Mycobacterium canettii is intrinsically resistant to both pyrazinamide and pyrazinoic acid

Abstract

Sir, Pyrazinamide is part of the standard short-course regimen that enables the treatment of drug-susceptible tuberculosis (TB) within 6 months (in combination with isoniazid, rifampicin and ethambutol). It is the only first-line TB drug that is being considered as part of future regimens, which have just completed Phase 2A trials [pyrazinamide in combination with bedaquiline (also known as Sirturo, TMC207 or R207910), PA-824 or PA-824/moxifloxacin]. These are designed to shorten the treatment duration for drug-susceptible TB, as well as multidrugresistant and extensively drug-resistant TB, which is paramount to reduce both the further spread of TB and the significant direct and opportunity costs to patients. The ability of these regimens to treat TB effectively might be compromised when applied to members of the Mycobacterium tuberculosis complex (MTBC) that are intrinsically resistant to one or more of the aforementioned drugs. In this context, Mycobacterium canettii is of particular importance, given that we have recently raised the possibility that this taxon might be intrinsically resistant to PA-824. In addition, it is intrinsically resistant to pyrazinamide despite lacking non-synonymous mutations in pncA (Rv2043c), encoding the pyrazinamidase that converts the prodrug pyrazinamide into its active form pyrazinoic acid. – 6 Recently, it has been shown that pyrazinoic acid inhibits the ribosomal protein S1 RpsA (Rv1630), thereby affecting transtranslation. When interrogating the recently published genome of the first M. canettii strain isolated (CIPT 140010059), we found two non-synonymous mutations in rpsA (T5A and T210A) in addition to two synonymous changes (P9P and E457E), which might constitute the basis for the intrinsic pyrazinamide resistance, particularly given that the Thr at codon 5 was previously found to be mutated in a pyrazinamide-resistant strain without pncA mutations. To test this hypothesis, we sequenced rpsA and pncA in a collection of 71 pyrazinamide-susceptible and -resistant isolates that encompassed representatives of the global diversity of the MTBC [M. tuberculosis genotypes Beijing, Cameroon, Delhi/ Central Asian, East African Indian, Ghana, Haarlem, Latin-American-Mediterranean, New-1, S-type, Turkey, Uganda I and II, Ural and X-type; Mycobacterium africanum West African I (a and b) and West African II; Mycobacterium bovis; M. canettii; Mycobacterium caprae; Mycobacterium microti; and Mycobacterium pinnipedii (see the Methods section and Table S1, available as Supplementary data at JAC Online)]. The three M. canettii strains from our reference collection shared the aforementioned mutations with CIPT 140010059 and displayed pyrazinoic acid MICs of 200 mg/L compared with 50 mg/L for the pyrazinamidesusceptible H37Rv reference strain (Table 1), suggesting that the mutations in rpsA or in an as yet unknown target of pyrazinamide/pyrazinoic acid were responsible for its intrinsic resistance. In contrast, we found that the four M. bovis isolates, which were intrinsically resistant to pyrazinamide, had pyrazinoic acid MICs that were lower or equal to the MIC for the H37Rv reference strain. This is in line with prior results that found that BCG is fully susceptible to pyrazinoic acid and becomes susceptible to pyrazinamide upon complementation of the mutated pncA (H57D) with a wild-type copy, thereby restoring the ability to convert pyrazinamide into pyrazinoic acid. Interestingly, the M. bovis strains harboured a non-synonymous rpsA mutation (A440T) that was shared with both M. caprae isolates in this study, as well as previously sequenced M. bovis isolates (AF2122/97 and 06-01518) and all BCG strains sequenced to date (China, Denmark 1331, Mexico 1931, Moreau RDJ, Pasteur 1173P2, Russia, Tice and Tokyo 172). However, none of the M. africanum, M. microti or M. pinnipedii isolates in this study had this particular mutation, suggesting it was a phylogenetically informative marker for the M. caprae/M. bovis/BCG group of animal-adapted strains, rather than a marker for pyrazinamide resistance. We found a further non-synonymous mutation in rpsA (D37A) that did not correlate with pyrazinamide resistance in one M. tuberculosis Cameroon strain. Finally, we identified the known synonymous R212R mutation in rpsA, which had been proposed as a marker for lineage 2 (East Asian) strains, whereas a second study had concluded that this was not a lineage-specific mutation. The fact that only the six Beijing isolates in our collection harboured this mutation supported the conclusion of the former study. In the Republic of Djibouti, M. canettii is responsible for 11% of active TB cases, which are clinically and radiographically indistinguishable from TB caused by other members of the MTBC. However, more data regarding its prevalence and mode of transmission are required. Our observation that M. canettii was resistant to pyrazinoic acid and harboured rpsA mutations suggested that altered trans-translation might be the mechanistic basis for the intrinsic pyrazinamide resistance of this pathogen.