Repositioning fusidic acid for tuberculosis: semisynthesis of analogues and impact of mycobacterial biotransformation on antibiotic activity

Abstract Tuberculosis (TB), caused by Mycobacterium tuberculosis (Mtb), is one of the leading causes of death globally, especially in low and middle-income countries. TB is primarily a curable disease, with chemotherapy predicated on a combination of four drugs. The increase in multiple forms of drugresistant TB is a major cause for concern, underpinning the importance of a continuous pipeline of new anti-TB agents. Drug repositioning – that is, the optimization of existing drugs for new therapeutic indications – has shown promise in expanding the therapeutic options for TB chemotherapy. Fusidic acid (FA), a natural product-derived antibiotic, has modest in vitro antimycobacterial activity. Through a multi-disciplinary approach combining aspects of chemistry and biology, this study investigated the pharmacological and physicochemical properties of FA that might be exploited for optimization of FA as a lead compound for TB drug discovery. FA is a weak carboxylic acid, and it was hypothesised that the carboxylic acid moiety limits its permeation of the complex mycobacterial cell wall. Therefore, this study aimed to identify novel FA analogues with improved permeation properties and designed to act as potential prodrugs. By modifying the C-3 hydroxyl and the carboxylic acid moiety, alkyl and aminoquinoline derivatives were covalently fused to FA through ester and amide coupling reactions to generate hybrids and/or potential prodrugs (Figure 1). Figure 1: Fusidic acid and representative ester and amide derivatives. iv Turbidimetric solubility assays revealed that most of the hybrids were 5- to 10-fold less soluble than FA. Moreover, the resultant amides and esters were found to be less potent than FA. Through the testing of 1:1 physical molar mixtures of the constituent pharmacophores, the contribution of the individual scaffolds to the antimycobacterial activity of the hybrids was evaluated. The aminoquinoline derivatives were found to reduce the overall antimycobacterial activity of the FA-aminoquinoline hybrids. The poor aqueous solubility of these compounds may have precluded the accurate determination of their antimycobacterial activity, as some of the compounds precipitated out of the aqueous media used for the drug susceptibility tests. The FA hybrids were also evaluated against two mammalian cell lines and found to be cytotoxic at very low concentrations. 1:1 physical molar mixtures of the selected aminoquinoline derivatives and FA revealed that the aminoquinoline scaffold contributed to the cytotoxicity of the hybrids. This unsatisfactory cytotoxic profile was a basis for the discontinuation of these hybrids and only the C-3 prodrugs were progressed to subsequent assays. Mtb is an intracellular pathogen, and in vitro macrophage models of Mtb infection that attempt to simulate host infection conditions are widely used to assess the efficacy of antimycobacterial compounds in disease-relevant conditions. FA and selected analogues were evaluated for their intracellular efficacy, and found to reduce bacterial burden in infected THP-1 cells at concentrations two or five times higher than their inhibitory concentrations in broth cultures. FA and three ester analogues, GKFA16, GKFA17 and GKFA61, as well as two amides, AW23 and AW25, were investigated for the role of mycobacterial metabolism in their efficacy. In Mtb cultures, the two esters were hydrolysed to FA to achieve a higher concentration of FA than that seen in incubations with FA, suggesting that they are prodrugs. In contrast, GKFA61 and the two C-21 amides were not hydrolysed to FA in Mtb cultures. Although the mode of action of FA in other organisms is well-established, there are no published reports on its antimycobacterial target and mechanism(s) of resistance (MoR). Through the generation of spontaneous resistant mutants, the target of FA and the corresponding MoR in Mtb were confirmed. A substitution mutation (H462Y) in the fusA1-encoded Elongation Factor G (EF-G) conferred resistance to FA and several derivatives. No cross-resistance was observed between FA and other standard anti-TB drugs in vitro. Notably, the fusA1 mutation conferred slight hypersensitivity to streptomycin, another v protein synthesis inhibitor, indicating its potential exploitation as a synergistic partner in TB drug combination therapy. Target-based whole cell screening was conducted using an anhydrotetracyclineresponsive knockdown strain of FA underexpressing EF-G, with consequent hypersensitivity observed for FA and three analogues. This provided further confirmation of EF-G as the target of FA in Mtb. In conclusion, these findings demonstrate that FA can be used as a template for repositioning in TB drug discovery.