Date Published: June 1, 2016
Publisher: Public Library of Science
Author(s): Kan Lin, Kathryn M. O’Brien, Carolina Trujillo, Ruojun Wang, Joshua B. Wallach, Dirk Schnappinger, Sabine Ehrt, Helena Ingrid Boshoff.
Mycobacterium tuberculosis (Mtb) must cope with exogenous oxidative stress imposed by the host. Unlike other antioxidant enzymes, Mtb’s thioredoxin reductase TrxB2 has been predicted to be essential not only to fight host defenses but also for in vitro growth. However, the specific physiological role of TrxB2 and its importance for Mtb pathogenesis remain undefined. Here we show that genetic inactivation of thioredoxin reductase perturbed several growth-essential processes, including sulfur and DNA metabolism and rapidly killed and lysed Mtb. Death was due to cidal thiol-specific oxidizing stress and prevented by a disulfide reductant. In contrast, thioredoxin reductase deficiency did not significantly increase susceptibility to oxidative and nitrosative stress. In vivo targeting TrxB2 eradicated Mtb during both acute and chronic phases of mouse infection. Deliberately leaky knockdown mutants identified the specificity of TrxB2 inhibitors and showed that partial inactivation of TrxB2 increased Mtb’s susceptibility to rifampicin. These studies reveal TrxB2 as essential thiol-reducing enzyme in Mtb in vitro and during infection, establish the value of targeting TrxB2, and provide tools to accelerate the development of TrxB2 inhibitors.
Endogenous oxidative stress represents an inevitable challenge for microbes adapted to an aerobic lifestyle . In addition, pathogens like Mycobacterium tuberculosis (Mtb) are confronted with exogenous oxidative stress imposed by the host . The production of antimicrobial oxidants is a critical host defense mechanism against Mtb [3,4]. Patients with germline mutations in phagocyte NADPH oxidase resulting in an impaired macrophage respiratory burst are predisposed to mycobacterial diseases including tuberculosis . Mice lacking inducible nitric oxide synthase succumb to Mtb infection much faster than their wild type littermates . The reactive oxygen and nitrogen species generated by these host enzymes can inactivate microbial iron-dependent enzymes, damage lipids and destroy DNA [1,6].
The paucity of targets that are both biologically validated and susceptible to inhibition by drug-like small molecules, i.e. “druggable”, is a major bottleneck in antimycobacterial drug development. Mtb’s thioredoxin reductase TrxB2 has recently been shown to be druggable, yet its biological evaluation has not advanced beyond the prediction of its essentiality for growth of Mtb on standard agar plates . Auranofin inactivates thioredoxin reductase in vitro but has multiple targets in bacteria, including in Mtb [18,30]. It was thus unknown how the specific inhibition of TrxB2 would affect Mtb in different environments including those encountered during acute and chronic infections. We addressed these questions using genetic strategies and found that inactivating TrxB2 quickly eradicated Mtb during the acute and, importantly, the chronic phase of mouse infection, validating TrxB2 as a valuable target for therapeutic intervention. Deliberately leaky TrxB2 knockdown mutants revealed that a TrxB2 inhibitor may synergize with rifampicin. Treatment combinations of rifampicin and a TrxB2 inhibitor could thus reduce the required drug dosage and limit the frequency of resistant mutants as shown for the synergistic action of carbapenems and rifampicin . We used a leaky TrxB2 mutant to determine the specificity of two TrxB2 inhibitors. The MIC of ebselen was not affected by partial TrxB2 depletion, suggesting that ebselen inhibits Mtb growth primarily through targets other than TrxB2. Ebselen has been shown to bind covalently to a cysteine residue located near the antigen 85 complex (Ag85C) active site and may thereby disrupt the biosynthesis of the mycobacterial cell envelope [32,33]. Auranofin was significantly more active against TrxB2-depleted Mtb than wild type indicating that it exerts its antimycobacterial activity at least partially through inhibiting TrxB2. However, auranofin exhibits a higher affinity for human thioredoxin reductase than for bacterial enzymes . Furthermore, auranofin, an FDA-approved anti-rheumatic drug, has immunosuppressive activities by inhibiting NF-κB signaling and decreasing the production of nitric oxide and pro-inflammatory cytokines, which are critical for anti-TB immune responses [35,36]. It also has anti-tumor activity through inhibition of proteasome-associated deubiquitinases [37–39]. The catalytic mechanisms of mammalian and bacterial thioredoxin reductases are significantly different and the crystal structure of TrxB2 has been solved [6,20]. It should thus be possible to identify inhibitors that are more specific for TrxB2 than auranofin. We expect the leaky TetON mutants we constructed for this study to facilitate the identification of such inhibitors.