• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Oct 2, 2012; 109(40): 16004–16011.
Published online Sep 10, 2012. doi:  10.1073/pnas.1214188109
PMCID: PMC3479555
Inaugural Article
Medical Sciences

Nonsteroidal anti-inflammatory drug sensitizes Mycobacterium tuberculosis to endogenous and exogenous antimicrobials


Existing drugs are slow to eradicate Mycobacterium tuberculosis (Mtb) in patients and have failed to control tuberculosis globally. One reason may be that host conditions impair Mtb’s replication, reducing its sensitivity to most antiinfectives. We devised a high-throughput screen for compounds that kill Mtb when its replication has been halted by reactive nitrogen intermediates (RNIs), acid, hypoxia, and a fatty acid carbon source. At concentrations routinely achieved in human blood, oxyphenbutazone (OPB), an inexpensive anti-inflammatory drug, was selectively mycobactericidal to nonreplicating (NR) Mtb. Its cidal activity depended on mild acid and was augmented by RNIs and fatty acid. Acid and RNIs fostered OPB’s 4-hydroxylation. The resultant 4-butyl-4-hydroxy-1-(4-hydroxyphenyl)-2-phenylpyrazolidine-3,5-dione (4-OH-OPB) killed both replicating and NR Mtb, including Mtb resistant to standard drugs. 4-OH-OPB depleted flavins and formed covalent adducts with N-acetyl-cysteine and mycothiol. 4-OH-OPB killed Mtb synergistically with oxidants and several antituberculosis drugs. Thus, conditions that block Mtb’s replication modify OPB and enhance its cidal action. Modified OPB kills both replicating and NR Mtb and sensitizes both to host-derived and medicinal antimycobacterial agents.

Some bacterial infections can be cured with a single dose of an antibiotic, and most others can be cured with administration of one drug over several days or weeks. In contrast, routine treatment of drug-sensitive tuberculosis (TB) takes 2 mo of therapy with four drugs and an additional 4 mo with two drugs to reduce the 2-y relapse rate below 5%. The difficulty of completing prolonged treatment is a major reason for emergence of drug resistance. When the infecting strain is resistant to isoniazid and rifampin, the two drugs recommended for all 6 mo of treatment, cure often requires 2 y of daily administration of toxic, expensive, second-line agents that are often unavailable at the point of care. When the causative strain is additionally resistant to a quinolone and an aminoglycoside, the resultant “extensively drug-resistant” TB was fatal to 80% of patients in a leading center (1), although complex multidrug regimens have achieved higher cure rates in populations not previously exposed to the additional drugs (2). In addition to sharing air with someone with TB, leading risk factors for contracting the disease are malnutrition, HIV infection, diabetes, and exposure to smoke from cigarettes or cooking fires (3). These epidemiological challenges exacerbate problems of inadequate diagnostic technology and limited access to drug susceptibility testing and to drugs. Control of the pandemic is not in sight (3).

It is widely hypothesized that treatment of TB is protracted because nonreplicating (NR) subpopulations of bacilli are phenotypically tolerant to drugs that were selected for activity against replicating (R) Mycobacterium tuberculosis (Mtb) (4). Mtb can occupy diverse microenvironments in the host. Evidence from auxotrophs, analyses of gene expression, and direct and indirect biochemical measurements in vivo or ex vivo in experimental animals and people suggest that such environments expose Mtb to acid, hypoxia, reactive nitrogen intermediates (RNIs), oxidative stress, carbohydrate deficiency, and metal starvation or intoxication, and require Mtb to metabolize fatty acids or cholesterol (517). In vitro, many of the same conditions can make Mtb relatively refractory to killing by the standard agents, except for pyrazinamide, which is only effective at a low pH.

Thus, there is a need for a high-throughput screen (HTS) for compounds that kill Mtb when the Mtb has been rendered NR by a combination of physiologically relevant host-imposed conditions. We were encouraged to devise such a screen by recent discoveries of a class of compounds that kill Mtb only when it is NR (18), an antibiotic in clinical use for other infections that kills NR Mtb better than R Mtb (19), and a compound that kills NR and R Mtb equally well (20). Unfortunately, only one of those compounds is an approved drug, and even if it were of proven utility in TB, its price would preclude its use by most of those who need it. We decided to screen other existing drugs that are not regarded as antiinfectives for those that kill NR Mtb. Here, we report finding such a drug in an HTS that combined four host-imposed conditions, some of which converted the drug into a form active on both R and NR Mtb.


Screen for Compounds That Kill NR Mtb, R Mtb, or Both.

We set out to identify drugs that can kill Mtb in the face of replication-inhibiting conditions. HTSs depend on robots that are difficult to accommodate in the biological safety level (BSL) 3 conditions required for work with pathogenic strains of Mtb. We took advantage of mc26220, a [increment]panCD[increment]lysA double-auxotrophic strain of Mtb H37Rv (a kind gift of W. Jacobs, Jr., Albert Einstein College of Medicine, New York), which has been found to cause no disease when injected into immunocompetent or immunodeficient mice, guinea pigs, or monkeys (21, 22). This strain has been deemed to be safe for use in BSL2 laboratories by the Institutional Biosafety Committees of the Albert Einstein College of Medicine and Weill Cornell Medical College, as approved by the US National Institutes of Health. Provision of pantothenate and lysine allows the auxotroph to grow like WT in vitro. The HTS under R conditions identified 24 actives that have known anti-Mtb activity; minimal inhibitory concentrations (MICs) were determined for 11 and were similar when tested against WT Mtb and the auxotroph (Table S1), validating use of the auxotroph.

It is a challenge to detect compounds that kill NR Mtb when the criterion for death is inability to replicate. We solved this problem by conducting the assay in two stages. In the first stage, we halted Mtb’s replication by incubation in 96- or 384-well microplates in modified Sauton’s minimal medium at pH 5.5 in 1% O2 and 5% CO2 with 50 μM butyrate or isobutyrate as the sole carbon source in the presence of 0.5 mM nitrite. These conditions prevented growth yet led to little or no decline in cfu over 6 d (Fig. S1A). We exposed Mtb to test compounds at 12.5 μM for 6 d and then diluted the contents of each well 21-fold into Middlebrook 7H9 medium containing dextrose and glycerol as carbon sources at pH 6.6 in 21% O2 and 5% CO2 without nitrite, conditions that support Mtb’s exponential replication. After 7–10 d, we recorded the OD in each well in comparison to negative and positive control cultures that contained 0.25% DMSO or 12.5 μM rifampin, respectively, in the first stage. This two-stage screen had an average Z′ value of 0.7. In a parallel screen, we tested the same compounds against R Mtb in a one-stage assay using the same medium as for the second stage of the two-stage assay. The one-stage screen had an average Z′ value of 0.9. Based on the distribution of results (Fig. S2), “hits” were defined as compounds that produced 60% inhibition of growth compared with the DMSO controls. We screened overlapping sets of past and present pharmaceuticals in the Spectrum and Johns Hopkins University collections (23), numbering ~3,600, along with ~2,000 natural products of bacterial and plant origin, with a confirmed hit rate of 2.7%. Of the actives among these ~5,600 compounds, 35% were active only against R Mtb, 17% only against NR Mtb, and 48% against both (Fig. 1A). Most were antiinfectives already known to kill Mtb; cytotoxic drugs, such as 5-fluorouracil and gliotoxin; or compounds likely to be toxic, such as the quinones juglone and lawsone. One active drug stood out because it fit none of those categories.

Fig. 1.
HTS of drugs and natural products for compounds that kill R or NR Mtb. (A) One of two independent screens of ~5,600 compounds that were scored as active against R Mtb, NR Mtb, both, or neither. OPB (result circled) was uniquely active in NR conditions. ...

Oxyphenbutazone Selectively Kills NR Mtb.

Oxyphenbutazone (OPB) (Fig. 1B), an extensively used nonsteroidal anti-inflammatory drug (NSAID), inhibited growth of NR Mtb by 100% at 12.5 μM but not at up to 200 μM when tested against R Mtb in liquid cultures (Fig. 1A). The closely related pyrazolidinediones phenylbutazone (PB), suxibuzone, and sulfinpyrazone (Fig. 1B) did not kill either NR or R Mtb (Fig. 1A). Results using OD as a read-out were confirmed by cfu assays with pure, resupplied OPB (Fig. 1 C and D). Killing by OPB was concentration-dependent, time-dependent, and extensive against both the auxotrophic strain (Fig. 1 C and D) and WT Mtb (Fig. S1B). In contrast, neither OPB nor PB was bactericidal to Escherichia coli, Salmonella enterica var. Typhimurium, Pseudomonas aeruginosa, Staphylococcus aureus, or Candida albicans up to 100 μM, nor was OPB cytotoxic to monkey kidney epithelial (Vero) cells or murine bone marrow-derived macrophages at 200 μM.

Next, we determined the individual contributions of the conditions used to prevent Mtb from replication. At pH 7, OPB was inactive in the presence or absence of butyrate or nitrite with either 1% or 21% O2. OPB’s mycobactericidal potency was markedly increased as the pH was lowered from 5.5 to 4.5 (Fig. 2A), the level in the phagosome of activated macrophages (8). OPB’s mycobactericidal activity was modestly enhanced by nitrite, but even though decreasing the pH enhances the rate of RNI production from nitrite, activity became largely nitrite-independent at pH 4.5 (Fig. 2B). Butyrate also increased activity. The activity at low pH was indistinguishable at O2 levels of 21% and 1%.

Fig. 2.
NR conditions that support mycobactericidal activity of OPB. We tested all 16 sets of conditions (pH 5.5 or 7.0, ±NaNO2, ±butyrate, 1% or 21% O2). Those with greatest impact were pH 5.5, as illustrated by further lowering the pH in the ...

Reactions of OPB Under Conditions That Impose Nonreplication.

Mild acidity, RNIs, and fatty acid might alter either OPB or Mtb. As monitored by liquid chromatography (LC)-MS and NMR, OPB was stable at pH 7.0 with or without butyrate, nitrite, or diethylenetriamine NONOate (DETA-NO), which donates NO at pH 7.0. In contrast, at pH 5.0, OPB rapidly and quantitatively converted to a species whose mass [MH+ = 341.27 absolute mass units (amu)] was consistent with the addition of an oxygen to OPB (MH+ = 325.15 amu) (Fig. 3A). Nitrite increased the rate of conversion (Fig. 3A). We identified the unknown species as 4-butyl-4-hydroxy-1-(4-hydroxyphenyl)-2-phenylpyrazolidine-3,5-dione (4-OH-OPB) (Fig. 3 B and C). After a brief incubation in aqueous solution at 37 °C, 4-OH-OPB underwent opening of the N-N bond in the pyrazolidinedione ring to form a quinonimine (Fig. 3 B and C), a Michael acceptor. NMR spectra for 4-OH-OPB and its ring-opened quinonimine were consistent with those described by Dekkers et al. (24). The addition of a hydroxyl to the pyrazolidinedione ring was consistent with a large 13C shift of the C4 carbon from 42 ppm (seen in OPB) to 81 ppm (seen in 4-OH-OPB) using 2D 13C-heteronuclear single-quantum coherence (HSQC) and 2D 13C-heteronuclear multiple-bond correlation (HMBC) spectra.

Fig. 3.
4-Hydroxylation of OPB under NR conditions. (A) Impact of acid and nitrite on 4-hydroxylation. OPB was incubated for 1 h (hatched bars) or 24 h (solid bars) in R conditions or at pH 5.0 ± NaNO2 ± butyrate, all of which are NR conditions. ...

Bioactivity of 4-OH-OPB.

Because OPB rapidly formed 4-OH-OPB under conditions in which OPB was mycobactericidal, we synthesized 4-OH-OPB, which proved to be equipotent to OPB in killing NR Mtb (Fig. 3D). The potency of each was modestly enhanced by omission of BSA (Fig. 3D). Although butyrate did not contribute to conversion of OPB to 4-OH-OPB, it did enhance the mycobactericidal activity of 4-OH-OPB on NR Mtb.

In contrast to OPB, 4-OH-OPB killed R Mtb in liquid culture with an MIC of 100 μM in the presence of 0.5% BSA and an MIC of 25 μM at lower BSA concentrations (Fig. 3E). Assays performed independently at SRI International confirmed that the MIC of 4-OH-OPB against drug-sensitive R Mtb H37Rv was 26 μM and that 4-OH-OPB had comparable MICs against R Mtb strains that were resistant to streptomycin (13 μM), ethionamide (13 μM), isoniazid (26 μM), p-aminosalicylate (26 μM), kanamycin (26 μM), or ethambutol (26 μM).

The bactericidal spectrum of 4-OH-OPB was remarkably narrow, with weak activity against Mycobacterium smegmatis (MIC of 200 μM in 7H9 lacking BSA and >200 μM in 7H9 with oleic acid, albumin, dextrose, catalase (OADC) or albumin, dextrose, sodium chloride (ADN) supplement) and S. aureus (MIC of 50 μM). 4-OH-OPB was inactive on E. coli, S. enterica var. Typhimurium, P. aeruginosa, and C. albicans up to 200 μM, even when the tests with C. albicans were conducted at pH 5.5 or 4.5 (Fig. S3).

Identification of Intramycobacterial Targets of OPB and 4-OH-OPB.

Targeted metabolomics of R Mtb growing on filters on drug-containing agar (25) and LC-MS analysis of lysates of NR Mtb in liquid culture demonstrated that Mtb exposed to PB contained no detectable intrabacterial PB after 24 h. In contrast, Mtb cultured with OPB contained both OPB and 4-OH-OPB. We used the filter-based method to study the kinetics of uptake. OPB was detectable in Mtb by 15 min, the first time point tested. The concentration increased steeply over 60 min and gradually over the next 23 h (Fig. 4A).

Fig. 4.
Uptake of OPB and impact on Mtb metabolites. (A) Rapid OPB uptake into Mtb from agar plates containing 100 μM OPB. Ion counts were normalized using residual peptides in each sample and a standard curve of OPB in DMSO. (B) Structure of OPB heptadeuterated ...

In addition to OPB and 4-OH-OPB, several chemical species appeared in OPB-treated Mtb that were not observed in vehicle-treated Mtb (Table S2). None matched metabolites in the Kyoto Encyclopedia of Genes and Genomes database. To test if these were covalent adducts of OPB or its derivatives with Mtb metabolites, we synthesized an analog of OPB, 4-(butyl-2,2,3,3,4,4,4-d7)-1-(4-hydroxyphenyl)-2-phenylpyrazolidine-3,5-dione (OPB-d7) (Fig. S4) containing seven deuterium atoms in place of hydrogens on the butyl chain, conferring an increased mass of 7.0439 amu (Fig. 4B). Numerous molecular species were undetectable in untreated Mtb, present in OPB-treated Mtb, and present but increased in mass by 7.0439 amu (±5 ppm error) in Mtb treated with OPB-d7, as illustrated in Fig. 4C. The characterization of over 30 major species that fit these criteria is summarized in Table S2. These features identified them as adducts with OPB or its derivatives. The most abundant species had masses corresponding to predicted reaction products (Figs. S5S7) of the quinonimine forms of OPB and 4-OH-OPB with the NAc-cysteine (NAC)–ligated disaccharide mycothiol (MSH) and/or its precursor, NAC (Fig. 4D and Table S2), the former as confirmed by in vitro reaction of 4-OH-OPB with authentic MSH (Fig. 4E). Where tested, these tentative assignments were strongly supported by the masses of fragments in MS/MS (Table S2). Thus, Mtb detoxifies OPB and 4-OH-OPB, in part, by forming covalent adducts with the intracellular thiols NAC and MSH.

Among the other most conspicuous metabolomic changes, chemical species with properties of flavin nucleotides were abundant in untreated Mtb but depleted in OPB-treated Mtb. An enzyme-coupled fluorometric assay confirmed depletion of the flavin pool (Fig. 4F).

OPB Sensitizes Mtb to Oxidants and Exogenous Antimicrobials.

MSH, an abundant thiol in Mtb, functions like glutathione (26). Genetic disruption of its biosynthesis sensitizes Mtb to oxidative stress (27). Consistent with competition for MSH, sublethal 4-OH-OPB treatment rendered R Mtb eightfold more sensitive than untreated Mtb to growth inhibition by H2O2 and fourfold more sensitive to the oxidants plumbagin, methyl viologen, and copper (12). Deletion of genes of the MSH biosynthetic pathway sensitizes Mtb to diverse antibiotics (27, 28). To see if OPB also sensitizes Mtb to other agents, we rescreened the original compound library in the presence of a single sublethal concentration of OPB. Of the compounds that appeared to synergize with OPB, we identified those that appeared to synergize with 4-OH-OPB as well. We picked four of them, including three TB drugs, for checkerboard analysis. 4-OH-OPB exhibited robust synergy with all four, p-aminosalicylate, fenamisal, nitrofurazone, and PA-824 (29), as defined by fractional inhibitory concentrations of <0.5 (Fig. 5 AD). In contrast, there was no synergy with amikacin or with isoniazid.

Fig. 5.
Synergy between 4-OH-OPB and known TB drugs in killing Mtb. Synergistic killing of R Mtb by 4-OH-OPB (0 μM, black circles; 10 μM, red circles) with the following drugs: p-aminosalicylate (PAS) (A), fenamisal (B), nitrofurazone (C), and ...

OPB in Mice.

Mice were given OPB by gavage at a dose (20 mg/kg of body weight) far above standard practice in humans (4–8 mg/kg in a 70-kg person). The serum concentration peaked at only 2 μM with a half-life of 3 h. The i.p. injections of 300–600 mg/kg of OPB, about 100-fold the human dose, achieved peak serum levels of only 160–205 μM with a half-life <2 h. In contrast, standard oral dosing of humans with OPB was reported to result in a peak serum concentration of 77–308 μM with a half-life of 50–75 h (3034). Thus, exposure to OPB was vastly lower in mice than in humans, making the mouse an unsuitable model in which to test OPB.


We identified OPB, an inexpensive, off-patent NSAID, as a narrow-spectrum antimycobacterial agent. OPB’s mycobactericidal activity was conferred by conditions that impose nonreplication on Mtb. The most important of these conditions, low pH, is obtained in the phagosome of activated macrophages (8) and at inflammatory sites. RNIs also contributed to the activity of OPB. RNIs are produced by human macrophages in tuberculous lesions (35). In granulomas, hypoxia may limit RNI production by host enzymes, but O2-starved Mtb respires nitrate (which is abundant in body fluids) and produces its own nitrite (36, 37). Mild acid converted OPB to a form that was active against R Mtb as well as NR Mtb, and RNIs accelerated the conversion.

The primary product of OPB exposed to mild acidity and RNIs was 4-OH-OPB. OPB did not kill R Mtb in liquid cultures, although we confirmed an observation made over 40 y ago (38) that OPB was active against R Mtb on agar plates (MIC of 10 μg/mL). In contrast, 4-OH-OPB killed R Mtb both in liquid culture and on agar plates, and it killed NR Mtb as well. Although PB was not detectable in Mtb over a 24-h period, OPB was detectable in Mtb within 15 min of exposure. When Mtb was exposed to OPB under NR conditions, 4-OH-OPB also accumulated rapidly in the bacteria. Provision of fatty acid enhanced OPB’s mycobactericidal activity in the presence of BSA. Fatty acids may displace OPB from BSA. Alternatively, OPB may sensitize Mtb to the toxicity of fatty acids (9, 39, 40).

The inability of PB to kill Mtb can be attributed to two features. First, it failed to accumulate within Mtb. Second, PB’s lack of a phenolic hydroxyl precludes 4-OH-PB from forming a quinonimine. Thus, 4-OH-PB is less chemically reactive than 4-OH-OPB.

OPB or its derivatives, most likely the 4-OH-OPB quinonimine, depleted flavins. Many enzymatic reactions could be compromised as a result. Moreover, OPB or its products formed adducts with MSH and NAC, an intermediate in MSH synthesis. This may contribute to OPB’s sensitization of Mtb to acid, oxidant stress, fatty acids in combination with low pH and RNIs, and certain antibiotics (27, 41). Numerous other adducts were formed that we did not identify. The mechanism of killing is thus likely to be multifactorial.

With respect to its mycobactericidal activity, OPB can be considered a prodrug. However, OPB does not require an enzyme for its activation. This, plus its multiplicity of targets, may explain why our extensive efforts to select OPB-resistant mutants of Mtb failed, such that the frequency of resistance appears to be <10−9.

OPB’s antimicrobial spectrum was narrow, and its cytotoxic action against mammalian cells was minimal. In part, this may reflect that conversion to the more reactive 4-OH-OPB quinonimine depends on conditions that are likely to be relatively restricted to certain pathological microenvironments. However, OPB did not kill C. albicans even at pH 4.5 or 5.5, where formation of 4-OH-OPB was rapid and extensive, and preformed 4-OH-OPB was inactive on C. albicans and other microbes. Investigators studying 4-OH-OPB as an anti-inflammatory agent reported it to be nontoxic in mice at serum concentrations <20 mM (42) and to be safe in people (43). OPB’s narrow antimicrobial and cytotoxic spectrum may be defined chiefly by differential abilities of various cells to take up, export, or metabolize OPB and 4-OH-OPB.

PB was introduced into clinical medicine in 1949 (32). It was soon demonstrated that humans metabolize PB, in part, to OPB, and OPB was introduced as a drug a few years later. PB and OPB were both widely used until less toxic but more costly NSAIDs replaced them in many countries. However, PB and OPB are still used in regions where cost governs access. Gastrointestinal distress with OPB is common (~10%) but avoidable with coated tablets or antacids (44). Less frequent are rash, gastric ulcers, hypersensitivity reactions, Na+ retention, hepatitis, renal disorders, leukopenia, agranulocytosis, and aplastic anemia. The most serious toxicities are more frequent in the elderly. The incidence of acute bone marrow failure was estimated at less than 1 in 50,000 (45) and at 1 in 66,000 (44). In the setting of drug-resistant TB, the major toxicities of OPB should be considered in the context of the dire prognosis and compared with those of second-line drugs, which averaged 16% in a recent report (46). Of practical concern, most patients with TB who need second-line drugs fail to receive them because of cost. OPB is exceedingly inexpensive.

We expected that some patients experiencing pain and fever associated with TB may have been given OPB for symptomatic relief. We found three such reports involving a total of 84 patients, who were reported to experience clearance of drug-resistant Mtb from sputum (47), faster gain of weight (48), or improved tolerance to conventional therapy (49). These studies neither establish nor exclude that OPB’s apparent impact in patients with TB reflected an anti-inflammatory action, a microbiological action, augmentation of the microbiological action of other drugs, or some combination of these effects. Nonetheless, in 84 subjects receiving 300–600 mg of OPB per day for 2–6 mo, addition of OPB to standard TB treatment did not lead to additional toxicity but, instead, was reported to confer clinical benefit.

Although the mouse is the model of convention and convenience for preclinical tests of candidate anti-TB drugs, differences in metabolism of some drugs between mice and humans are too great for the mouse to be of use in modeling these drugs’ exposure in humans. This proved to be the case with OPB. In cases in which the mouse is pharmacokinetically unsuitable to test a particular agent that otherwise holds promise and has already been used safely in patients with TB, it is our belief that the benefit-to-cost ratio of clinical trials in patients with TB will exceed that of experiments in large animal models.

In conclusion, we devised an HTS that finds compounds that kill Mtb whose replication is inhibited by a combination of four pathophysiologically relevant conditions, extending the initial observation that antiinfectives need not be uniquely or preferentially active on R bacteria (18). We demonstrated that use of an Mtb auxotroph allows robotic screening under BSL2 conditions. This may encourage others who lack access to robotics in BSL3 conditions to conduct screens against Mtb. Testing known drugs, we identified OPB, a widely used, inexpensive, relatively safe NSAID with narrow-spectrum mycobactericidal activity and a history of clinical tolerability and potential benefit in patients with TB. We demonstrated that OPB’s mycobactericidal mechanism involves condition-dependent conversion to reactive species that deplete Mtb’s thiols and flavins. OPB sensitizes Mtb to killing by two important sets of factors: the same host chemistries that make Mtb phenotypically resistant to many conventional anti-TB agents and some of those same drugs. Addition of OPB to currently available treatments may benefit patients with drug-resistant TB while we wait for the reemergent TB drug pipeline to deliver affordable new agents.

Materials and Methods

Strains and Growth Conditions.

WT Mtb H37Rv was cultivated in Middlebrook 7H9 medium with 0.2% glycerol, tyloxapol (0.02%), and 10% ADN supplement. Mtb strain mc26220 [increment]panCD[increment]lysA was passaged in Middlebrook 7H9 supplemented with glycerol (0.5%), OADC, tyloxapol (0.02%), casein hydrolysate (CAS) amino acids (0.05%), l-lysine (240 μg/mL), and pantothenate (24 μg/mL). R conditions included incubation in 20% O2 and 5% CO2. NR conditions used minimal Sauton’s-based medium [per liter: 0.5 g of KH2P04, 0.5 g of MgS04, and 0.05 g of ferric ammonium citrate (this trace amount [0.005%] of citrate does not support growth)] supplemented with BSA (0.5%), NaCl (0.085%), tyloxapol (0.02%), l-lysine (240 μg/mL), pantothenate (24 μg/mL), and butyrate (50 μM), omitting glycerol, citrate, and asparagine from Sauton’s recipe, at pH 5.5 with 0.5 mM freshly prepared NaNO2. NR conditions included incubation at 37 °C in chambers (BioSpherix) at 1% O2 and 5% CO2. The following microbes were grown in 150 μL in 96-well microtiter plates at 37 °C: uropathogenic E. coli TOP10 (Luria broth), S. enterica var. Typhimurium (Luria broth), S. aureus American Type Culture Collection (ATCC) 29213 (Mueller–Hinton broth), P. aeruginosa PAO1 (Luria broth), and C. albicans ATCC 90028 (YM broth) at pH 5.5.


HTS data were managed using Collaborative Drug Discovery software and JChem for Excel and MarvinView (ChemAxon). For HTS against NR Mtb, 100 μL of the NR medium with 0.5 mM NaNO2 was added to 96-well plates (Corning), followed by known drugs and bioactives in the collections from Prof. J. Liu (Johns Hopkins University, ~2,000), Spectrum (~2,000), and Analyticon (~2,000) as 0.5 μL of 5-mM stocks in DMSO. Mtb mc26220 in log phase (~0.5 A580) was washed twice with PBS containing tyloxapol (0.02%; PBS-Ty) and resuspended at an OD580 of 0.4 in NR medium containing 0.5 mM NaNO2 for 30–60 min. Mtb suspension (100 μL) was dispensed to each well, such that the final volume was 200 μL, OD580 of 0.2, 0.25% DMSO, and ~12.5 μM test agent. Wells were mixed, and plates were incubated under 1% O2 and 5% CO2 in stacks between hydration plates containing only PBS. After 7 d, the Mtb in each well was resuspended and diluted 21-fold (10 μL into 200 μL) into fresh R medium. Test agents were thus carried over at ~0.6 μM. After 10–14 d of incubation at 37 °C in 20% O2 and 5% CO2, cells in outgrowth plates were resuspended and an aliquot (100 μL) was transferred to clear-bottomed, black, 96-well microtiter plates (Greiner) for OD580 determination by bottom-read. The HTS against R Mtb used R medium in a single-stage assay with an initial OD of 0.1 and 7–8 d of incubation in 20% O2 and 5% CO2. During the course of this study, we converted to assays in black, clear-bottomed, 384-well plates (Greiner). To a final assay volume of 50 μL, we added either 0.25 μL or 0.50 μL of compounds at 5 mM in DMSO. Inhibition of outgrowth was calculated as 100 − (100[ODexp − ODrifampin/(ODDMSO − ODrifampin)]), where ODexp is the OD of the experimental well and ODDMSO and ODrifampin are means for 32 negative and positive controls on each plate, respectively.

The MIC was defined as the lowest concentration at which there was no increase in OD beyond the starting value of 0.01. Synergy experiments were performed in R medium containing 0.005% BSA. Fractional inhibitory concentrations for wells containing various concentrations of drug A and drug B were determined as the fractional inhibitory concentration = ([A]/MICA + [B]/MICB) when inhibition is ≥90% in the simultaneous presence of compound A and compound B, each used below its own MIC. MICs determined at SRI International were based on resazurin reduction after 6 d of exposure of ~104 Mtb per well to eight different concentrations of 4-OH-OPB in 7H9 medium with 10% OADC.

Vero monkey kidney cells (ATCC CRL-1587) seeded at 104 cells per 200 μL per well in 96-well microtiter plates were incubated at 37 °C in DMEM with 4.5 g/L glucose, 2 mM l-glutamine, 1 mM sodium pyruvate, 100 U/mL penicillin, 100 μg/mL streptomycin, 0.5 mg/mL gentamicin, and 10% FBS. Cells were washed once with PBS, and medium was added containing 2% FBS and compound (0–100 μM). After 48 h, viability was measured by tetrazolium reduction (MTS assay; Promega). Bone marrow-derived macrophages were obtained from femurs of C57BL/6 mice differentiated for 7 d in DMEM with 10% L-cell conditioned medium and 10% FBS.

Special Reagents.

OPB was resupplied (MP Biomedicals), and its purity was confirmed by LC-MS, IR spectroscopy, and NMR spectroscopy. PB was from Sigma. MSH was a kind gift of R. Fahey and G. Newton (University of California, San Diego, CA).

Analysis of Metabolites.

Mtb was incubated for 24 h under NR conditions with or without OPB or 4-OH-OPB, centrifuged, washed twice with PBS-tyloxapol, resuspended in 500 μL of ice-cold methanol/acetonitrile/water (40:40:20 ratio), lysed by mechanical homogenization (3 × 30 s) with cooling on ice, filter-sterilized using 0.2-μ filters, and frozen at −80 °C until analysis. Volatiles were removed with a Speedvac (Labconco; 30 min, room temperature). Samples were then loaded on a Dionex 3000 nano-LC system coupled with a Thermo Scientific LQ-Orbitrap XL Mass Spectrometer providing mass measurement error <5 ppm. As standards, we reacted authentic MSH with 4-OH-OPB. The flavin adenine dinucleotide pool (oxidized plus reduced) was measured using a kit (BioVision).

NMR Spectroscopy.

1H and 13C NMR spectra were recorded at 11.7 T on a Bruker 500-MHz Avance III spectrometer equipped with a broad-band gradient probe. Spectra were externally referenced to tetramethylsilane (TMS) in appropriate solvent for 1D spectra. For 2D spectra, 1H and 13C frequencies were referenced to the internal DMSO solvent signal calibrated to TMS (50). Kinetic experiments were performed at 308 K. Spectra were acquired using 0.5-mM samples of OPB or 4OH-OPB in PBS prepared in D2O at pH 5.0. Samples were warmed before initiation of kinetics. The time course was initiated on addition of an aliquot of 1 M NaNO2 in D2O to give a final NaNO2 concentration 10-fold higher than the sample concentration. The 1D 1H spectra were collected at intervals for up to 18 h. Natural abundance 13C chemical shifts were obtained at 298 K using 2D 13C-HSQC (51) and 2D 13C-HMBC (52) spectra. Natural abundance data were collected at sample concentrations from 25 to 100 mM using 25%, 58%, and 75% d6-DMSO solutions mixed with PBS prepared in D2O at pH 5.0. Spectra were processed, analyzed, and plotted using the NMR processing software MestReNova 5.3 (Mestrelab Research). Signal intensities were integrated for kinetic analysis of decay and growth rates. Curves were fitted and plotted to single exponential models using Grace 5.1 (GNU general public license) software.


Uptake of gavaged OPB was determined by Quest Pharmaceutical Sciences under an institutional animal care and use committee-approved protocol. OPB was dissolved in 5% DMSO/95% of 1% methylcellulose in water and administered to adult female C57BL/6 mice. Plasma OPB concentrations were analyzed by LC/MS/MS. Serum concentrations of OPB following i.p. injection were determined in-house. To each sample (100 μL) was added 20 μL of citrate/phosphate buffer (0.1 M, pH 3) and 600 μL of chloroform. Mixed samples were centrifuged at 1,800 × g for 20 min at 4 °C. The recovered chloroform layer was evaporated, and the residue was dissolved in 20 μL of methanol and 2.5 μL of DMSO for HPLC analysis.

Supplementary Material

Supporting Information:


We thank M. Larsen and W. R. Jacobs, Jr. (Albert Einstein College of Medicine) for the auxotrophic Mtb; T. Parker (Tuberculosis, Leprosy, and Other Mycobacterial Diseases Section, Respiratory Diseases Branch, Division of Microbiology and Infectious Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health) for access to Resources for Researchers services at SRI International and Johns Hopkins University; Kathy N. Williams and Opokua Amoabeng (Johns Hopkins University) for assistance with the mouse studies; R. Fahey and G. Newton (University of California, San Diego) for a MSH standard; D. J. Rouse (RTA International) and The Global Alliance for Tuberculosis Drug Development for PA-824; H. N. Sultan (The Rockefeller University HTS Resource Center) for help with HTS; K. Rhee, L. P. S. de Carvalho, and S. Chakraborty for the metabolomic study on filter-grown cultures; G. Sukenick (NMR Analytical Core) and G. Yang (Organic Synthesis Core, Memorial Sloan Kettering Cancer Center) for help with analysis and synthesis of OPB analogs; S. Ekins (Collaborative Drug Discovery) for cheminformatics work not included here; Alfonso Mendoza-Losana (GlaxoSmithKline) for an experiment not included; J. Liu (Johns Hopkins University) for sharing a collection of pharmaceutical agents; R. Elliott and K. Duncan (Bill and Melinda Gates Foundation) and M. Reidenberg, S. Ehrt, E. Talley, G. Lin, R. Bryk, and L. P. S. de Carvalho (Weill Cornell Medical College) for advice; H. A. Brown, K. Parikh, J. Pale, and S. Muffly (Weill Cornell Medical College) and L. Solla (Cornell University) for literature review; and F. Barany (Weill Cornell Medical College) for access to BSL2 facilities and robotics. This work was supported by the Tuberculosis Drug Accelerator Program of the Bill and Melinda Gates Foundation and by the Abby and Howard P. Milstein Program in Chemical Biology of Infectious Disease. The Department of Microbiology and Immunology (Weill Cornell Medical College) is supported by the William Randolph Hearst Foundation.


The authors declare no conflict of interest.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1214188109/-/DCSupplemental.


1. Chan ED, Strand MJ, Iseman MD. Treatment outcomes in extensively resistant tuberculosis. N Engl J Med. 2008;359:657–659. [PubMed]
2. Mitnick C, Horsburgh CR., Jr Encouraging news for multidrug-resistant tuberculosis treatment. Am J Respir Crit Care Med. 2010;182:1337–1338. [PubMed]
3. Nathan C. Taming tuberculosis: A challenge for science and society. Cell Host Microbe. 2009;5:220–224. [PubMed]
4. Young DB, Duncan K. Prospects for new interventions in the treatment and prevention of mycobacterial disease. Annu Rev Microbiol. 1995;49:641–673. [PubMed]
5. Gould TA, van de Langemheen H, Muñoz-Elías EJ, McKinney JD, Sacchettini JC. Dual role of isocitrate lyase 1 in the glyoxylate and methylcitrate cycles in Mycobacterium tuberculosis. Mol Microbiol. 2006;61:940–947. [PubMed]
6. Muñoz-Elías EJ, McKinney JD. Mycobacterium tuberculosis isocitrate lyases 1 and 2 are jointly required for in vivo growth and virulence. Nat Med. 2005;11:638–644. [PMC free article] [PubMed]
7. MacMicking JD, et al. Identification of nitric oxide synthase as a protective locus against tuberculosis. Proc Natl Acad Sci USA. 1997;94:5243–5248. [PMC free article] [PubMed]
8. MacMicking JD, Taylor GA, McKinney JD. Immune control of tuberculosis by IFN-gamma-inducible LRG-47. Science. 2003;302:654–659. [PubMed]
9. Vandal OH, Pierini LM, Schnappinger D, Nathan CF, Ehrt S. A membrane protein preserves intrabacterial pH in intraphagosomal Mycobacterium tuberculosis. Nat Med. 2008;14:849–854. [PMC free article] [PubMed]
10. Schnappinger D, et al. Transcriptional adaptation of Mycobacterium tuberculosis within macrophages: Insights into the phagosomal environment. J Exp Med. 2003;198:693–704. [PMC free article] [PubMed]
11. Marrero J, Rhee KY, Schnappinger D, Pethe K, Ehrt S. Gluconeogenic carbon flow of tricarboxylic acid cycle intermediates is critical for Mycobacterium tuberculosis to establish and maintain infection. Proc Natl Acad Sci USA. 2010;107:9819–9824. [PMC free article] [PubMed]
12. Gold B, et al. Identification of a copper-binding metallothionein in pathogenic mycobacteria. Nat Chem Biol. 2008;4:609–616. [PMC free article] [PubMed]
13. Gold B, Rodriguez GM, Marras SA, Pentecost M, Smith I. The Mycobacterium tuberculosis IdeR is a dual functional regulator that controls transcription of genes involved in iron acquisition, iron storage and survival in macrophages. Mol Microbiol. 2001;42:851–865. [PubMed]
14. Rodriguez GM, Voskuil MI, Gold B, Schoolnik GK, Smith I. ideR, An essential gene in mycobacterium tuberculosis: Role of IdeR in iron-dependent gene expression, iron metabolism, and oxidative stress response. Infect Immun. 2002;70:3371–3381. [PMC free article] [PubMed]
15. Timm J, et al. Differential expression of iron-, carbon-, and oxygen-responsive mycobacterial genes in the lungs of chronically infected mice and tuberculosis patients. Proc Natl Acad Sci USA. 2003;100:14321–14326. [PMC free article] [PubMed]
16. Walters SB, et al. The Mycobacterium tuberculosis PhoPR two-component system regulates genes essential for virulence and complex lipid biosynthesis. Mol Microbiol. 2006;60:312–330. [PubMed]
17. Miner MD, Chang JC, Pandey AK, Sassetti CM, Sherman DR. Role of cholesterol in Mycobacterium tuberculosis infection. Indian J Exp Biol. 2009;47:407–411. [PubMed]
18. Bryk R, et al. Selective killing of nonreplicating mycobacteria. Cell Host Microbe. 2008;3:137–145. [PMC free article] [PubMed]
19. de Carvalho LP, Lin G, Jiang X, Nathan C. Nitazoxanide kills replicating and nonreplicating Mycobacterium tuberculosis and evades resistance. J Med Chem. 2009;52:5789–5792. [PubMed]
20. Darby CM, Nathan CF. Killing of non-replicating Mycobacterium tuberculosis by 8-hydroxyquinoline. J Antimicrob Chemother. 2010;65:1424–1427. [PubMed]
21. Sambandamurthy VK, et al. Long-term protection against tuberculosis following vaccination with a severely attenuated double lysine and pantothenate auxotroph of Mycobacterium tuberculosis. Infect Immun. 2005;73:1196–1203. [PMC free article] [PubMed]
22. Larsen MH, et al. Efficacy and safety of live attenuated persistent and rapidly cleared Mycobacterium tuberculosis vaccine candidates in non-human primates. Vaccine. 2009;27:4709–4717. [PMC free article] [PubMed]
23. Chong CR, et al. Identification of type 1 inosine monophosphate dehydrogenase as an antiangiogenic drug target. J Med Chem. 2006;49:2677–2680. [PubMed]
24. Dekkers DWC, Aarden LA, Ten Brinke JA. 2010. Quinoneimines of malonic acid diamides. US Patent Appl 7,663,000.
25. de Carvalho LP, et al. Metabolomics of Mycobacterium tuberculosis reveals compartmentalized co-catabolism of carbon substrates. Chem Biol. 2010;17:1122–1131. [PubMed]
26. Newton GL, et al. Distribution of thiols in microorganisms: Mycothiol is a major thiol in most actinomycetes. J Bacteriol. 1996;178:1990–1995. [PMC free article] [PubMed]
27. Rawat M, et al. Mycothiol-deficient Mycobacterium smegmatis mutants are hypersensitive to alkylating agents, free radicals, and antibiotics. Antimicrob Agents Chemother. 2002;46:3348–3355. [PMC free article] [PubMed]
28. Rawat M, et al. Targeted mutagenesis of the Mycobacterium smegmatis mca gene, encoding a mycothiol-dependent detoxification protein. J Bacteriol. 2004;186:6050–6058. [PMC free article] [PubMed]
29. Singh R, et al. PA-824 kills nonreplicating Mycobacterium tuberculosis by intracellular NO release. Science. 2008;322:1392–1395. [PMC free article] [PubMed]
30. Perel JM, Snell MM, Chen W, Dayton PG. A study of structure-activity relationships in regard to species difference in the phenylbutazone series. Biochem Pharmacol. 1964;13:1305–1317. [PubMed]
31. Aarbakke J. Clinical pharmacokinetics of phenylbutazone. Clin Pharmacokinet. 1978;3:369–380. [PubMed]
32. Okonek S. Intoxication with pyrazolones. Br J Clin Pharmacol. 1980;10(Suppl 2):385S–390S. [PMC free article] [PubMed]
33. Okonek S, Reinecke HJ. Acute toxicity of pyrazolones. Am J Med. 1983;75(5A):94–98. [PubMed]
34. Gaucher A, Netter P, Faure G, Schoeller JP, Gerardin A. Diffusion of oxyphenbutazone into synovial fluid, synovial tissue, joint cartilage and cerebrospinal fluid. Eur J Clin Pharmacol. 1983;25(1):107–112. [PubMed]
35. Nathan C. 2006. Role of iNOS in human host defense. Science 312:1874–1875; author reply 1874–1875.
36. Malm S, et al. The roles of the nitrate reductase NarGHJI, the nitrite reductase NirBD and the response regulator GlnR in nitrate assimilation of Mycobacterium tuberculosis. Microbiology. 2009;155:1332–1339. [PubMed]
37. Tan MP, et al. Nitrate respiration protects hypoxic Mycobacterium tuberculosis against acid- and reactive nitrogen species stresses. PLoS ONE. 2010;5:e13356. [PMC free article] [PubMed]
38. Hoffmann K, Onöz E. Inhibitory effect of oxyphenbutazone against Mycobacterium tuberculosis in vitro. Arzneimittelforschung. 1969;19:241–242. (in German) [PubMed]
39. Dubos RJ. The effect of organic acids on mammalian tubercle bacilli. J Exp Med. 1950;92:319–332. [PMC free article] [PubMed]
40. Iwai K, Tsukuma S, Yoshida T, Naito M. Influence of serum concentration and pH of Kirchner’s liquid medium on the growth of tubercle bacilli. Acta Tuberc Jpn. 1965;14(2):65–74. [PubMed]
41. Buchmeier NA, Newton GL, Fahey RC. A mycothiol synthase mutant of Mycobacterium tuberculosis has an altered thiol-disulfide content and limited tolerance to stress. J Bacteriol. 2006;188:6245–6252. [PMC free article] [PubMed]
42. Tjotta E, Klaveness J. 2008. US Patent Appl 2008/0287390 A1.
43. Ten Brinke A, et al. 4-Hydroxy-oxyphenbutazone is a potent inhibitor of cytokine production. Eur Cytokine Netw. 2005;16:144–151. [PubMed]
44. Huskisson EC. Antiinflammatory drugs. Semin Arthritis Rheum. 1977;7(1):1–20. [PubMed]
45. Mills JA. Nonsteroidal anti-inflammatory drugs (first of two parts) N Engl J Med. 1974;290:781–784. [PubMed]
46. Carroll MW, et al. Frequency of adverse reactions to first- and second-line anti-tuberculosis chemotherapy in a Korean cohort. Int J Tuberc Lung Dis. 2012;16:961–966. [PubMed]
47. Stupenengo RH, Wendy CD. Experience with a nonspecific antiinflammatory agent in pulmonary tuberculosis. Sem Med. 1963;123:114–117. (in Spanish) [PubMed]
48. Doyle D. Oxyphenbutazone in pulmonary tuberculosis. S Afr Med J. 1966;40:743–744. [PubMed]
49. Schmidt W. 1970. Treatment of tuberculous pleuritis using oxyphenbutazone. Prax Pneumonol 24(3):169–172 (in German)
50. Gottlieb HE, Kotlyar V, Nudelman A. NMR chemical shifts of common laboratory solvents as trace impurities. J Org Chem. 1997;62:7512–7515. [PubMed]
51. Davis AL, Keeler TJ, Laue ED, Moskau D. Experiments for recording pure-absorption heteronuclear correlation spectra using pulsed field gradients. J Magn Reson. 1992;98:207–216.
52. Cicero DO, Barbato G, Bazzo R. Sensitivity enhancement of a two-dimensional experiment for the measurement of heteronuclear long-range coupling constants, by a new scheme of coherence selection by gradients. J Magn Reson. 2001;148:209–213. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...