The ddn Trp20Stop Mutation and Its Association with Lineage 4.5 and Resistance to Delamanid and Pretomanid in Mycobacterium tuberculosis

ABSTRACT

High-confidence resistance mutations for new and repurposed anti-TB drugs, such as delamanid (DLM) and pretomanid (Pa), are rare and more data are needed in order to correctly interpret the results generated by genotypic drug susceptibility testing. In this study performed on clinical Mycobacterium tuberculosis complex isolates, we report that in the Swedish strain collection the ddn mutation Trp20Stop is found exclusively among DLM and Pa resistant (Pa MIC >16 mg/L) isolates assigned to lineage 4.5.

INTRODUCTION

The approval and introduction of the nitroimidazoles delamanid (DLM) and pretomanid (Pa) in the treatment of drug resistant tuberculosis (TB) has recently improved the available options for clinicians treating patients with multidrug-resistant (MDR) or extensively drug-resistant (XDR) TB. Both DLM and Pa are prodrugs which need activation by a deazaflavin-dependent nitroreductase encoded by ddn (Rv3547) and ddn mutations have been reported in several DLM and Pa resistant TB isolates (1–3). It should however be noted that not all ddn mutations necessarily confer cross-resistance to DLM and Pa (in a previous study, a clinical strain with the ddn mutation Ser78Tyr was resistant to Pa, but susceptible for DLM) (4).

In 2021, a study which investigated the genetic diversity in ddn (as well as other loci implicated in DLM and Pa resistance) among more than 30,000 Mycobacterium tuberculosis isolates, identified the ddn stop mutation Trp20Stop in 17 isolates reported as drug susceptible (2). Eleven of them belonged to lineage 4.5 and six to lineage 5.

There is currently no established critical concentration for Pa in the MGIT 960 system, but a recent study which determined the MIC for Pa among isolates within the M. tuberculosis complex reported that none of the lineage 4 isolates included in the study (n = 113) had a Pa MIC higher than 0.5 mg/L (5). The critical concentration for DLM in the Bactec MGIT 960 system has on the other hand been set to 0.06 mg/L and susceptible isolates generally have a DLM MIC well below that concentration (6, 7).

In the M. tuberculosis complex strain collection at the Public Health Agency of Sweden, we identified the ddn mutation Trp20Stop in 11 strains (isolated between 2016 and 2020) deriving from 10 different TB patients (the whole-genome sequencing and the sequence analysis are described in Supplemental file 1). Among these isolates, nine were reported as phenotypically susceptible to the first line anti-TB drugs and two as monoresistant to isoniazid. All isolates with ddn_Trp20Stop were assigned to lineage 4.5 (Euro-American). Eight of the 10 patients infected with an isolate harboring ddn_Trp20Stop had their origin in Afghanistan. Importantly, no additional loss-of-function mutations were identified in the proposed DLM and Pa resistance genes fgd1, fbiA, fbiB, fbiC and fbiD (and ndh for DLM) among the isolates harboring ddn_Trp20Stop (8, 9).

In order to evaluate the effect of ddn_Trp20Stop on resistance to DLM and Pa, all isolates with this specific mutation (n = 11) were phenotypically tested for DLM and Pa resistance in the Bactec MGIT 960 system (details on the phenotypic drug susceptibility testing [DST] are available in Supplemental file 1). As a control group, we included 13 consecutive lineage 4.5 isolates (two isolates deriving from the same patient) lacking the ddn mutation in the phenotypic DST.

The MGIT results showed that all isolates with the ddn_Trp20Stop were resistant to DLM at 0.06 mg/L and Pa at 1.0 mg/L, while the 13 lineage 4.5 isolates with a wild type ddn were susceptible to DLM and Pa at both concentrations tested for each drug (Table 1). Subsequent Pa MIC-determination (2 to 16 mg/L) was performed on all isolates with ddn_Trp20Stop and showed that all these isolates had a Pa MIC of >16 mg/L.

TABLE 1.

ddn genotype, phenotypic DST results for DLM/Pa, additional phenotypes and lineage for the isolates which were tested phenotypically for DLM and Pa resistance^a

ID	ddn	DLM 0.03 mg/L	DLM 0.06 mg/L	Pa 0.5 mg/L	Pa 1.0 mg/L	Pa MIC (mg/L)	Additional phenotype	Lineage
H37Rv 27294	WT	S	S	S	S	<2		4.8 Euro-American
SEA201600059	Trp20Stop	R	R	R	R	>16	INH-R	4.5 Euro-American
SEA201600109	Trp20Stop	R	R	R	R	>16		4.5 Euro-American
SEA201600229b	Trp20Stop	R	R	R	R	>16		4.5 Euro-American
SEA201700050b	Trp20Stop	R	R	R	R	>16		4.5 Euro-American
SEA201800144	Trp20Stop	R	R	R	R	>16		4.5 Euro-American
SEA201900100	Trp20Stop	R	R	R	R	>16		4.5 Euro-American
SEA201900143	Trp20Stop	R	R	R	R	>16		4.5 Euro-American
SEA202000009	Trp20Stop	R	R	R	R	>16		4.5 Euro-American
SEA202000050	Trp20Stop	R	R	R	R	>16		4.5 Euro-American
SEA21-00009	Trp20Stop	R	R	R	R	>16	INH-R	4.5 Euro-American
SEA21-01048	Trp20Stop	R	R	R	R	>16		4.5 Euro-American
SEA202000011	WT	S	S	S	S	NA	INH-R	4.5 Euro-American
SEA202000020	WT	S	S	S	S	NA		4.5 Euro-American
SEA202000060	WT	S	S	S	S	NA		4.5 Euro-American
SEA202000061c	WT	S	S	S	S	NA		4.5 Euro-American
SEA202000077	WT	S	S	S	S	NA	INH-R	4.5 Euro-American
SEA202000091	WT	S	S	S	S	NA		4.5 Euro-American
SEA202000094	WT	S	S	S	S	NA		4.5 Euro-American
SEA202000130	WT	S	S	S	S	NA		4.5 Euro-American
SEA202000144	WT	S	S	S	S	NA		4.5 Euro-American
SEA202000208	WT	S	S	S	S	NA		4.5 Euro-American
SEA202000218c	WT	S	S	S	S	NA		4.5 Euro-American
SEA202000225	WT	S	S	S	S	NA		4.5 Euro-American
SEA202000229	WT	S	S	S	S	NA		4.5 Euro-American

^aMIC, MIC; WT, wild type; R, resistant; S, susceptible; INH, isoniazid; NA, not available.

^bIsolates from the same patient, no SNP difference.

^cIsolates from the same patient, differing by two SNP’s.

To delineate the population structure of the lineage 4.5 isolates, we performed a SNP (Single-Nucleotide Polymorphism) analysis based on the Ion Torrent data and generated an approximately maximum-likelihood phylogenetic tree (Fig. 1), also including 55 additional Swedish lineage 4.5 isolates without ddn_Trp20Stop (no DLM or Pa phenotypes were available for these isolates). The SNP analysis indicated a clear separation of the isolates with and without ddn_Trp20Stop.

FIG 1.

Phylogeny of 79 lineage 4.5 isolates (including 55 reference isolates without phenotypic result for DLM and Pa). Green dots indicate isolates with ddn_Trp20Stop, yellow triangles indicate isolates without ddn_Trp20Stop that were tested phenotypically for DLM and Pa resistance (control group), and yellow dots indicate the reference isolates lacking phenotypic result for DLM and Pa.

In 2021, WHO released its “Catalogue of mutations in Mycobacterium tuberculosis complex and their association with drug resistance” (10), and even though the publication constitutes an important framework for the development of genotypic assays for DST of M. tuberculosis, data on so called high-confidence mutations conferring resistance to the new and repurposed drugs (among them DLM and Pa) is still scarce. ddn_Trp20Stop was for example only reported in two isolates (both resistant to DLM) and obtained a final confidence grading of “uncertain significance.” Although the present study included a fairly low number of isolates (restricted to Sweden) and lacked allelic exchange experiments, the obtained results do indicate that ddn_Trp20Stop should be considered a high-confidence marker for DLM and Pa resistance. The results also support the introduction of new expert rules regarding loss-of-function mutations in nonessential genes implicated in DLM resistance, as already suggested in the catalogue (10).

The mutation’s confinement to lineage 4.5 in the Swedish M. tuberculosis complex strain collection, and more specifically to what seems to be an otherwise susceptible (or at most INH-monoresistant) subgroup within this lineage, hints that this specific 4.5 sublineage is intrinsically resistant to both DLM and Pa. However, since the mutation has also been reported in lineage 5 and 4.3.2 (2, 11), ddn_Trp20Stop cannot be added to the list of lineage- and species-specific resistance mutations already characterized in the M. tuberculosis complex (12). A previous study has in fact argued that the probability of spontaneous ddn mutations occurring in otherwise susceptible isolates is significant (4). The finding of identical ddn mutations, such as Trp20Stop, in different lineages might then be expected, and this further emphasizes the need for routine DST and surveillance.

Finally, the observation that the majority of the patients with lineage 4.5 isolates had their origin in Afghanistan indicates that this lineage might be prevalent among the Afghani TB cases and could constitute a reservoir of DLM and Pa resistant isolates.