M. tuberculosis has a clonal population structure and its different strains/genotypes are notably different in their pathogenic characteristics which in turn may correlate with their enhanced transmission capacity, and wider dissemination hence epidemiologic significance. Previously, a remarkable difference between the major M. tuberculosis strains circulating in Russia was demonstrated in terms of their drug resistance association, in vitro growth features, and in the murine model^[6-8]. In particular, genetically related strains of the LAM family demonstrated differences in their virulence and lethality in mice that appeared correlated with dynamic changes in their geographic distribution in European Russia over 20 years^[8]. On the other hand, recently described MDR-associated subtypes of the ancient sublineage of the notorious Beijing genotype in the Asian part of Russia demonstrated contrasting virulence and lethality in mice^[7] furthermore one of them, a so-called “Buryat” genotype (Beijing 14717-15-cluster) was claimed to be the most lethal Russian strain characterized to date.

Here, we concisely report the most interesting interdependent insights from our just completed study with clinical implications. The C57BL/6 mice were infected with different M. tuberculosis strains; the control group remained untreated while the other group received chemotherapy and the whole experiment lasted for 200 days. Four to six months course of anti-TB drugs is currently recommended by WHO as a short course to treat MDR and XDR-TB and as a result, 85% of TB patients can be cured^[9]. Furthermore, this time is sufficiently long to study pathological changes in the lungs of resistant C57BL/6 mice during progression of TB disease, survival of infected animals as well as possible changes in microbial genes under selective pressure of the TB drugs.

We used the C57BL/6 mouse model and intravenous tail injection of the bacterial suspension according to the well-established and validated methodology previously used by us and others to study M. tuberculosis virulence^[3,7,10,11]. The M. tuberculosis strains represented the most epidemically relevant global genotypes, Beijing and LAM, and were previously comprehensively characterized by drug susceptibility testing, whole genome sequencing (WGS), and in the same murine model^[7,8]. The strains were MDR or pre-extensive drug-resistant (pre-XDR) according to the current WHO definition^[9] but differed in virulence and lethality in mice (Table 1).

All experimental procedures were carried out following the National guidelines “Rules for working with laboratory rodents and rabbits” and were reviewed and approved by the Ethical Committee of the St. Petersburg Research Institute of Phthisopulmonology (Protocol 48.1 of May 20, 2019).

C57BL/6 male mice (weight 16-18 g) were used in all experiments and were obtained from the Andreevka laboratory animal nursery (Moscow region, Russia). The animals were kept under the conditions of a certified animal facility at the St. Petersburg Research Institute of Phthisiopulmonology using NexGen Mouse IVC Cage & Rack system with built-in ventilation and air conditioning system. Before the start of the study, laboratory animals were quarantined for 14 days. A mycobacterial suspension was prepared ex tempore from three-week-old strains. The infecting dose was 10⁶ CFU in 0.2 ml of saline buffer per mouse. A suspension of mycobacteria was inoculated into the lateral tail vein of the animals.

Eight infected animal groups were formed, 20 mice per group, and two groups per each strain. The first infected group was a control one, and the other infected group was receiving an adequate chemotherapy regimen based on the knowledge of the phenotypic and genotypic resistance of the infecting strains (Table 1). The anti-tuberculosis therapy regimen consisted of 4 new-generation or repurposed drugs: moxifloxacin (7 mg/kg), linezolid (10 mg/kg), bedaquiline (6.7 mg/kg for 2 weeks, followed by 3.9 mg/kg for the rest of experiment) and perchlozone (12 mg/kg). This treatment regimen was formulated based on the doses generally recommended to treat TB in humans^[12-14]. All drugs were administered orally using atraumatic metal probes, 5 times a week and the drug solutions were prepared on the day of treatment.

The treatment started at day 10 post-infection and lasted for 177 days. After 77 and 177 days of anti-tuberculosis therapy, animals (6 per strain-determined group of mice) were euthanized and examined for different pathology-related characteristics: lung weight coefficient, lung lesions, bacterial load of the lungs, and histology.

Bacterial isolates were cultured from the lungs of the treated mice (one isolate per animal) and euthanized after 77 and 177 days of treatment. Drug susceptibility testing (DST) of the strains was carried out using the method of absolute concentrations (Order No. 109 of the Ministry of Health of the Russian Federation) and/or using the modified proportion method and automated system BACTEC MGIT 960 system (Becton Dickinson, Sparks, Md.) according to the manufacturer’s instructions and WHO recommendations^[9]. The critical drug concentrations used were 1.0 μg/mL for streptomycin, 0.1 μg/mL for isoniazid, 5.0 μg/mL for ethambutol, 1.0 μg/mL for rifampicin, 100 μg/mL for pyrazinamide, 1.0 μg/mL for amikacin, 2.5 μg/ mL for capreomycin, 2.0 μg/mL for ofloxacin, 5 μg/mL for ethionamide, 0.25 and 1.0 μg/mL for moxifloxacin, 1.0 μg/mL for linezolid, and 1.0.μg/mL for bedaquiline. Drug susceptibility testing was not performed for perchlozone because no critical concentrations have been approved and perchlozone susceptibility testing has not yet been implemented^[15].

Bacterial DNA was extracted using the Cetyltrimethylammonium bromide-based method and submitted to the whole genome sequencing (WGS). Whole genome, paired-end sequencing on the HiSeq2500 platform was done using NEBNext Ultra and PhiX Control v3 kits (Illumina). DNA libraries were prepared using ultrasound DNA fragmentation and NEBNext Ultra DNA Library Prep Kit for Illumina (New England Biolabs). Data for the M. tuberculosis sequenced genomes (fastq files) were deposited in the NCBI Sequence Read Archive (project number PRJNA1028254).

SAM-TB online tool (https://samtb.uni-medica.com/index) was used for SNP calling and genotypic detection of drug resistance. MDR, pre-XDR, and XDR phenotypes were defined according to the updated WHO definitions: MDR are strains resistant to isoniazid and rifampicin; pre-XDR - resistant to isoniazid, rifampicin, fluoroquinolone; XDR - resistant to isoniazid, rifampicin, fluoroquinolone plus bedaquiline and/or linezolid. Geneious 9.0 package (Biomatters Ltd) was additionally used for mapping the reads to the genome of reference strain H37Rv (NC_00962.3) and analysis of the drug resistance genes for possible emergence of mutant alleles.

In brief, after 77 days of anti-TB therapy, in comparison with the respective infected control groups, a decrease was recorded in lung weight coefficients, bacterial load of the lungs, and abundance of lung lesions. A complete treatment course of 177 days significantly reduced the contamination of the lungs with M. tuberculosis and the prevalence of tuberculous lung damage decreased in all treated groups.

We note that M. tuberculosis strains were isolated from the lungs of euthanized animals even after 5.5 months of treatment. From each strain-determined group, there were 6 mice/isolates analyzed per each time-point except for the group infected with low-virulent strain 6691 where 4 isolates were recovered after 5.5 months of treatment. In this sense, 177 days of uninterrupted and adequate chemotherapy could not eradicate the pathogen and provide sterile immunity but this is not unexpected. A recent human TB study demonstrated this and highlighted how ambitious (unrealistic) the WHO plan to eliminate tuberculosis is. Patterson et al.^[16] quantified viable aerosolized M. tuberculosis from TB clinic attendees in the high-burden area in South Africa and M. tuberculosis was detected in bioaerosols in a notable proportion (~30%) of confirmed TB patients after completion of the standard 6-month treatment.

We further performed DST and WGS of the M. tuberculosis isolates recovered from the lungs of euthanized treated mice after 2 and 5.5 months of treatment. The phenotypic DST revealed that the infecting strains did not change the initial resistance profiles (Table 1) and remained susceptible to the administered drugs moxifloxacin, bedaquiline, and linezolid.

The generated short sequencing reads (fastq files) were analyzed using the SAM-TB web-based pipeline to detect possible mutations in the M. tuberculosis genes known to be associated with resistance to the administered drugs. This online tool permits a comprehensive genotypic drug-resistance prediction for 17 antituberculosis drugs based on the TB Profiler tool and CRyPTIC Consortium and 100 000 Genomes Project followed by further correction and curation^[17].

We additionally mapped the sequencing reads to the H37Rv reference genome with Geneious package. We checked manually the reads mapped to the genes gyrA, and gyrB genes (fluoroquinolones resistance), rplC, rrs, rrl (linezolid resistance), Rv0678, atpE, pepQ, and Rv1979c (bedaquiline resistance). To increase the sequencing depth, we pooled together isolates from each strain-determined group together and as a result, the sequencing depth was ~1600x. No mutant alleles were identified in linezolid and bedaquiline resistance genes even at a negligible percentage. Regarding fluoroquinolone resistance, no mutations in gyrA/gyrB were found in isolates from mice infected with LAM 4542 and both Beijing strains. A gyrA mutation (D94G) was initially present in one LAM strain 7074 but no additional gyrA/gyrB mutations emerged in this strain during the experiment. Summing up, no mutations associated with resistance to the administered drugs emerged during 5.5 months of treatment even at a minor percentage. This finding is remarkable given the known emergence of resistance mutations in consecutive isolates recovered from human patients during not only long-term but also relatively short-term treatment, e.g. with bedaquiline^[18,19]. However, in real life, unfortunately, human patients may and do have multiple episodes of interrupted chemotherapy for any reason^[9,18-21], and this makes a great difference compared to the fully compliant mice receiving scientifically sound adequate chemotherapy.

Natural death of mice was also recorded during the experiment. The survival analysis showed that no mice died during 200 days in three of the four strain-determined animal groups that received chemotherapy (Figure 1). Only in one group of mice infected with hypervirulent and highly-lethal strain 396, the lethal cases were recorded but less frequently compared to the untreated group.

Figure 1.  Survival curves for eight mouse groups infected with different M. tuberculosis strains and treated or not treated with anti-TB chemotherapy.

To conclude, the correct chemotherapy regimen and complete compliance resulted in efficient treatment, the absence of emerging resistance to the administered drugs, and drastically reduced lethality in all treated groups of animals. Our findings highlight that early diagnosis including detection of resistance determinants, hence early choice of correct chemotherapy regimen and patient compliance are the key factors to control TB. Indirectly, this implies that the WHO resistance mutation catalogue^[22] should be further updated and improved to provide the highest sensitivity and specificity of the early genotypic detection of drug resistance in clinical samples from TB patients.

At the same time, lethal cases were still observed in the mice infected with hypervirulent strain 396 (Beijing 14717-15 subtype), in spite of adequate treatment. This implies that truly personalized medicine should consider not only drug resistance but other pathobiological properties of M. tuberculosis strains to begin with virulence. In this view, research focused on the so-called anti-virulence drugs is most relevant and justified.