Copper-Boosting Compounds: a Novel Concept for Antimycobacterial Drug Discovery

Abstract

We and others recently identified copper resistance as important for virulence of Mycobacterium tuberculosis. Here, we introduce a high-throughput screening assay for agents that induce a copper hypersensitivity phenotype in M. tuberculosis and demonstrate that such copper-boosting compounds are effective against replicating and nonreplicating M. tuberculosis strains.

TEXT

The finding that the innate immune system employs copper ions to eliminate bacterial infections and the establishment of copper resistance as a virulence factor of Mycobacterium tuberculosis (1–3) lead us to consider a drug screening system for compounds that induce copper hypersensitivity in M. tuberculosis.

In a first step, we identified Hartmans-de Bont (HdB) minimal medium (4) supplemented with 0.5% glucose as the carbon source and 0.02% tyloxapol as the anticlumping agent as the optimal medium for this drug screening effort. Direct bactericidal effects of copper ions against M. tuberculosis in HdB medium were observed at copper concentrations between 15 and 25 μM (Fig. 1A). This concentration range correlates well with physiologically achievable copper levels observed in humans (15 to 25 μM) (5). Optimal growth of M. tuberculosis in HdB medium was achieved at copper concentrations between 7.5 and 12.5 μM (Fig. 1A). To conduct the screen under biosafety level 2 (BSL2) conditions, we used M. tuberculosis mc²6230 (ΔRD1, ΔpanCD) (6), an attenuated mutant of M. tuberculosis H37Rv. The attenuated strain showed the same tolerance toward copper as the virulent strain (data not shown), and none of the deleted genes are known or suspected to influence copper resistance. In the final assay setup, selectivity for copper-boosting compounds was achieved by screening compounds simultaneously in the presence and absence of copper.

Fig 1.

Establishment of screening parameters and molecular probes. (A) Susceptibility of M. tuberculosis to copper (added as CuSO₄) in HdB medium. Copper-dependent activity of neocuproine (B) and bathocuproine (C) against M. tuberculosis in copper-free (gray line) or copper-supplemented (12.5 μM; black line) HdB medium. Molecular structures are depicted within the graphs. (D) Assay-specific Z′-factor of 0.89 was determined using untreated (control) and copper-neocuproine (Cu/Neo; 10 μM/3 μM)-treated cells. Copper only (Cu; 10 μM)- and neocuproine only (Neo; 3 μM)-treated cells were included as controls.

In the optimized assay format, all compounds were mixed with copper prior to the addition of medium and cells to enable the formation of potential copper complexes. The final cell density (optical density at 600 nm [OD₆₀₀]) was 0.02 in a total volume of 200 μl. The 96-well plates were incubated for 7 days at 37°C. Then, 40 μl alamarBlue dye mixture (50% alamarBlue, 5% [vol/vol] Tween 80) was added, and incubation was continued for another 24 h before plates were analyzed. Conversion of the indicator dye resazurin into resorufin with a fluorescence emission peak at 590 nm served as a marker for bacterial growth and viability (7–9).

Next, we looked for chemical probes to optimize our assay. We distinguish at least two functional classes of such compounds: (i) copper resistance pathway inhibitors which target specific proteins that mediate copper resistance and (ii) copper complexing agents that cross the mycobacterial outer membrane barrier and thereby possibly increase the intracellular copper content. Because both compound classes enhance the bactericidal properties of copper ions, we henceforth refer to them as copper-boosting compounds. While specific inhibitors of copper resistance proteins have yet to be identified, we reasoned that known copper complexing agents may provide an exploitable resource for assay validation purposes. Neocuproine and the structurally related compound bathocuproine represent copper complexing agents with opposing membrane permeability profiles (10). Our assay revealed that the membrane-permeable neocuproine exhibited growth-inhibiting activity against M. tuberculosis in a copper-dependent manner (Fig. 1B). In contrast, the membrane-impermeable derivative bathocuproine failed to inhibit growth of M. tuberculosis in the absence or presence of copper (Fig. 1C). Thus, neocuproine and bathocuproine provided excellent positive and negative controls that were used to determine the quality of the developed assay. The assay-specific Z′-factor (11) was determined using untreated and neocuproine-copper (3 μM/10 μM)-treated cells. Copper only (10 μM)- and neocuproine only (3 μM)-treated cells were included as controls to access edge and drift effects, signal uniformity, and overall assay performance. The assay was characterized by a Z′-factor of 0.89, which indicates excellent assay performance and reliability (Fig. 1D). Although neocuproine served as an important tool during assay development, an unfavorable cytotoxicity profile prevents its use in human therapy (10).

To test the M. tuberculosis-based assay under automated conditions on a high-throughput screening platform, we performed a limited pilot screen using a random selection of compounds from our in-house small molecule library (Chembridge, Inc.). As copper chelation was identified as a potential mechanism to enhance the antibacterial properties of copper ions against M. tuberculosis, we included 20 reported copper chelators in the pilot screen to increase the likelihood of identifying hits.

Our pilot screen identified the membrane-permeable bis-thiosemicarbazones ATSM and GTSM (12) as novel copper-boosting compounds (Fig. 2A). ATSM has been developed for tumor imaging and diagnosis and is currently in clinical trials (13–15). In the absence of copper, ATSM at concentrations of up to 10 μM had no apparent inhibitory effect on M. tuberculosis. However, with 10 μM copper, ATSM had a 50% inhibitory concentration (IC₅₀) of ∼0.6 μM (∼0.16 μg/ml) and an IC₉₀ of ∼2.5 μM (∼0.65 μg/ml). The anti-M. tuberculosis activity of ATSM was confirmed in detailed experiments using virulent M. tuberculosis H37Rv (Fig. 2B). GTSM, an ATSM analogue, already exhibited anti-M. tuberculosis activity in regular HdB medium. Importantly, copper supplementation enhanced this activity even further. The copper-dependent IC₅₀ for GTSM was ∼60 nM (∼0.014 μg/ml), and its IC₉₀ was experimentally determined as ∼300 nM (∼0.07 μg/ml) (Fig. 2C). Using a previously published in vitro model for latent M. tuberculosis (16), we found that GTSM, but not ATSM, killed nonreplicating M. tuberculosis at a concentration of 2.5 μM (∼0.58 μg/ml) in the presence of 10 μM copper (Fig. 2D). GTSM, tested at 0.3 and 2.5 μM, had no activity in the absence of copper (Fig. 2D).

Fig 2.

Copper-dependent activity of bis-thiosemicarbazones against M. tuberculosis. (A) Molecular structure of ATSM [diacetylbis(N(4)-methyl-3-thiosemicarbazone); MW, ∼260.38 g/mol] and GTSM [glyoxalbis(N(4)-methyl-3-thiosemicarbazone); MW, ∼232.33 g/mol]. Copper-dependent activity of ATSM (B) and GTSM (C) in the absence (gray lines) or presence (10 μM; black lines) of copper. (D) Activity against nonreplicating M. tuberculosis was evaluated 8 days after compound exposure by spotting 5 μl of each well on Middlebrook 7H10 medium plates. Growth was assessed after 16 days at 37°C. Rifampin (RIF) was included as the copper-independent inhibitor control (16).

Tolerability of ATSM and GTSM is implied by their use in various clinical settings and animal experiments (14, 17) and was confirmed by us on mouse peritoneal macrophages (Fig. 3), which were obtained according to published protocols (18). The in vitro therapeutic indices on peritoneal mouse macrophages were >20 for both compounds, indicating that such compounds could potentially advance into drug candidate status.

Fig 3.

Cytotoxicity of ATSM and GTSM. Mouse peritoneal macrophages were exposed to 10 μM ATSM or GTSM in 96-well plates. Viability was accessed 24 h posttreatment by alamarBlue (A) according to the recommendations of the manufacturer (AbD Serotec) and by light microscopy (B). Magnification, 400-fold.

In summary, we provide a drug screening assay that specifically distinguishes between copper-dependent and copper-independent anti-M. tuberculosis activities. We demonstrate that small molecules can induce copper hypersensitivity in M. tuberculosis and identified copper complexation as a potential mechanism by which this can be achieved. The discovery of ATSM and GTSM is the first step toward the development of research tools that may enable us to evaluate if copper resistance mechanisms of M. tuberculosis provide for an exploitable new drug target. As copper binding drugs are in clinical use to treat Wilson's disease (penicillamine) or alcohol dependence (disulfiram) and are investigated as anticancer and anti-HIV drugs (19–21), the here-described copper-dependent anti-M. tuberculosis activities of ATSM and GTSM may provide a novel opportunity to therapeutically exploit the bactericidal properties of copper ions.