberculosis

Indolcarboxamide Is a Preclinical Candidate for Treating Multidrug-Resistant Tuberculosis

See allHide authors and affiliations

Science Translational Medicine  04 Dec 2013:
Vol. 5, Issue 214, pp. 214ra168
DOI: 10.1126/scitranslmed.3007355


New chemotherapeutic compounds against multidrug-resistant Mycobacterium tuberculosis (Mtb) are urgently needed to combat drug resistance in tuberculosis (TB). We have identified and characterized the indolcarboxamides as a new class of antitubercular bactericidal agent. Genetic and lipid profiling studies identified the likely molecular target of indolcarboxamides as MmpL3, a transporter of trehalose monomycolate that is essential for mycobacterial cell wall biosynthesis. Two lead candidates, NITD-304 and NITD-349, showed potent activity against both drug-sensitive and multidrug-resistant clinical isolates of Mtb. Promising pharmacokinetic profiles of both compounds after oral dosing in several species enabled further evaluation for efficacy and safety. NITD-304 and NITD-349 were efficacious in treating both acute and chronic Mtb infections in mouse efficacy models. Furthermore, dosing of NITD-304 and NITD-349 for 2 weeks in exploratory rat toxicology studies revealed a promising safety margin. Finally, neither compound inhibited the activity of major cytochrome P-450 enzymes or the hERG (human ether-a-go-go related gene) channel. These results suggest that NITD-304 and NITD-349 should undergo further development as a potential treatment for multidrug-resistant TB.


Mycobacterium tuberculosis (Mtb) is the causative agent of tuberculosis (TB), a disease that causes 1.4 million to 2.0 million deaths per annum. In 2011 alone, more than 9 million new TB cases and 1.4 million deaths were reported due to TB (1). In the past decade, there has been an alarming increase in TB cases that are either multidrug-resistant (MDR), extensively drug-resistant (XDR), or totally drug-resistant (1). There is also a high burden of TB-HIV co-infection, and treatment of these cases using first-line TB drugs such as rifampicin is challenging because of drug-drug interactions with anti-HIV drugs (2). Hence, there is an urgent need to identify new compounds that are active against a wide variety of drug-resistant TB strains and also have a good safety profile and limited inhibition of major drug-metabolizing enzymes.

Recently, efforts from various groups have generated several new compounds that are active against TB, namely, PA-824, TMC207, OPC67683, BTZ043, and Q-203 (37). PA-824, OPC67683, and TMC207 (bedaquiline) have been shown to be effective in reducing the bacterial load of infected TB patients (8). For the first time in four decades, a new chemical entity, bedaquiline, has been approved by the U.S. Food and Drug Administration for the treatment of MDR TB patients as a part of combination therapy (9). Presently, there are several candidates and regimens in phase 2 and 3 clinical trials; many of them are repurposed classes of compounds ( (10). However, there is a worrying gap corresponding to advanced preclinical and phase 1 drug candidates that needs to be addressed to compensate for possible attrition and clinical drug resistance (8).

Here, we have identified and characterized indolcarboxamides as a new class of antitubercular bactericidal agents equally active against drug-sensitive and drug-resistant clinical TB isolates. The two advanced lead candidates, NITD-304 and NITD-349, showed favorable oral pharmacokinetic (PK) properties in rodents and dogs and were efficacious in mouse models of both acute and chronic Mtb infection. Further, in vitro and in vivo safety profiling including rat toxicology studies confirmed the therapeutic potential of NITD-304 and NITD-349.


Phenotypic screen and identification of indolcarboxamides

A whole-cell–based high-throughput screening (HTS) of nearly 2 million compounds from the Novartis collection was carried out against Mycobacterium bovis BCG (Bacille Calmette-Guérin) using ATP (adenosine 5′-triphosphate) as a surrogate marker for growth, resulting in ~6000 hits with >50% inhibition at 10 μM concentrations at a hit rate of 0.27% (fig. S1). Frequent hitters, scaffolds of known anti-TB compounds, compounds containing undesirable functional groups, and compounds that are toxic to mammalian cells [Huh7 or HepG2 cytotoxicity minimum inhibitory concentration (IC50) <10 μM] were deprioritized, resulting in ~3000 confirmed Mtb cell active hits with an MIC50 <10 μM. The whole-cell screening efforts identified several series of compounds, one of which belonged to indole-2-carboxamides, that is, N-cyclohexyl-4,6-dimethyl-1H-indole-2-carboxamide. Indolcarboxamides are a valuable series of compounds having relatively low molecular weight and are simple to synthesize. Recently, two groups have reported on the anti-TB activity of indole-2-carboxamide (11, 12). We have synthesized several indolcarboxamides to optimize activity against Mtb and improve metabolic stability (13).

Metabolically stable indole-2-carboxamides with dichloro substitutions on the indole ring, namely, cis (NITD-218) and trans (NITD-219) isomers, have similar Mtb activities (Fig. 1A). The difluoro substitution on the indole ring of trans isomer led to NITD-324, which retained moderate potency (MIC50, 0.16 μM). Further introduction of the gem-dimethyl group on the four positions of the cyclohexane ring (NITD-304 and NITD-349) resulted in significant improvement in Mtb potency (MIC50, 15 to 23 nM) and also gem-dimethyl group removed cis and trans isomerism. A detailed structure-activity relationship and structure-property relationship study of indole-2-carboxamides has been described by Kondreddi et al. (13). Here, NITD-218, NITD-219, NITD-324, NITD-304, and NITD-349 (Fig. 1A) were evaluated for in vivo activity and used to elucidate the mechanism of action against Mtb.

Fig. 1. Chemical structures and in vitro activity of indolcarboxamides.

(A) Chemical structures of indolcarboxamide analogs used in this study. MIC50 values (μM) for the indolcarboxamide analogs against Mtb are shown in parentheses. (B) Kill kinetic analysis of NITD-304 and isoniazid (INH) against the H37Rv strain of Mtb. (C) Intracellular activity of NITD-304 against Mtb in activated THP-1 macrophages. IC90 and IC99 values are indicated by stippled lines. Both kill kinetic analysis and intramacrophage analysis were performed in duplicate, and results are shown as mean values with SEs.

Microbiological profiling of indolcarboxamides

Indolcarboxamides proved to be highly potent anti-mycobacterial compounds with MIC50 of 15 and 23 nM for NITD-304 and NITD-349, respectively, against virulent Mtb H37Rv (Fig. 1A and table S1). These compounds are at least 10 times more potent than isoniazid (MIC50, 0.33 μM) and also PA-824 (MIC50, 0.4 μM) (14), but are comparable to the activity of TMC207 (MIC50, 50 nM) (3). The lead candidates NITD-304 and NITD-349 showed bactericidal activity against in vitro replicating Mtb and also were active against intramacrophage Mtb (Fig. 1, B and C, and fig. S2, A and B). Kill kinetic analysis of these compounds showed both concentration- and time-dependent killing of Mtb cells with 3- to 4-log colony-forming unit (CFU) reduction within 3 days of treatment. The cidal activity profile of NITD-304 was similar to that of isoniazid for which rapid killing was noticed at concentrations greater than 0.2 μM. Indolcarboxamides showed a narrow spectrum of antibacterial activity, with no activity against tested Gram-positive and Gram-negative bacterial species, including selected efflux pump–defective mutants (table S1). Indolcarboxamides were also shown to be active against both slow-growing (Mtb, M. bovis BCG) and fast-growing (Mycobacterium smegmatis) mycobacterial species (table S1), indicating that the molecular target of these compounds may be restricted to mycobacterium. In the Wayne model of nonreplicating persistent Mtb, indolcarboxamides were less effective (NITD-304 >27 μM), which implies that these molecules block an essential step in active metabolism of Mtb, similar to isoniazid.

Indolcarboxamides are active against MDR TB clinical isolates

Indolcarboxamides were tested against nine different isolates of MDR TB strains that are distributed into six prominent clusters based on synonymous single-nucleotide polymorphisms representing the global Mtb populations (15). The minimum concentration required to inhibit 99% growth of the diverse drug-resistant clinical isolates by NITD-218, NITD-219, NITD-304, and NITD-349 was in a similar range to that of wild-type Mtb H37Rv (Table 1), indicating a new mode of action for this series of compounds. The MIC activity of NITD-304 and NITD-349 against various MDR Mtb strains ranged from <0.04 to 0.08 μM, indicating their potential for use against MDR Mtb strains.

Table 1. Activity of indolcarboxamide analogs against a panel of drug-resistant TB clinical isolates.

MIC99 concentrations were determined by the pellet formation method. I, isoniazid; M, moxifloxacin; PZA, pyrazinamide; R, rifampicin; S, streptomycin.

View this table:

NITD-304 and NITD-349 are orally bioavailable in preclinical species

NITD-304 and NITD-349 showed high permeability and moderate in vitro metabolic clearance in mouse and human hepatic microsomes (Table 3). NITD-304 and NITD-349 were subjected to in vivo PK studies in mouse. Intravenous administration of NITD-304 and NITD-349 (5 mg/kg) to mice displayed a large volume of distribution (Vss) (4.1 and 2.2 liters/kg, respectively), moderate total systemic clearance, and long elimination half-life (Table 2). Taking advantage of the high lipophilic nature of NITD-304 and NITD-349, a lipid-based microemulsion preconcentrate (MEPC) solution formulation was developed for oral treatment (table S2). Oral administration of NITD-304 at 25 and 75 mg/kg resulted in rapid absorption (Tmax 0.75 and 1.5 hours) and good oral bioavailability (53 and 37%). A further threefold increase in dosage resulted in nearly a twofold higher Cmax (maximum plasma concentration) and exposure [the area under the plasma drug concentration–time curve (AUC) for 24 hours after dosing]. The concentrations in mice plasma were well above those needed for in vitro potency. Likewise, NITD-349 also displayed a good exposure and oral bioavailability in mouse (Table 2). In agreement with their large Vss, both NITD-304 and NITD-349 showed good partition in the lungs, at the site of Mtb infection.

Table 2. In vivo PK analysis of NITD-304 and NITD-349 in mouse, rat, and dog.

Cmax, maximum plasma concentration; Tmax, time of peak plasma concentration; AUC, area under the curve (t = 0 to 24 hours); F, absolute oral bioavailability.

View this table:
Table 3. In vitro PK and safety profiling data for NITD-304 and NITD-349.

M/R/D/H, mouse/rat/dog/human; PAMPA, parallel artificial membrane permeability assay; SI, selectivity index.

View this table:

NITD-304 and NITD-349 were administered to rat and dog for in vivo PK analysis to enable in vivo safety evaluation and to calculate human dose projections. Both compounds were dosed orally in MEPC formulation. In rats, NITD-304 and NITD-349 exhibited large volume of distribution, low total systemic clearance, and long elimination half-life (~6 hours) (Table 2). Oral dosing of rats with NITD-304 showed higher exposure and Cmax compared to mice (Table 2). Similar observations were made with NITD-349. In dogs, NITD-304 and NITD-349 exhibited large volume of distribution, moderate total systemic clearance, and a long elimination half-life (13 to 31 hours) (Table 2). The Cmax and AUC achieved with NITD-304 and NITD-349 were overproportional between the doses of 3 and 10 mg/kg (Table 2). Overall, these results indicated that NITD-304 and NITD-349 have adequate PK properties in preclinical species that enabled in vivo efficacy and safety studies.

Acute and established in vivo mouse efficacy studies

In the acute murine efficacy model, animals were orally treated with a daily dose of NITD-304 or NITD-349 for 4 weeks, 1 week after intranasal infection with low-dose Mtb (~103 CFU). All animals tolerated 1 month of daily dosing with NITD-304 or NITD-349 and showed a dose- and time-dependent in vivo activity (Fig. 2, A and B, and table S3). Treatment of mice with NITD-304 at doses of 12.5 and 50 mg/kg resulted in 1.24- and 3.82-log CFU reduction in lung tissue, respectively, compared to untreated control animals. Similarly, NITD-349 (12.5 and 50 mg/kg) resulted in 0.9- and 3.4-log CFU reduction in lung tissue. The minimum effective dose (MED) is defined as the lowest dose capable of preventing significant bacterial multiplication compared to lung bacterial load before treatment initiation. The MEDs for NITD-304 and NITD-349 were 37.5 and 25 mg/kg, respectively. The Cmax and exposure achieved in plasma of treated animals at the MED dose for NITD-304 were 2 μM and 19.5 μM/hour, respectively. Similarly, for NITD-349, the Cmax and exposure were 2.4 μM and 25.7 μM/hour, respectively.

Fig. 2. In vivo efficacy of indolcarboxamides in mice.

(A and B) Determination of the MED for NITD-304 (A) and NITD-349 (B) given orally in a mouse model of acute Mtb infection. (C) Efficacy of NITD-304 and NITD-349 compared with ethambutol and rifampicin (RIF) in a mouse model of established Mtb infection.

In vivo efficacy of NITD-304 and NITD-349 was further assessed in an established infection mouse model in which animals were treated 1 month after low-dose intranasal Mtb infection with the compound (25 or 100 mg/kg, daily) (Fig. 2C and table S3). Two weeks of treatment with rifampicin (10 mg/kg), ethambutol, NITD-304, or NITD-349 at 100 mg/kg each reduced the bacterial burden by 1.29, 0.75, 1.11, and 1.10 logs, respectively. Thus, after 2 weeks of treatment, the efficacy of NITD-304 and NITD-349 was comparable to the first-line TB drug rifampicin and was better than ethambutol (table S3). Four weeks of treatment at 100 mg/kg with NITD-304 and NITD-349 resulted in 2.14- and 2.38-log CFU reductions, respectively. Therefore, results in acute and chronic mouse efficacy models suggest that NITD-304 and NITD-349 have potent in vivo activity against Mtb (table S3), justifying further preclinical safety characterization.

Preclinical safety profiling of NITD-304 and NITD-349

Considering the long duration of TB treatment, any cytotoxicity, cardiotoxicity, or mutagenicity liabilities would preclude progression of the compound to clinical testing. In cytotoxicity assays, both compounds showed no activity (CC50 >20 μM) and a selectivity index of >1000 (the ratio of CC50/MIC50) (Table 3). Multiple anti-Mtb drug candidates (for example, TMC207, moxifloxacin, and OPC67683) have potential cardiotoxicity liabilities due to hERG (human ether-a-go-go related gene) channel inhibition (8). For both indolcarboxamides, hERG binding and patch clamp IC50 were >30 μM, consistent with a low risk of cardiotoxicity (Table 3). The mini-AMES analysis of NITD-304 and NITD-349 did not show intrinsic mutagenic potential in the absence or presence of rat liver S9 mix for metabolic activation. Further, in vitro toxicity assessment with nearly 40 biochemical assays including a panel of human G protein (heterotrimeric guanine nucleotide–binding protein)–coupled receptors, phosphodiesterases, proteases, and ion channels showed no inhibition (table S4).

In vivo mouse efficacy studies established that daily treatment of NITD-304 or NITD-349 (100 mg/kg) for 1 month was well tolerated. The histopathological analysis of mouse tissues after 1 month of treatment showed no evidence of treatment-associated renal or hepatic toxicity. Further, exploratory 2-week rat toxicology studies were performed with NITD-304 and NITD-349 at a daily dosing of 50 and 300 mg/kg. All rats were monitored for changes in body weight, food consumption, and peripheral blood micronucleus activity, and clinical and histopathological analyses were undertaken. All rats tolerated oral administration of NITD-304 or NITD-349 (50 and 300 mg/kg) daily for 14 days with no mortality or morbidity. In rats, NITD-304 doses of 50 and 300 mg/kg resulted in 12- and 15-fold higher exposure in plasma, respectively, compared to the exposure achieved in mouse at a MED of 37.5 mg/kg. Likewise, NITD-349 doses of 50 and 300 mg/kg resulted in 19- to 23-fold higher plasma exposure than a MED in mouse of 25 mg/kg.

Indolcarboxamides induce the iniB transcriptional response

To understand the mechanism of action of hits from the cell-based phenotypic screen, a luciferase reporter cell-based assay driven by the iniB (isoniazid inducible gene) promoter was developed. The iniB gene is part of the iniBAC operon, which encodes genes that are induced specifically by a broad range of antibiotics that inhibit cell wall biosynthesis (16, 17). The M. bovis BCG iniBlucA-E luciferase reporter driven by the iniB promoter was constructed and validated with known TB drugs. NITD-304, NITD-349, and NITD-219 induced the transcription of iniB in the same way as other mycobacterial cell wall inhibitors (Fig. 3A). Both NITD-304 and NITD-349 induced the luciferase reporter at concentrations five times lower than NITD-219, corresponding to their cellular potency, suggesting that indolcarboxamides potentially may inhibit mycobacterial cell wall biosynthesis.

Fig. 3. Mechanism of action of indolcarboxamides.

(A) Indolcarboxamides induced the iniB transcriptional response, which was measured by luciferase reporter activity driven by the iniBAC promoter in M. bovis. (B and C) Mycolate lipid profiles of Mtb after drug exposure. Mtb MIC50 values for NITD-219, ethambutol, and isoniazid are 0.14, 1.5, and 0.5 μM, respectively. The concentration and fold MIC values used in the experiments are shown. (B) Thin-layer chromatographic profiles of mycolic acid methyl esters (MAMEs) isolated from bacterial cell pellets after removal of free lipids. (C) Lipid profiles of CHCl3-CH3OH extracted lipids in the cell pellet showing TDM and TMM after exposure to the indolcarboxamide NITD-219, ethambutol, or isoniazid. (D) Quantitation of TDM and TMM concentrations in lipid extracts plotted as a percentage of those in untreated controls.

Indolcarboxamides block attachment of mycolates to the bacterial cell wall

To narrow down the molecular pathway that could be affected by indolcarboxamide in cell wall biosynthesis, we explored the effect of indolcarboxamides on mycolic acid biosynthesis using [14C]acetate metabolic labeling. Treatment of cells with 7× or 15× MIC of NITD-219 and NITD-304, respectively, did not affect the concentrations of total mycolic acid pools (fig. S3A). Under similar conditions, PA-824 treatment resulted in loss of ketomycolates and accumulation of hydroxymycolates, as reported earlier (7). However, after the removal of noncovalently associated lipids by 2:1 chloroform/methanol extraction, saponification of the covalently attached cell wall lipids revealed that all three classes of arabinogalactan-anchored mycolates consisting of α-mycolates, ketomycolates, and methoxymycolates were depleted by a 1.4× MIC of NITD-219 (Fig. 3B). In contrast, isoniazid, a known mycolic acid biosynthesis inhibitor, resulted in complete loss of arabinogalactan-anchored mycolates (Fig. 3B).

Trehalose monomycolate (TMM) and trehalose dimycolate (TDM) play an important role during mycolation of the cell wall (18). Hence, we analyzed the effect of indolcarboxamides on TMM and TDM. As reported earlier, TDM and TMM concentrations in the extractable lipid fraction accumulated after treatment with ethambutol (Fig. 3C) (19) because of inhibition of arabinogalactan biosynthesis by ethambutol. In contrast, NITD-219 treatment resulted in concentration-dependent loss of TDM with a concomitant accumulation of TMM (Fig. 3C). As expected, the spots corresponding to TMM and TDM were absent in isoniazid-treated cells. Densitometric analysis of the image further established the concentration-dependent accumulation of mycolates in TMM and the decrease of mycolates in TDM. Treatment of Mtb cells with 9× MIC of NITD-219 resulted in an accumulation of nearly 350% more TMM (Fig. 3D). These results showed that indolcarboxamides inhibit attachment of mycolates to the cell wall with a concomitant accumulation of TMM.

Indolcarboxamides target MmpL3, an essential TMM membrane transporter

To elucidate the genetic basis of the mechanism of action of indolcarboxamides, we isolated and characterized spontaneously resistant Mtb mutant strains with resistance against NITD-218 and NITD-324. The observed frequency of resistance in the H37Rv Mtb strain against a 10× MIC of NITD-218 and NITD-324 was 1 × 10−8, which was nearly two log orders lower than that for isoniazid (1 × 10−6) and similar to that for rifampicin (3 × 10−8). Three independent Mtb mutant strains resistant against NITD-218 and NITD-324 were at least 4- to 10-fold more resistant than the parental H37Rv strain and also were resistant to all substituted indolcarboxamide analogs tested (Table 4). All of these mutants were equally sensitive to streptomycin, isoniazid, and other tested anti-TB drugs, suggesting a unique mechanism of action. Further lipid profiling with the mutant 218-B Mtb strain revealed that there were no significant changes in cell wall–bound mycolates and extractable TDM and TMM in the presence of a 4× MIC of NITD-219 (fig. S3, B and C). However, using the wild-type H37Rv strain, a 1.4× MIC of NITD-219 resulted in rapid depletion of cell wall–bound mycolates and accumulation of TMM (Fig. 3, B and C).

Table 4. Cross-resistance and whole-genome sequencing analysis of indolcarboxamide resistance mutants of Mtb.

TM, transmembrane region.

View this table:

Whole-genome sequencing of six mutant strains identified a single-nucleotide polymorphism in the mmpL3 gene (encoding the mycobacterial membrane protein large) in all mutants compared to the parental strain Mtb H37Rv. However, in the Mtb-218-A mutant strain, two adjacent missense mutations encoding V683G and V684G mapped to mmpL3 (Table 4). MmpL3 is an essential membrane protein belonging to the resistance, nodulation, and division (RND) family, with 11 homologs present in H37Rv (20). MmpL3 is ~950 amino acids in length with 11 probable transmembrane segments. All indolcarboxamide resistance mutations including L189, G253, T311, S591, V683, and V684 have been mapped to the transmembrane regions of MmpL3 (figs. S4 and S5 and Table 4). MmpL3 is involved in transport of the essential macromolecule TMM that is required for cell wall biosynthesis (21, 22). The fact that mutations in MmpL3 confer resistance to indolcarboxamides raises the possibility that besides exporting TMM, MmpL3 may also act as a drug efflux pump. Nevertheless, sensitivity of Mtb to NITD-304 and NITD-349 was unaltered in the presence of the broadly active efflux pump inhibitors reserpine and verapamil, suggesting that mutations in the mmpL3 gene did not cause resistance through drug efflux (table S1).


The 4,6-dichloro-N-(4,4-dimethylcyclohexyl)-1H-indole-2-carboxamide NITD-304 and the 4,6-difluoro analog NITD-349 are promising compounds with activity against Mtb and are superior to many existing TB drugs. In vitro bactericidal activity of indolcarboxamides is comparable to that for the frontline TB drug isoniazid. NITD-304 and NITD-349 showed potent activity against both drug-sensitive and MDR clinical Mtb isolates. Despite their low aqueous solubility and high lipophilicity, both compounds displayed favorable PK properties after oral administration in mouse, rat, and dog. NITD-304 and NITD-349 were efficacious in mouse models of both acute and established infection. In the established infection model, after 2 weeks of drug treatment, the efficacy of NITD-304 and NITD-349 was comparable to that for the first-line TB drug rifampicin and was better than that for ethambutol. In addition, neither compound showed a risk for hERG-mediated cardiotoxicity. Further, in vitro and in vivo safety assessment of both candidate compounds including 2 weeks of exploratory toxicology studies in rat demonstrated that NITD-304 and NITD-349 are potential preclinical candidates.

TB is the leading cause of death among HIV-infected people, and the treatment of HIV-TB patients is complex because of drug-drug interactions between antiretroviral (ARV) and anti-TB drugs. Because any new TB drug has to be clinically administered in combination with other antitubercular or ARV medications, the absence of drug-drug interactions is critical for clinical development (8). NITD-304 and NITD-349 did not inhibit the major CYP450 isoenzyme 3A4 in both reversible and time-dependent inhibition assays and also did not induce hPXR (human pregnane X receptor) activation (Table 3). Metabolite analysis of unlabeled NITD-304 in rat, dog, and human liver microsomes showed no major breakdown of the compound in any of these three species. These results demonstrate that NITD-304 and NITD-349 neither inhibit nor stimulate CYP enzymes, suggesting a low potential for drug-drug interactions.

Mycobacterial lipid profiling and whole-genome sequencing of indolcarboxamide-resistant mutant strains suggested that indolcarboxamides target MmpL3, an essential membrane-bound protein involved in the transport of TMM. MmpL proteins belong to the RND family of membrane proteins, which are involved in the transport of diverse ligands (2224). The Mtb genome encodes 13 mmpL homologs, of which mmpL3 is essential for mycobacterial survival (20). The topology of the MmpL3 protein, predicted using the transmembrane hidden Markov model (25), suggests that it has 11 transmembrane helices with a cytoplasmic N-terminal region and a long C-terminal extracellular region with three soluble domains. All mutations conferring resistance to the indolcarboxamides have been mapped to the transmembrane regions (TM2, TM3, TM5, TM8, and TM10) of MmpL3 (figs. S4 and S5 and Table 4). The fact that all the mutated residues are located in the transmembrane domains of MmpL3 is not surprising given the hydrophobic nature of the indolcarboxamides, as well as the physiological substrate TMM. In the absence of a good structural homology model for MmpL3, it is challenging to determine how these six mutations mediate resistance to indolcarboxamides. Further, structure-activity relationship analysis of indole-2-carboxamides revealed a direct correlation between potency and clogP of indolcarboxamide analogs (13). Hence, we propose that indolcarboxamides bind to MmpL3 and inhibit the essential TMM transport function resulting in reduction of mycolate attachment to the cell wall, thus weakening the mycobacterial cell wall leading to cell lysis (fig. S5). The mmpL3 gene is highly conserved in each of the mycobacterial genomes analyzed including Mycobacterium leprae. M. leprae is the causative agent of leprosy, and it has undergone a high degree of genetic decay (26). All of the residues involved in indolcarboxamide resistance, that is, L189, G253, T311, S591, V683, and V684, are highly conserved in the MmpL3 of M. leprae (fig. S5), suggesting the potential use of indolcarboxamides against leprosy.

MmpL3 seems to be a promiscuous target for many potential TB compounds (27), namely, SQ109 (22), AU1235 (21), BM212 (28), benzimidazole (29), THPP, spiro analogs (30), and indole-2-carboxamides (12). These diverse chemical entities have a broad range of anti-Mtb activity, and in vivo efficacy in a mouse model of TB has been demonstrated for SQ109, BM212, THPP, and spiro analog (fig. S6). Although there are many new chemical entities known to target the MmpL3 of Mtb, biological validation with potent in vivo efficacy, good oral bioavailability, and safety is often lacking. It is intriguing that inhibitors of MmpL3 have such diverse chemical structures and mutations because most of them mapped to the transmembrane regions of the protein. The obvious common feature of all the compounds is that they are highly lipophilic (clogP >4.2) and either neutral or basic. Nearly half of the reported MmpL3 inhibitors contain lipophilic non-aromatic substituents such as adamantine or cycloalkyl moieties including NITD-304 (fig. S6). These findings have led us to hypothesize that the ligand-binding portion of MmpL3 is a large hydrophobic pocket that accommodates neutral or basic lipophilic compounds; however, this requires further experimental validation. Hence, designing sufficiently potent indolcarboxamides in the optimal lipophilicity range for clinical development may be challenging.

Notably, the recently approved candidate drug for MDR TB treatment bedaquiline has a clogP of 6.41 with a molecular weight of 555 daltons. Likewise, other promising candidates in clinical development have high clogP such as OPC67683 (clogP 5.2) and SQ109 (clogP 5.8). SQ109, also known to target MmpL3, showed poor oral bioavailability (4% in mouse at an oral dose of 25 mg/kg) and is currently being evaluated in a phase 2A early bactericidal activity study (31, 32). The oral bioavailability for NITD-304 and NITD-349 in a MEPC formulation ranged from 15 to 76% and from 38 to 100%, respectively, in three preclinical species. In rats, NITD-304 or NITD-349 dosing of 50 mg/kg resulted in a ~12- to 23-fold higher exposure compared to the exposure achieved with a MED in mice. Allometric scaling principles have been applied for NITD-304 and NITD-349 using PK parameters from three preclinical species (mouse, rat, and dog) to predict human PK (33, 34). Furthermore, using efficacious exposure at the MED dose in mice, the predicted preliminary efficacious dose for NITD-304 and NITD-349 in human is calculated to be 340 to 370 mg/day. Thus, despite a high clogP and low solubility, NITD-304 and NITD-349 reached adequate oral exposure with a favorable predicted human efficacious dose of ~5 mg/kg per day.

NITD-304 and NITD-349 are promising preclinical candidates for treating TB with the potential to fill the gap in the global TB drug discovery portfolio. The prospect of combining indolcarboxamides with existing TB drugs or candidates offers hope for drug regimens that may be better tolerated, require shorter treatment duration, and have fewer drug-drug interactions compared to existing regimens. A goal of the TB Alliance is to establish new drug regimens containing the best combination of new drugs to shorten TB therapy duration, as well as to combat MDR and XDR TB (35, 36). Further evaluation of NITD-304 and NITD-349 is warranted. The indolcarboxamides may ultimately be useful for treating TB in combination either with conventional TB drugs or with promising candidates also currently under development.


Strains and growth conditions

M. bovis BCG Pasteur [American Type Culture Collection (ATCC) 35734], M. smegmatis MC2 155, Mtb H37Rv (ATCC 27294), and derivative strains were maintained in Middlebrook 7H9 Broth medium supplemented with 0.05% Tween 80 and 10% ADS (albumin dextrose saline) supplement.

High-throughput cell-based screen, MIC determination, and kill kinetic analysis

HTS of Novartis compound collection was carried out with M. bovis BCG strain, and MIC50 against Mtb H37Rv was determined as previously described (37). For determining growth inhibition against nine diverse MDR Mtb clinical isolates (15), pellet formation method was used (14). MIC is defined as the minimum concentration of the drug required to inhibit 50% of H37Rv growth or 99% of growth in MDR clinical isolates after 5 or 10 days of incubation, respectively. Five milliliters of H37Rv culture (1 × 107 CFU/ml) was incubated with varying concentrations of compounds for 6 days at 37°C, and an aliquot of culture was plated onto Middlebrook 7H11 agar plates; the CFU were enumerated after incubating plates for 3 weeks in a 37°C incubator.

Generation of indolcarboxamide-resistant mutants and whole-genome sequencing

Mtb H37Rv strain was plated on 7H11 agar plates containing 10× and 20× MIC50 of the test compounds. After 3 weeks of incubation at 37°C, single isolated colonies were propagated in drug-free 7H9 broth. The resistance phenotype to the indolcarboxamides was confirmed by testing MIC50 values against NITD-218, -219, -324, -304, and -349. Whole-genome sequencing of the spontaneous resistant mutants was carried out as described earlier (37).

PK properties and oral bioavailability

CD-1 female mice, Wistar rats, and Beagle dogs were used for PK studies. All procedures involving mice were reviewed and approved by the Institutional Animal Care and Use Committee of the Novartis Institute for Tropical Diseases. NITD-304 and NITD-349 were formulated in a suitable formulation as described in table S2. The plasma concentrations of compounds at various time points were determined by liquid chromatography–mass spectrometry analysis (Supplementary Methods). PK parameters were determined with Watson LIMS by noncompartmental analysis.

Efficacy studies in mouse models of acute and established infection

Acute in vivo efficacy studies were carried out as described previously (37). For the established efficacy model, BALB/c mice were infected intranasally with 103 Mtb H37Rv; after 4 weeks of infection, animals were orally treated for 4 weeks. Rifampicin (10 mg/kg) and ethambutol (100 mg/kg) were used as controls. Bacterial load in lungs (mean ± SD from six mice per group and per time point) was analyzed at 2 and 4 weeks after treatment by enumerating CFU. Statistical evaluation was done with one-way analysis of variance (ANOVA), and the data were analyzed with Tukey’s multiple comparison test. Statistical significance was accepted with P values <0.05.



Fig. S1. Phenotypic HTS cascade to identify the indolcarboxamide series.

Fig. S2. Anti-TB activity of NITD-349.

Fig. S3. Effect of indolcarboxamides on mycolic acid profiles in Mtb H37Rv and indolcarboxamide-resistant mutants.

Fig. S4. Multiple sequence alignment of MmpL3 protein from Mtb, M. bovis BCG, and M. leprae.

Fig. S5. A proposed mechanism of action for indolcarboxamides.

Fig. S6. Summary of all reported chemical entities showing mutations in MmpL3.

Table S1. Activity of indolcarboxamides against various mycobacterial species and broad-spectrum Gram-positive and Gram-negative bacteria.

Table S2. Summary of formulations used for NITD-304 and NITD-349 during in vivo mouse efficacy, PK, and toxicology studies.

Table S3. Summary of lung Mtb CFU reduction in efficacy studies of acute and established infection in mice.

Table S4. Effect of NITD-304 and NITD-349 on ligand binding to a panel of recombinant receptors.

References (3850)


  1. Acknowledgments: We thank B. Robertson, Imperial College London, for luciferase plasmid constructs; M. Cynamon for mouse efficacy studies; B. Kreisworth for TB clinical isolates; and G. De Pascale, M. Sachdeva, and J. Leeds for broad-spectrum antibacterial testing. We thank other colleagues from Novartis Institute for Tropical Diseases (NITD) and Novartis Institutes for BioMedical Research, Basel and Cambridge, for their support. Funding: This work was funded and supported by NITD and Genomics Institute of the Novartis Research Foundation. S.K.K. is a Novartis-funded postdoctoral research fellow. Author contributions: U.H.M., S.P.S.R., L.R.C., P.B., B.H.T., S.H.N., P.T., S.H.L., B.H.L., Z.C., A.K.C., K.P., K.G., K.K., and D.B. performed/analyzed mycobacterial cell-based assays; U.H.M., S.P.S.R., B.H.T., W.S.B., and J.W. performed/analyzed whole-genome sequencing; U.H.M. and S.P.S.R. performed/analyzed lipid profiling; U.H.M., S.P.S.R., and S.K.K. performed/analyzed cell-based reporter assay; S.B.L., A.G., S.B., L.Z., G.F., F.B., and V.D. performed/analyzed animal PK studies; M.W. analyzed rat toxicology studies; U.H.M., S.P.S.R., M.H., M.N., S.R., and P.G.S. performed/analyzed mouse efficacy studies; S.P.S.R., N.L.M., K.K., A.K.C., P.B., K.P., R.G., T.D., and J.J. performed/analyzed HTS data; R.R.K., J.J., A.K.C., and P.W.S. synthesized indolcarboxamide analogs; U.H.M., S.P.S.R., L.R.C., P.B., M.W., F.B., V.D., R.G., T.D., P.W.S., and T.T.D. supervised and directed the work; U.H.M., S.P.S.R., and S.B.L. wrote the paper. All authors discussed the results and commented on the manuscript. Competing interests: R.R.K., J.J., and P.W.S. are named as inventors on patent applications related to this work: patent #PCT/IB2013/058318, “Indolcarboxamide derivatives and uses thereof.” Data and materials availability: All requests for compounds are subject to a material transfer agreement.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article