*3.1. Isoniazid—Target: Mycolic Acids*

*Discovery.* In 1949, Colin Hinshaw and Walsh McDermott went to Germany to investigate reports that TB patients were successfully treated with a new synthetic compound, Conteben [36]. Conteben, a thiosemicarbazone also known as thiacetazone and Tibione (see thiacetazone section below), had been discovered by Gerhard Domagk at Bayer in West Germany. Hinshaw and McDermott returned to the USA with a supply of Conteben. After testing of Conteben in US hospitals, they concluded that "a prompt and thorough series of experimental and clinical trials in the United States" was justified along with experimenting with other thiosemicarbazones [37]. Two US pharmaceutical companies, Hoffman-La Roche and E. R. Squibb & Sons, quickly developed a series of thiosemicarbazone analogs but none showed better activity than Conteben [38–40] until the benzene ring in Conteben was replaced with a pyridine ring leading to the simultaneous discovery of Rimifon at Hoffman-La Roche [39] and Nydrazid at Squibb [41]. In parallel, Domagk at Bayer developed the thiosemicarbazone analog Neoteben [42]. Rimifon, Nydrazid and Neoteben had antiTB activity that far exceeded streptomycin, PAS and any other analog synthesized so far and share the same chemical structure: 4-pyridinecarboxylic acid hydrazide. Hoffman-La Roche, Squibb and Bayer had simultaneously and independently discovered the most potent TB drug at the time: isoniazid (INH). Ironically, none of the pharmaceutical companies could patent their discovery, as INH had been synthesized 40 years earlier by two Polish graduate students Hans Meyer and Josef Mally [36]. The results with INH were so striking that chemotherapy became the leading route for TB treatment, and soon after the sanatoria closed. Almost 70 years later, INH remains the cornerstone of TB chemotherapy for the treatment of drug-susceptible and latent *M. tuberculosis* infections.

*Activity and toxicity.* INH is a first-line TB drug. INH is an oral, highly soluble in water (140 g/L), bactericidal drug with a minimum inhibitory concentration (MIC) ranging from 0.1 to 0.7 μm against *M. tuberculosis*. In vitro, INH rapidly reduces the number of *M. tuberculosis* bacteria by 2 to 3-log10s during the first four days of treatment [43]. This bactericidal activity is only observed in exponentially growing *M. tuberculosis* cultures. INH has no activity in stationary or persistent *M. tuberculosis*. In mice, a similar pattern is observed, where INH is bactericidal only on actively dividing *M. tuberculosis* [44]. INH is readily absorbed and reaches concentrations in tissues and organs above its MIC.

The main adverse effect of INH is hepatotoxicity. TB patients more likely to develop liver damage when taking INH are slow acetylators [45,46]. INH is acetylated into the inactive molecule AcINH by the human *N*-acetyltransferase-2 (NAT2), which is expressed mostly in the liver and gastrointestinal tract. Genetic polymorphisms in NAT2 renders this acetylation reaction either slow or fast, dividing TB patients between slow and fast acetylators. INH metabolism also involves an amidase metabolizing INH into an hydrazine (Hz), which is then acetylated by NAT2 to form a toxic acylhydrazine (AcHz) and a non-toxic diacylhydrazine (DiAcHz) [46]. Hz, AcHz and their metabolites generated by the

liver microsomal cytochrome P450 enzymes have been linked to liver toxicity [47,48]. Fast acetylators have a lower risk of liver toxicity by producing more of the DiAcHz metabolite and less of the AcHz metabolite(s) [49].

*Mode of action.* INH enters *M. tuberculosis* cells by passive diffusion as a prodrug. A prodrug is the ideal drug where the compound has, if possible, no effect on eukaryotic cells, yet, once activated by a pathogen-specific enzyme, leads to the death of the pathogen [50]. Curiously, the TB pharmacopeia is composed of an array of prodrugs approved for clinical therapy as well as in the developmental phase [51]. INH is activated by the mycobacterial catalase peroxidase KatG (Rv1908c) [52] into what is most likely an isonicotinoyl radical that reacts with nicotinamide adenine dinucleotide (NAD+) to form the INH-NAD adduct [53–56]. This adduct binds to and inhibits InhA [54,57–60], the NADH-dependent enoyl-ACP reductase [61,62] of the FAS-II system [63], leading to the inhibition of mycolic acid biosynthesis and mycobacterial cell death [60,64–66]. Although the mechanism of INH action seems rather straightforward, elucidation of the molecular details took almost fifty years [67].

*Resistance.* Shortly after the antimycobacterial activity of INH was published in 1952, the first report of INH-resistant clinical isolates appeared less than one year later [68]. In 1954, Gardner Middlebrook demonstrated that INH-resistant mutants isolated in vitro were catalase-negative [69]. It will then take another 40 years to discover the genetic basis of this phenotype and decipher the main mechanism of resistance to INH: mutation in *katG,* the gene encoding the INH activator. Zhang and colleagues showed that 1) complementation of an INH-resistant *Mycobacterium smegmatis* mutant with a single copy of *katG* restored INH susceptibility [52]; 2) two highly INH-resistant clinical isolates had *katG* deletion [52]; and 3) *M. tuberculosis* INH-resistant mutants regained INH susceptibility when transformed with *katG* [70]. Since then, more than 300 *katG* mutations covering 99% of the gene's length (*katG* has 2223 base pairs (bp)) have been identified in INH-resistant laboratory and clinical strains [71]. The most frequent mutation in clinical isolates is the Ser315Thr. Actually, each of the three bases of the serine codon (AGC) can be mutated leading to Asn, Arg, Ile, Gly or Leu amino acid change. Mutations in KatG alter its catalase peroxidase and oxidase activities causing a defect in KatG's ability to activate INH [72–74]. Thus, most *M. tuberculosis* clinical strains carrying *katG* mutations are highly resistant to INH [75]. KatG enzymatic activities can also be disrupted by mutations in *furA* (*Rv1909c*), a gene encoding a ferric uptake regulation protein and a negative regulator of *katG* transcription. Isogenic strains carrying the mutations a-10c and g-7a in the intergenic region between *katG* and *furA* had reduced *katG* expression leading to a decrease in INH oxidase activity and a modest increase in INH resistance [76].

Resistance to a drug can occur through either target mutation (preventing the binding of the drug to its target) or target overexpression (titration of the drug). This is the case for the second most common mutations in INH-resistant clinical isolates: mutations in *inhA* and its promoter region. The c-15t mutation in the *inhA* promoter region increases *inhA* mRNA levels by 20-fold resulting in higher InhA protein levels and an 8-fold MIC increase in *M. tuberculosis* [60]. This mutation is found in about one third of the INH-resistant clinical isolates but more often in XDR TB cases than MDR or INH-monoresistant TB cases, suggesting that the c-15t mutation could be a marker for XDR TB [77]. In contrast to *katG* mutations, clinical strains with *inhA* mutations (either in the gene or in the promoter region) have a low INH resistance phenotype [75]. There are another 20 different mutations identified in the *inhA* promoter region [71]. Mutations in the target of INH *inhA*, an essential gene, are rare, with only 17 identified so far [71]. The first *inhA* mutation (Ser94Ala) was isolated in vitro in *M. smegmatis* during a screening for mutants co-resistant to INH and ethionamide (ETH), a second-line TB drug [57]. This mutant led to the hypothesis that InhA was the primary target of both INH and ETH [57]. Introduction of the Ser94Ala mutation into wild-type *M. tuberculosis* H37Rv was shown to be sufficient to confer INH and ETH resistance [60] and to decrease the binding of the INH-NAD adduct to InhA [60,78]. These observations strongly supported the conclusion that INH and ETH target InhA. The Ser94Ala mutation has been found in INH-resistant *M. tuberculosis* clinical isolates carrying no mutations in *katG* [75,79–81].

Mutations in many other genes have been identified in INH-resistant laboratory and clinical strains such as *kasA, mshA, ndh, nudC, ahpC,* and *nat* (the *M. tuberculosis N*-acetyltransferase) to cite a few, but very often these mutations were present in clinical strains already carrying a *katG* or *inhA* mutation or were also found in INH-susceptible strains questioning their roles in INH resistance [71].

*Area of investigation.* For the past 70 years, INH has had an essential role in TB treatment and control. One of INH downsides is its lack of activity against dormant/persistent *M. tuberculosis*. The reasons for this lack of activity in the dormant form of *M. tuberculosis* are still up for debate; however, it is known that KatG has limited activity in the dormant state [44]. With KatG being a major factor in INH resistance and a potential player in INH shortcoming in dormant *M. tuberculosis*, new InhA inhibitors that would not require activation by KatG have been actively sought [82–89]. Compounds such as GlaxoSmithKline's thiadiazole GSK693 [90] or the diazaborine AN12855 [91,92] are promising leads with good oral bioavailability, low toxicity, activity against *katG*-deficient *M. tuberculosis* and in vivo efficacy similar to INH.
