Emergence of multi-drug-resistant Mycobacterium tuberculosis in Niger: A snapshot based on whole-genome sequencing

Background Among other West African countries experiencing the high endemicity of deadly tuberculosis, the situation in Niger is poorly evidenced by microbiological investigations. Methodology/Principal findings The study of 42 isolates of Mycobacterium tuberculosis from Niger by whole genome sequencing using Illumina iSeq technology yielded four M. tuberculosis lineages: Indo-Oceanic L1 (n = 1) (2.3%), East-Asian (n = 1) (2.3%), East-African Indian L3 (n = 2) (4.7%) and Euro-American L4 (n = 38) (90.4%). The sub-lineage L4.1.3 comprising 18 isolates (47.3%) was predominant, followed by the L4.6.2.2 sub-lineage (Cameroon genotype, n = 13 isolates) (34.2%). Investigating drug resistance profile for 12 antibiotics found 8/42 (19%) pan-susceptible isolates and 34/42 (81%) resistant isolates; with 40/42 (95.2%) isolates being susceptible to clofazimine-bedaquiline. Conclusions/Significance These unprecedented data from Niger highlight the dynamics of tuberculosis transmission and drug resistance in Niger and may assist tuberculosis control in this country which continues to support a high burden of tuberculosis.


Introduction
Niger is a country in West Africa with an estimated 23-million inhabitants. It is experiencing high endemicity of deadly tuberculosis, with 11,485 new and relapsing cases, including 87% of pulmonary forms, which were notified to the World Health Organization in 2019. In Niger, the incidence of multidrug-resistant (MDR)/rifampicin resistance (RR) was estimated at 2.6% (1. 2-4.5) per 100,000 inhabitants [1] and in 2019, patients between 25-34 years old were the most affected age group, with a male predominance [1]. As in most limited resource countries, knowledge of the epidemiology of tuberculosis in Niger has been based solely on recording the number of cases as well as baseline demographic characteristics of patients.
The actual incidence of tuberculosis in the various regions of Niger, therefore, is poorly known; and the Mycobacterium tuberculosis (M. tuberculosis) complex (MTBC) species and lineages responsible for laboratory-documented cases, are also poorly known, limiting preventive public health work. The lineage of cases caused by M. tuberculosis demonstrates distinct geographical associations worldwide [2,3]. The genetic diversity of circulating strains of the M. tuberculosis complex may have implications for public health, as observed previously in several studies in West African countries [4,5].
Here, we investigated a collection of MTBC isolates by WGS to gain knowledge of the M. tuberculosis species, lineages and sub-lineages circulating in Niger, and tentatively correlated them to anonymised demographic data.

Study collection
This retrospective study included all the positive cultures of mycobacteria routinely made for the diagnosis of pulmonary tuberculosis from respiratory tract specimens collected from patients suspected of having pulmonary tuberculosis, and from monitoring of multidrug resistance patients in Niger in 2016 and 2017. Respiratory tract specimens routinely addressed to the Laboratoire National de Référence des IST/VIH et de la Tuberculose, Niamey, Niger from five regions (Maradi, Niamey, Tahoua, Tillaberi and Zinder) were subjected to smear microscopic examination after auramine-O staining, decontamination using the modified Petroff method [16], and culture on Löwenstein-Jensen medium, prepared according to the manufacturer's instructions (Merck, Darmstadt, Germany) (Fig 1). Any colonies were verified by Ziehl-Neelsen staining, stored at -20˚C and shipped to MEPHI, IHU Méditerranée Infection, Marseille, France for further investigation, as described below. Colonies which did not stain by Ziehl-Neelsen staining were not included in this study. No clinical samples were specifically collected for the present study, samples have been collected as part of the patients' routine medical management. We obtained permission from the head of the laboratory to carry out any investigation deemed useful on the strains of mycobacteria from our internal library.

Isolate culture and DNA sequencing
All the laboratory investigations involving non-inactivated isolates were performed in the biosafety level 3 laboratory, at the IHU Méditerranée Infection, Marseille, France. Isolates were subcultured on Middlebrook 7H10 medium incorporating the OADC medium enrichment (oleic acid-albumin-dextrose-catalase) prepared according to the manufacturer's instructions (Becton Dickinson, Franklin Lakes, USA). Cultures were incubated in an aerobic atmosphere for up to 21 days at 37˚C and visually inspected for colonies weekly. A loopful (equivalent to ten 2-microliter inoculation loops) of colony biomass was collected in a 1.5-mL tube containing 200 μL sterile phosphate buffered saline (PBS), and was heat-inactivated at 100˚C for one hour [17]. Total DNA was extracted by vortexing the suspension with glass powder (Sigma-Aldrich, St. Louis, MO, USA) using a FastPrep apparatus (MP Biomedicals, Santa Ana California, USA) followed by a Qiagen kit with EZ1 DNA Tissue Kit (Qiagen) according to the manufacturer's recommendations (Courtaboeuf, France) and eluted in 50-μL volume [18]. The Illumina iSeq library was prepared as previously described [19]. Briefly, DNA (1 ng) was fragmented in a mix containing 5 μL of Amplicon Tagment Mix in the presence of Tagment DNA Buffer (Nextera XT Library prep Kit, Illumina) for five minutes at 55˚C, in an ABI 2720 Gen-eAmp PCR System Thermal Cycler (Applied Biosystems, Foster City, CA, USA) in a 20 μL volume. Then, 5 μL of Neutralize Tagment Buffer was added before centrifugation for one minute at 2,800g and five minutes' incubation at room temperature, then indexed and amplified in a 50 μL volume, followed by 18 cycles of PCR-index reaction, in the presence of Nextera XT Index Kit V2 (Nextera, San Diego, USA). A first purification was performed using Agencourt Ampure XP beads (Beckman Coulter, Villepinte, France) in a 0.8 ratio of beads followed by two washes with 80% alcohol and elution in 52.5 μL of RSB buffer. The library concentration was measured in Agilent 2100 Bioanalyzer (Thermo Fisher Scientific), then diluted to 100 μL RSB buffer in the presence of 10 μL-volume of Phix (50 pM). Finally, the diluted libraries (50 pM) were denatured and sequenced on the iSeq 100 sequencer (Illumina) in a single 17.5-h run providing 2x150-bp long reads.
Study strains are available in the CSUR collection at IHU laboratory with specific CSUR number of each sample (S1 Table).

Genome typing and cluster identification
After reads were stored for the delay of analysis, kaiju with default parameters [20] was used to detect for contamination level using NCBI BLAST nr-non-redundant protein database including bacteria, archaea and viruses (2021-02-24 (52 GB). Overall qualities before and after trimming sequencing reads were evaluated using FastQC [21] and multiqc [22]. Then, Trimmomatic tool [23] was used to remove residual Illumina adapters and Illumina-specific sequences. Species, lineages and sub-lineages were identified directly based on output iSeq read using the Tb-profiler (TB_v0.1.3) (https://tbdr.lshtm.ac.uk/) and MTBseq [24] with default settings by specific mapping to reference M. tuberculosis H37Rv (NC_000962.3). MTBseq was used to recover the statistical mapping data of output sequencing reads using M. tuberculosis H37Rv (NC_000962.3) as the reference genome (S1 Table). In addition, a local SNPs database constituted as previously reported, was used to identify Beijing sub-lineages [12,25,26]. Genome sequences were assembled using SPAdes version 3.13.1 [27] and annotated using Prokka version 1.12 [28]. The Roary pangenome pipeline [29] was used to generate the core-genes alignment of 42 clinical strains to create the phylogenetic tree in order to represented study strains using following parameters: 80% minimum identity for blastp, and a gene detected in 99% of isolates to be recorded as a core gene. Phylogenetic trees based on core genome were generated using FastTree 2.1.10_1 software (https://ngphylogeny.fr/tools/tool/ 271/form) using GTR model and bootstrap option with 1000 replicates.
Regarding bedaquiline-clofazimine susceptibility profile, we detected no mutation in the atpC, atpE and rv0678 genes. However, two L14R mutations (deleterious using PROVEAN with a score of -4.483) and D286G (neutral using PROVEAN with a score of -1.389) were detected in the rv1979c gene for M. tuberculosis Z204 (pre-MDR) belonging to sub-lineage L1.1.1. Other D283G mutations (deleterious using PROVEAN with a score of -2.935) were detected in the rv2535c gene for M. tuberculosis Z237 belonging to sub-lineage L4.1.3 (MDR). A specific amino acid insertion R473_V474insR (neutral using PROVEAN with a score of -0.092) was detected in the rv1979c gene, in common with M. tuberculosis sub-lineage L4.6.2.2 (S2 Table). These results showed that the M. tuberculosis sub-lineage L2.2.9 was detected as MDR.

Discussion and conclusion
This first ever WGS-based analysis of clinical M. tuberculosis isolates in Niger provided an overview of the population structure and genetic distribution of M. tuberculosis complex isolates from pulmonary and pleural fluid patients and a TB susceptibility profile in Niger.
This study offered an unprecedented yet preliminary opportunity to map the geographical distribution of the 42 M. tuberculosis isolates in Niger (Fig 1). Only M. tuberculosis sub-lineage L.4.6.2.2 was observed in the regions of Zinder (two isolates) and Tillabéri (one isolate) and only M. tuberculosis sub-lineage L4.  Table).
The Indo-Oceanic lineage consisted of one sub-lineage L1.1.1 (n = 1; one other drug-resistant type isolate), and East-African Indian L3 (n = 2; 2 sensitive isolates) in this study, and it has been reported in Niger the low prevalence of these lineages [37].
The East Asian genotype L2 (Beijing) includes one isolate of the M. tuberculosis sub-lineage L2.2.9 (B0/W148) (2.3%). To our knowledge this data has been reported for the first time in this study and suggests that the Beijing L2.2.9 strains are also commonly described as an emerging genotype in West African countries such as Niger. Moreover, recent round-trip migrations to and from China may also have played an important role in emerging and rising frequency of Beijing lineage strains in Niger.
Regarding second-line drugs, this study showed that there is no resistance to amikacin and kanamycin, which allows its use in treatment. In addition, 95.2% of isolates were susceptible to the antibiotics clofazimine and bedaquiline. These data support the repurposing of anti-leprosy antibiotics as antituberculosis treatments [51]. More, the specific amino acid insertion detected was previously reported as specific insertion for sub-lineage L4.6.2.2 (Genotype Cameroon) [52]. Despite the small sample size, our study provides insight into the genomic diversity of lineages of M. tuberculosis and the drug-resistant epidemic in Niger. This study indicates the need to be aware of infection control procedures in healthcare establishments and within the population, and to carry out genomic epidemiological surveillance in Niger.
Supporting information S1