Whole Genome Sequencing of Bulgarian Rifampicin Resistant Mycobacterium tuberculosis Strains

Whole genome sequencing of Bulgarian resistant tuberculosis strains. Abstract Introduction: The transmission of drug-resistant tuberculosis is one of the greatest challenges facing the global tuberculosis control. Aim: The aim of the study was to investigate the resent transmission of rifampicin resistant tuberculosis in Bulgaria and to describe the mutations related to the antimicrobials’ resistance using whole genome sequencing. Materials and methods: As part of an ECDC funded pilot study for evaluation of the systematic use of whole genome sequencing (WGS) of Mycobacterium tuberculosis (MTB) surveillance (EUSeqMyTB), Bulgaria provided 65 rifampicin resistant isolates over a three years’ timeframe (2017-2019) representing 87.5% of the notified rifampicin resistant cases. Drug resistance prediction and relatedness analysis of the resistant isolates was performed in collaboration with San Raffaele Scientific Institute, Milan, Italy. Results: Almost all of the isolates were identified as Euro-American lineage (96.9%); 18.5% of the isolates were found to be resistant to fluoroquinolones, but no mutations conferring resistance to bedaquiline or linezolid could be identified. Less than half (43.3%) of the isolates were clustered (<5 SNPs distance) into a total of seven national SNP-based clusters, while a total of six isolates were found to be part of different cross-border clusters. All clustered cases originated from Bulgaria. Conclusions: WGS has proven to be a reliable tool for surveillance and tracing of recent transmission of tuberculosis and has the potential for resistance prediction for most of the antituberculosis drugs.


INTRODUCTION
Transmission of drug-resistant tuberculosis (TB) is one of the greatest challenges in the TB control. Globally in 2019, almost half a million individuals had rifampicin or multidrug-resistant tuberculosis (RR/MDR-TB) (3.3% of the new TB cases and 17.7% of previously treated TB cases). [1] Although the TB incidence tend to decline in the recent years, Bulgaria remains a high priority country with notification rate of 19.1 per 100 000 in 2019, which is twice as high as that of the European Union/European Economic Area (EU/EEA) -9.6 per 100 000. [2] The bacteriologically confirmed RR/MDR-TB cases in EU/EEA for 2019 were 834 (3.4%) and 11 of them were reported from Bulgaria, which was 2.4% of all the cases with drug susceptibility testing (DST) result in the country. [2] Treatment of such patients is long (about 24 months) and expensive, a successful treatment outcome being hard to achieve -barely 45.7%. [2] Despite the implementation of reliable PCR techniques in the primary TB diagnosis settings worldwide, whole genome sequencing (WGS) stands out as the most sensitive typing method to trace transmission chains providing simultaneously drug resistance prediction to most of anti-tuberculous compounds.
In an ECDC funded pilot study for evaluation of WGS systematic use for M. tuberculosis surveillance (EUSe-qMyTB), about 75% of the rifampicin resistant M. tuberculosis isolates from 28 EU/EEA were sequenced in a three-year term (2017-2019). [3] The National Reference Laboratory of tuberculosis (NRL TB), Sofia, Bulgaria confirmed rifampicin or multidrug resistance in 65 strains of 60 TB patients for the three-year period from 2017 to 2019, collected from all over the country.

AIM
The aim of the study was to reveal the resent transmission of rifampicin-resistant tuberculosis in Bulgaria and to describe the mutations related to antimicrobials' resistance using whole genome sequencing.

MATERIALS AND METHODS
The isolate's selection criteria was any M. tuberculosis strain with resistance to rifampicin (i.e. RR-TB or MDR-TB) based on genotypic and phenotypic drug susceptibility testing, provided by the Bulgarian TB laboratory network and confirmed in NRL TB. The study period was from 2017 to 2019. In the summarized results the multiple cultures from the same patient were excluded from the relatedness analysis so that each patient was represented by a single isolate. The isolates were twice pseudo anonymized by unique codes and no identifiable personal data was gathered.
The drug susceptibility testing was performed in NRL TB Sofia by BACTEC MGIT 960 system for antituberculosis drugs as follows: rifampicin, isoniazid, streptomycin, ethambutol, ofloxacin, moxifloxacin, amikacin, kanamycin, capreomycin and linezolid to the current critical concentrations. [4,5] The Line Probe Assay Genotype MTB-DR plus/sl was used to detect the most common mutations associated with the resistance to the main first and second line antituberculosis drugs. [6,7] In order to summarize the data, we used descriptive statistical analysis.
WGS and data analysis were done at the San Raffaele Scientific Institute, Milan, Italy. For WGS-based relatedness analysis, there were performed core genome multilocus sequence typing (cgMLST) and the single nucleotide polymorphism (SNP)-based approach using the MTBseq pipeline. [8] The minimum spanning tree was calculated using Grapetree. [9] A cluster was defined as two or more isolates having a difference ≤5 SNPs.

RESULTS
For the period from 2017 to 2019, rifampicin or multidrug resistance was found in 65 strains of 60 Bulgarian TB patients. The coverage of the RR/MDR-TB strains, which were examined in Bulgaria, was estimated to be 87.5% according to ECDC notification data. [3] All rifampicin resistant TB cases were affected by pulmonary tuberculosis. The majority were males (n=41, 68.3%) with mean age 45.3 (range: 19 to 77 years). Most of the males were previously treated for tuberculosis (n=22) versus 19 new cases. Females were 19 (31.7%), with mean age 39.6 (range 23-71). Most of the females were new cases (n=14), versus 5 cases with previous treatment history.
The distribution of the RR/MDR-TB cases in the country for the three-year period is shown in Fig. 1. The most affected districts were Plovdiv (n=9), Montana (n=6), and Pernik (n=5). Most of the Bulgarian isolates were identified as Euro-American lineage (96.9%) and only 3.07% were Beijing (n=2). The relatedness analysis allowed identification of 26 patients (43.3%) as part of seven clusters (Fig. 2, Table 1). The size of the clusters varied from two to six patients. There were no cases of foreign origin involved in a cluster.
The transmission link for cluster A and one of the Montana's clusters -B, was the place of residence.
Cluster C contained family members with different households and addresses.
The northeastern cluster D had two patients with the same place of residence and two previously treated indi-  The second cluster of the Montana district (E) contained only new cases -evidence for ongoing transmission of MDR-TB. The epidemiological connection with the patient from Sofia province was not determined, but the two districts are near to one another.
Cluster F was formed from patients with various places of residence and half of them were previously treated. Two relatives were found in this group.
The Plovdiv cluster G was from previously treated individuals, originated from the town of Plovdiv and new remote patients with no clarified connection with the main cluster.
Out of sixty cases, eight (13.3%) were monoresistant to rifampicin, 86.7% were multidrug resistant tuberculosis. Resistance to fluoroquinolones (preXDR-TB) was found in 18.5% (n=12) of the isolates. The MDR-TB strains with resistance to fluoroquinolones and injectable second line drugs meeting the definition for XDR-TB (before the update in 2021) were three (5%) (Fig. 3). There was no detected resistance to bedaquiline or linezolid so far.   The most commonly detected mutations in the Bulgarian RR/MDR-TB strains were Ser450Leu in the rpoB gene, C-15T mutation in the inhA promoter region or Ser315Thr in katG gene, Met306Val in embB, and Ala90Val in gyrA gene. The resistance to pyrazinamide and streptomycin was WGS data provided comprehensive information about the resistance to the majority of the antituberculosis drugs: rifampicin, isoniazid, ethambutol, pyrazinamide, fluoroquinolones, linezolid, bedaquiline, and the injectable second-line drugs ( Table 2).
coded by various mutations. The genotype-phenotype correlation was the strongest for rifampicin and the injectable drugs (amikacin, kanamycin, capreomycin, streptomycin), and significantly weak for ethambutol.

DISCUSSION
The Beijing lineage is globally distributed due to its high virulence and transmissibility. It is most common in China and Mongolia (more than 80%) [10] and is associated with MDR-TB outbreaks all over the world, including in high burden countries like Ukraine [11] . Although the Beijing lineage keeps the second position of distribution in EU/EEA (about 30% of the sequenced strains) in Bulgaria, it represents only 3% amongst the MDR-TB cases. [3] The trend of scarce spreading remains steady with the years (3.2% for the period of 2007-2011) [12] and could be explained with the small numbers of imported TB cases in the country.
The results of GenoType MTBDR plus/sl and WGS analysis of the genetic regions related to resistance showed equal results for rifampicin, isoniazid, fluoroquinolones, and the injectable second line drugs. However, the prediction for bedaquiline or linezolid resistance by molecular method has been available only by WGS so far. The drug resistance pattern of the clinical isolates in general remains the same in the years with the distinctive mutation C-15T in the inhA promoter region, related to low-level isoniazid resistance. [13] WGS analysis failed in resistance prediction for two phenotypically isoniazid-resistant Bulgarian strains, probably because of mutation outside the analyzed genomic regions and other factors contributing to the resistance to this drug. [3,14,15] The phenotypic DST for ethambutol is considered not reliable and reproducible [16] and the contribution of each mutation to drug resistance is difficult to evaluate.
Since the testing of ofloxacin is not recommended due to dropping out of resistant-TB treatment regimens [17] and the cross-resistance levels with the other representatives of the fluoroquinolones are not utter, the results of phenotypic DST is not represented in the table above.
Since January 2021, the definition of XDR-TB has been updated [18] and already covers the resistance to rifampicin, isoniazid, fluoroquinolone(s), and bedaquiline/linezolid. There were no detected MDR-TB strains with such pattern of resistance among the sequenced Bulgarian samples in the conducted study, but a few were found in Italy. [19] Nevertheless, the considerable number of the pre-XDR-TB isolates in Bulgaria (18.5%) remains a concern.

CONCLUSIONS
The first detailed characterization of rifampicin resistant strains isolated on the territory of Bulgaria by WGS was performed. The conducted study highlighted the sustainable trend of RR/MDR-TB transmission within the country of strains with well-known features. So far, the Beijing lineage has not affected the dynamics of tuberculosis in Bulgaria. WGS has proven to be a reliable tool for surveillance but in order to assess the transmission dynamics, it cannot be used without additional epidemiological data. The method has the potential for resistance prediction after adaptation to the diagnostic requirements -sequencing from specimen, simplifying the procedure and lowering the costs.